article
stringlengths 0
745k
| abstract
stringlengths 0
92.7k
|
|---|---|
v4332 sagittarii ( v4332 sgr ) erupted in 1994 in a nova - like explosion which was recognized to be unusual ( martini et al .
the object showed a rapid post - outburst evolution to a cool m giant / supergiant which was uncharacteristic of a classical nova outburst .
the light curve of the object ( martini et al .
1999 ) showed a slow rise to a maximum visual magnitude of @xmath1 8.5 .
this was followed by a fast decline with the decay time for 2 and 3 magnitudes being 8@xmath21 and 12@xmath21 days respectively .
current interest in v4332 sgr was resurrected by the january 2002 eruption of v838 mon which showed the outburst characteristics of both objects to be quite similar ( munari et al .
v838 mon , iconized by its remarkable light - echo ( bond et al .
2003 ) , has been at the center of several , recent and ongoing studies .
there is a general consensus that v4332 sgr , v838 mon and m31 rv ( a red variable that erupted in m31 ; rich et al .
1989 ) may belong to a new and select class of eruptive variables whose post - outburst behavior is different on many counts from other classes of eruptive variables ( munari et al .
2002 ; banerjee @xmath3 ashok , 2002 ) .
the fundamental question regarding the cause of the explosion in such eruptive variables or quasi - novae is not completely understood .
plausible explanations for the outburst invoke merger of stars ( soker and tylenda 2003 ) and a star capturing its encircling planets ( retter and marom 2003 ) .
the cause of the outburst is still uncertain and there is a definite need to study these objects further to gain a better understanding of their nature and evolution .
+ in a recent work ( banerjee et al .
2003 ; hereafter bval ) we had shown that v4332 sgr has a very interesting near - ir spectrum at present .
several new bands of alo were detected and it was also shown that the spectral energy distribution ( sed ) of the object had undergone a striking change - a new dust shell having formed recently .
the present optical study shows that v4332 sgr is rising like a phoenix from the ashes .
our results , discussed below , show that there is a variety and richness in its spectrum that is rarely encountered in astronomical sources .
the optical observations were made on 29 september 2003 using the recently commissioned 2 m himalaya chandra telescope ( hct ) located at hanle , india .
spectroscopy and photometry were done using hfosc ( himalaya faint object spectrograph camera ) which uses a liquid - nitrogen cooled 2048x4096 pixel ccd with 15 micron square pixels as the detector .
it has a 10@xmath4x10@xmath4 unvignetted field in imaging mode and also has several grisms for low and intermediate resolution spectroscopy .
the spectra presented here were obtained at r @xmath1 870 using a slit 1.3@xmath5 wide .
the spectrum of v4332 sgr and the standard star feige 110 were obtained with exposure times of 15 and 10 minutes respectively .
the standard star spectrum was used to ratio the v4332 sgr spectrum to remove telluric / air glow lines .
the ratioed spectrum was finally multiplied by a smooth polynomial fit to the flux - calibrated spectrum of feige 110 as given in massey et al .
( 1988 ) - this gives the proper slope to the continuum of the v4332 sgr spectrum .
all spectra were wavelength calibrated using a fe - ar spectral lamp .
photometry in the @xmath0 bands , using hfosc , was done by taking multiple exposures in the @xmath0 bands .
photometric calibration was done by observing the landolt standard star sa110 -232 ( landolt 1992 ) . a cross - check for the calibration
was also done by calculating the magnitudes of another comparison star sa110 -230 .
there is a good agreement ( @xmath2 0.05 mag ) between the derived and listed magnitudes of the comparison star .
the mean air - mass at the time of observations was 1.40 and 2.3 for v4332 sgr and the standard star respectively .
extinction corrections were done using mean values of the extinction coefficient per unit airmass of 0.209 , 0.121 , 0.0823 and 0.0498 magnitudes for the @xmath6 and @xmath7 bands respectively for the hct observatory site .
the photometric and spectroscopic data were reduced using iraf .
the details of the photometry and the derived @xmath0 magnitudes for v4332 sgr are given in table 1 .
the spectrum of v4332 sgr covering the 5000 - 8300 spectral range is shown in figure 1 .
the identification of the lines and the equivalent widths of the atomic lines are given in table 2 .
some unidentified features are also listed with their observed wavelengths and equivalent widths . as can be seen , the striking feature of the spectrum is the great strength of the ki resonance doublet at 7665 and 7699 .
the blend of the nai doublet at 5890 and 5896 , although unresolved in the spectrum , is also very prominent .
we have surveyed the literature , as comprehensively as possible , for reported detections of ki and nai lines in emission .
although their occurrence is not too common ( ki detection is much rarer than nai ) they have been seen in a variety of objects viz .
comets , io , jupiter , the moon , rcb stars at minimum , a few be stars like hd 45677 , mwc 645 and p - cygni , around red giants like betelgeuse and around some n type stars . among these
, rcb stars in particular are seen to display strong nai lines during their minima ( rao et al .
1999 ) and ki to a much weaker extent .
in general , we find that the observed strength of the ki doublet in v4332 sgr is unusual and it could be among the strongest to be detected in an astronomical source .
the presence of the rbi lines in emission also appears to be rare .
although the rbi lines are blended with the neighboring tio@xmath8(3,4 ) band , we feel their identification is correct because of the good match between their observed and laboratory wavelengths .
several of the tio bands identified here , are also seen in emission in the peculiar red giant star u equulei ( barnbaum , omont , & morris 1996 ) .
we find that the average full width at half maximum ( fwhm ) of the ki doublet is greater than that of the instrument profile as obtained from two spectral lamp lines in the same wavelength region .
since gaussian fits to the ki lines and the instrument profile are found to give a good agreement , the intrinsic width of the ki lines can be obtained from where the subscripts refer to the intrinsic , observed and instrumental widths . for the observed values of @xmath10@xmath11 of 10.6 and @xmath10@xmath12 equal to 8.24 , we find that @xmath10@xmath13 is @xmath1 6.6 or 260 km / s .
the expected line width due to doppler broadening of the ki atoms , having a mass @xmath14@xmath15 , a kinetic temperature @xmath16 and a turbulent velocity @xmath17@xmath18 ( assumed to be gaussian ) is given by from equation 2 it is seen that thermal broadening will account for a negligible amount of the observed line width of 6.6 ( for e.g at @xmath16 = 1000k , the thermal broadening for the ki line is only 1 km / s ) . thus there is a large amount of velocity dispersion in the ki emitting gas .
a significant part of this broadening could be due to line - of - sight averaging of different velocity components in the ki shell in case it has an expansion or rotation velocity associated with it .
this aspect is discussed later in section 3.3 .
the ki doublet lines are expected to have a strength of 2:1 in case they are optically thin . however , the observed ratio is closer to unity for the ki doublet ( 1.1:1 ) indicating the lines are optically thick . following williams ( 1994 ) , the optical depth @xmath20 in the ki 7665 line can be calculated from where @xmath7@xmath22 and @xmath7@xmath23 are the observed intensities .
considering the observed equivalent widths of 193 and 176 for the ki lines ( table 2 ) to represent their intensities , we get a value of @xmath20 @xmath1 4.5 from equation 3 . where @xmath26@xmath15 is the number density of the ki atoms , @xmath27 is the oscillator strength for the 7665 transition ( @xmath27=0.335 ) and @xmath28@xmath29@xmath30 is the local line width given by equation 2 . using values of @xmath20 = 4.5 and @xmath28@xmath29@xmath30 = 3.4x@xmath31@xmath32 s@xmath33 ( corresponding to a intrinsic width of 6.6 ) , we derive a column density @xmath26@xmath15@xmath24 = 3x@xmath31@xmath34 @xmath35 . the value of @xmath26@xmath15@xmath24 can be used to calculate the mass of the ki region ( williams 1994 ) in case its geometry is better established in any future study - at present we are not too sure of the geometry as discussed in section 3.3 .
the sed of v4332 sgr is shown in figure 2 .
the @xmath0 fluxes have been computed after reddening corrections adopting @xmath36= 0.32 from martini et al.(1999 ) and using zero magnitude fluxes from bessell , castelli and plez ( 1998 ) .
the @xmath37 fluxes from bval are also plotted to show the sed over an extended wavelength range .
the present @xmath0 data establishes more definitely - as suggested in bval - that the hot component of v4332 sgr is well fit by a blackbody of 3250k .
the newly - formed dust component ( bval ) is also well - fitted by a 900k blackbody . associating
the 3250k component with the central star of v4332 sgr shows that it has an effective temperature corresponding to m5 type .
its luminosity class is uncertain because of distance uncertainties but as per our earlier estimate it is slightly over - luminous for a main sequence m5 object .
the sed also suggests that the central star has remained at a constant temperature of 3250k between the 1998 2mass observations ( bval ) and now .
if this is the quiescent state of the star , then it appears that the 1994 explosion has taken place on a m type star .
this gives a more definite classification , not available before , on the likely nature of the progenitor on which the outburst has taken place .
this should be an useful input for models investigating the cause of the outburst in quasi - novae .
it is difficult to conclude , based on the present data , whether v4332 sgr is a binary system .
the emission lines are in general found to be blue shifted by @xmath1 3 - a point already noticed in the high - resolution h@xmath38 line profiles of v4332 sgr obtained during outburst by martini et al .
the similar observed blueshift of the lines at two different epochs suggests the blueshift of the lines could be due to systemic motion .
radial velocity monitoring at higher spectral resolution ( than in the present studies ) would be helpful in establishing or ruling out binarity for v4332 sgr .
it may be pointed out that a hot b3v companion to the outbursting star has been reported in v838 mon ( wagner et al .
2003 ) . considering the large strength of the alkali metal lines in v4332 sgr
, we have investigated whether the m5 central object is some variant of a brown dwarf or a very low mass star .
brown dwarfs and very low mass stars show , like v4332 sgr , very strong resonance lines of the alkali metals - although in absorption ( for e.g. burgasser et al .
2003 and references therein ) .
in these objects , the nai and ki lines are the strongest while rbi and csi are progressively weaker ( rbi lines are present in our spectrum ; the csi doublet at 8521 , 8943 is not covered in our spectrum ) . further evidence that v4332 sgr type of objects could be related to very cool dwarfs
comes from the near - ir spectra of v838 mon which indicate that it could be a l giant ( evans et al .
we therefore compared the colors of v4332 sgr with a large sample of l and t type dwarfs whose @xmath39 magnitudes are available ( dahn et al . 2002 ) .
however , our results do not show v4332 sgr to have the colors of a l or t type dwarf .
the @xmath0 colors of the suggested m5 central star are not red enough . in the near - ir
, the comparison is made difficult by the contribution of the dust shell in v4332 sgr to the @xmath37 colors . while flaring activity on a brown dwarf - over a period of few hours - has been reported ( e.g. rutledge et al .
2002 ) , no instance is known of a nova - like eruption in them . since the ki / nai lines are seen in emission , and not in absorption , they can not be of photospheric origin but should originate from an extended envelope .
it is likely that some of the emission in the ki / nai lines could be caused by resonance scattering from the continuum of the 3250k central source .
in addition collisional excitation could also contribute to the observed line strength .
this is because the resonance lines have low excitation energies viz .
they are at 2.1ev , 1.61ev and 1.56ev above the ground level for nai , ki and rbi respectively .
collisions in a low temperature gas could excite the atoms to the upper levels .
further , the work of tsuji ( 1973 ) on expected molecular abundances of different species in stellar atmospheres shows that the alkali metals do not associate themselves into any molecular form at low temperatures in the range 1000 - 1500k .
thus the availability of large number of neutral atoms with low energy for their excitation could be responsible for the strong ki and nai lines .
however , we have presented here only a qualitative discussion and an estimate of the fractional contribution of collisional excitation vis - a - vis resonance scattering to the observed line strengths needs a detailed analysis . as the ki / nai lines are optically thick , yet do not show p - cygni profiles , they are unlikely to originate in a stellar wind .
two possibililities for the site of the ki emission are i ) a disc around the central source or ii ) the ejecta of the 1994 outburst .
a point in support of the latter case is that nai was seen in emission in the ejecta during the 1994 outburst ( martini et al 1999 ) .
however , the 1994 ejecta is expected to have a spatially resolved diameter of 2@xmath5 to 3@xmath5 based on the well - estimated expansion velocity of 200 - 300 km / s for the ejecta and an adopted distance estimate of 300pc to the object ( martini et al .
the present optical spectrum , suggests weakly that there could be extended emission along the slit at the positions of the ki and nai lines .
a high spatial resolution image of v4332 sgr would be invaluable to look for extended emission zone around the object .
it may be pointed out that the observed broadening of the ki lines could be accounted for by an expanding nova shell .
however , we favor the possibility of the ki gas being in an extended disc around v4332 sgr . in a recent observation from ukirt
, we have detected the @xmath40co fundamental band at 4.67 @xmath41 m strongly in emission ( again a rare phenomenon ) and also find a deep water ice absorption band at 3.1 microns ( paper under preparation ) .
the co - existence of several species , requiring successively cooler temperature conditions , in the same object viz .
neutral species like ki / nai , molecular species as alo , tio and co and finally a cold , solid - phase water ice component suggests a spatial stratification of the different species .
therefore , a possible and simplistic scenario for the geometry of v4332 could be a central source at 3250k surrounded by a dust shell .
the dust shell is either clumpy or optically thin since radiation from the central source is seen in the sed .
surrounding the central source is a disc with the atomic and molecular species in the inner parts and ice in the colder , outer regions .
if the gas in the disc has a keplerian velocity ( which would typically be @xmath1 150 km / s for a m5 star at a distance of 5 - 10@xmath24@xmath42 ) , the emission lines from the disc could get broadened substantially due to rotational motion .
this could account for the observed width of the ki lines in v4332 sgr .
incidentally the presence of a disc , if it has planetary bodies in it , creates the necessary background for the outburst mechanism of a star capturing its planets - as suggested by retter and marom ( 2003)- to become viable .
this work highlights the detection of strong emission lines of alkali metals in v4332 sgr .
the optical spectrum strengthens the idea - since nothing similar to it has been found earlier in a nova - that v4332 sgr belongs to a new class of eruptive variables . v4332 sgr is found to be an extremely interesting object , worthy of wider attention and studies .
the research work at prl is funded by the department of space , government of india .
we thank the staff of hct , hanle and crest , hosakote , that made these observations possible .
we are grateful to the referee , m. della valle , for his helpful comments .
lllllll + ut & band & exposure & integration & mag . ( error ) + & & time(s ) & time(s ) & + 12.199 & b & 50 & 250 & 20.04 ( 0.15 ) + 12.017 & v & 10 & 90 & 17.52 ( 0.14 ) + 11.955 & r & 4 & 80 & 16.31 ( 0.01 ) + 11.955 & i & 4 & 80 & 15.01 ( 0.03 ) + llllll + s.no & rest wave- & species & eq .
width + & length( ) & & ( ) + + 1 & 5197 & u.i & 40 + 2 & 5890,96 & nai & 210 + 3 & 6035 & u.i & 15 + 4 & 6086.4 & vo ( 0,1 ) ? & + 5 & 6159 & tio @xmath8@xmath4(0,0 ) & + & 6187 & `` & + & 6215 & '' & + 6 & 6403 & u.i & 16 + 7 & 6569 & tio @xmath8@xmath4(0,1 ) ? & + 8 & 6651.5 & tio @xmath8@xmath4(1,0 ) & + & 6681.1 & `` & + & 6714.4 & '' & + 9 & 6780 & u.i & 15 + 10 & 6843 & u.i & 4 + 11 & 7054.5 & tio @xmath8(0,0 ) & + & 7087.9 & `` & + & 7125.6 & '' & + 12 & 7197.7 & tio @xmath8@xmath4(1,1 ) & + 13 & 7664.9 & ki & 193 + & 7698.96 & ki & 176 + 14 & 7800.3 & rbi & 9 + 15 & 7861 & tio @xmath8(3,4 ) & + & 7907.3 & " & + 13 & 7947.6 & rbi & 5 + 14 & 7987 & u.i & 6 + | we present optical observations of the enigmatic nova - like variable v4332 sgr .
the importance of this object should not be understated since it is considered to be the possible prototype of a new class of eruptive variables .
these objects have been the subject of considerable studies at present primarily because of the spectacular eruption of v838 mon - another member of this class - recently in 2002 .
the cause of the outburst in such objects is not well understood .
our recent work has shown striking changes in the near - ir spectrum of v4332 sgr since its 1994 outburst .
the optical spectrum presented here confirms that v4332 sgr is indeed an unusual and extremely interesting object .
this spectrum , the first to be taken after a hiatus of nearly 10 years after the outburst , shows several lines in emission but is dominated by exceptionally strong emission in the resonance doublet of ki at 7665 and 7699 and to a slightly lesser strength in the unresolved nai doublet at 5890 and 5896 .
the ki lines are shown to be optically thick and considerably broadened .
we investigate the site of origin of the ki and nai emission . considering the strength of the alkali metal lines - seen at similar strength only in l and t type
dwarfs ( though in absorption ) - we discuss whether the outburst of v4332 sgr was an explosion on a l or t type dwarf .
however @xmath0 photometry does not support such a scenario but rather shows the central object of v4332 sgr to be a m - type star with a temperature of 3250k . |
given a pair @xmath4 of weights for @xmath5 with @xmath6 dominant , let @xmath7 $ ] denote the @xmath8-isotropic component of the irreducible representation @xmath9 with highest weight @xmath6 .
the _ weight ring _
@xmath10 is the graded ring @xmath11 $ ] ; it is the projective coordinate ring of the git quotient of the flag variety modulo the @xmath8twisted action of the maximal torus @xmath12 in @xmath5 .
we define the _ weight variety _ @xmath13 as @xmath14 weight varieties for arbitrary reductive lie groups ( not just those of type a ) were studied by a. knutson in his ph.d .
thesis @xcite .
knutson also studied the symplectic geometry of these spaces .
our first theorem ( theorem [ thm : krulldimensionbound ] below ) is that @xmath10 is generated in degree strictly less than the krull dimension of @xmath10 , provided that the degree - one piece @xmath7 $ ] is nonzero . the basic idea behind the proof is to show that the degree - one piece contains a system of parameters , and that the @xmath15-invariant of @xmath10 is negative .
( the @xmath15-invariant is the degree of the hilbert series , which is a rational function . )
the theorem then follows from the fact that @xmath10 is cohen macaulay .
the @xmath15-invariant is negative in all types ; however , for types other than type a , the degree - one piece sometimes fails to contain a system of parameters .
this condition is equivalent to the condition that all semistable flags lie in the supports of degree - one @xmath12-invariants ; see @xcite for a counterexample in @xmath16 .
it is well known ( cf .
@xcite ) that @xmath10 has a flat degeneration to the semigroup algebra @xmath17 of the semigroup of gelfand tsetlin patterns associated to semistandard tableaux of shape @xmath18 and content @xmath19 for @xmath20 . in particular , the ring @xmath17 is the graded ring associated to a filtration of @xmath10 by natural numbers .
generators for @xmath17 can be lifted to generators of @xmath10 , so one might hope that @xmath17 is relatively simple .
unfortunately , we find pairs @xmath21 for which @xmath17 has essential generators of degree exponential in @xmath3 .
our second main result ( theorem [ thm : gtdegreebound ] below ) is that , in the case where @xmath22 is a multiple of @xmath23 , the semigroup algebra @xmath24 has an essential generator of degree approximately @xmath25 .
this is in striking contrast to the lower bound of @xmath26 for @xmath27 that follows from our first theorem ( since the krull dimension of @xmath28 is @xmath29 ) .
this case is particularly interesting , because , via the gelfand
macpherson correspondence , it is the moduli space of @xmath3-tuples of points in the projective plane .
this is a remarkable example of how a semigroup algebra produced by a promising toric degeneration can fail to serve as an effective proxy for the original ring .
our motivation for studying the semigroup of gelfand tsetlin patterns was to imitate the method of @xcite , which studied the case of @xmath3 points on the projective line .
here one takes @xmath6 to be a multiple of the second fundamental weight @xmath30 .
it was shown in @xcite that the associated semigroup of gelfand tsetlin patterns is generated in degree @xmath31 .
we had hoped to use the same method in the case of @xmath3 points in the projective plane , but the second theorem indicates why this is not the right approach .
however , there might still be another toric degeneration that yields a bound better than the one in theorem [ thm : krulldimensionbound ] , perhaps among those discovered by caldero @xcite .
we thank harm derksen , ionut ciocan - fontanine , and mircea mustata for their invaluable advice .
in this section , we give an explicit description of @xmath32 .
let @xmath33 , and let @xmath34 denote the borel subgroup of @xmath5 consisting of the upper - triangular matrices in @xmath5 . fix a nontrivial dominant weight @xmath6 of @xmath5 .
we represent @xmath6 as a _ partition _ , _ i.e. _ , as a weakly decreasing sequence @xmath35 of nonnegative integers , with @xmath36 and @xmath37 .
let @xmath8 be a weight of @xmath5 such that @xmath38 $ ] is nonzero .
thus @xmath8 may be expressed as a sequence @xmath39 of nonnegative integers such that @xmath40 . by the borel
weil construction , the irreducible representation with highest weight @xmath6 is the finite - dimensional vector space @xmath41 where @xmath42 for @xmath43 .
the action of @xmath5 on @xmath9 is given by @xmath44 for @xmath45 .
we define @xmath46 multiplication in @xmath47 is the usual multiplication of functions @xmath48 .
as we will review in the next section , the dominant weight @xmath6 determines a line bundle @xmath49 such that the space @xmath50 of sections is isomorphic to @xmath51 .
the multiplication in @xmath47 coincides with multiplication of the corresponding sections .
the ring @xmath47 is the coordinate ring of the partial flag variety .
we define @xmath52 { \mathrel{\coloneqq}}\{f \in v_\lambda \mid \text{$f(tg ) = e^\mu(t)f(g)$ for all $ t \in { { \mathbb t}}$ , $ g \in { \mathsf{sl}}_n({{\mathbb c}})$}\}.\ ] ] we now define a @xmath8-twisted action of @xmath53 on @xmath47 . for @xmath54 of degree @xmath55 , the action of @xmath56 on @xmath57 is given by @xmath58 relative to this twisted action , the @xmath12-invariant subring of @xmath47 is exactly @xmath59.\ ] ] there is a unique @xmath60-linearization of @xmath61 .
this defines a canonical @xmath12-linearization of @xmath61 by restriction @xmath62 .
the above action of @xmath12 coincides with the canonical @xmath12-linearization twisted by @xmath8 .
a fundamental fact from the representation theory of @xmath5 is that @xmath51 has a basis indexed by semistandard tableaux of shape @xmath6 .
a _ young diagram _ of shape @xmath6 is a left - justified arrangement of @xmath63 boxes with @xmath64 boxes in the @xmath65th row .
for example , if @xmath66 , then the young diagram of shape @xmath6 is @xmath67 a _
semistandard tableaux _ of shape @xmath6 is a filling of each box in a young diagram of shape @xmath6 with a number from @xmath68 through @xmath3 such that the rows are weakly increasing and the columns are strictly increasing . for example , if @xmath69 then @xmath70 is a semistandard tableau of shape @xmath6 .
such a tableau @xmath71 determines a basis vector @xmath72 as follows .
write @xmath73 for the length of a column @xmath74 of @xmath71 .
we identify @xmath74 with the @xmath73-tuple of its entries , read from top to bottom . if @xmath75 , let @xmath76 be the function that returns the determinant of the @xmath77 submatrix consisting of rows @xmath78 and columns @xmath79 .
the basis vector @xmath80 is then defined by @xmath81 hence , in the example above , @xmath82 .
we can also describe the @xmath12-isotropic subspace @xmath38 $ ] in terms of semistandard tableaux .
the _ content _ of a tableau @xmath71 is @xmath83 if @xmath84 is the number of boxes in @xmath71 that contain the number @xmath65 for @xmath85 .
the subspace @xmath7 $ ] is the span of the @xmath86 such that @xmath71 has shape @xmath6 and content @xmath8 .
we will derive an upper bound on the degree in which @xmath10 is generated .
we generally follow the method of @xcite ( also explained in @xcite ) .
the idea is to find a homogeneous system of parameters , together with an upper bound on the @xmath15-invariant of the ring ; this yields an upper bound on the degree of a generating set .
we begin by referencing a result of @xcite and showing why this implies the existence of a system of parameters in degree one .
first we must introduce the notion of semistability . recall that @xmath34 is the borel subgroup of @xmath60 , and @xmath87 is the flag variety .
a function @xmath88 defines a section of the line bundle @xmath89 , where @xmath90 denotes the quotient of @xmath91 by the equivalence relation @xmath92 .
the projection @xmath93 is given by sending the equivalence class of @xmath94 to @xmath95 .
we define the @xmath8-twisted linearization of @xmath12 on @xmath61 by @xmath96 . given @xmath88 , we define a global section @xmath97 of @xmath61 by @xmath98 .
the map @xmath99 is an isomorphism @xmath100 .
the @xmath8-twisted torus action on @xmath61 defines an action on global sections , which coincides with the @xmath8-twisted action on @xmath9 that we earlier defined .
a flag @xmath95 is _ semistable _ if there is a positive integer @xmath55 and a @xmath12-invariant global section @xmath101 of @xmath102 such that @xmath103 .
that is , @xmath95 is semistable if and only if there is an @xmath104 and an @xmath105 $ ] such that @xmath106 .
it was shown in @xcite that we may take @xmath107 .
that is , @xmath95 is semistable if and only if there exists an @xmath108 $ ] such that @xmath106 .
we shall use this fact to show that there is a system of parameters within @xmath7 $ ] for @xmath10 .
we now recall some basic facts from commutative algebra .
our main references on cohen macaulay rings and modules are @xcite .
let @xmath109 be an algebraically closed field .
suppose that @xmath110 is a @xmath111-graded finitely - generated @xmath109-algebra with @xmath112 .
let @xmath113 denote the graded ideal generated by the positive degree homogeneous elements of @xmath110 .
then @xmath113 is the unique graded ideal such that all other graded ideals are contained within it .
a _ homogeneous system of parameters _ for @xmath110 is a set of homogeneous elements @xmath114 such that @xmath101 is the krull - dimension of @xmath110 and the ideal @xmath115 is @xmath113-primary . by (
* theorem 1.5.17 ) we have that @xmath114 is a homogeneous system of parameters if and only if @xmath110 is an integral extension of the subalgebra @xmath116 $ ] , and that this is the case if and only if @xmath110 is a finitely - generated @xmath116$]-module .
let the _ null cone _ @xmath117 be the subvariety of points in @xmath118 at which all positive - degree homogeneous elements of @xmath10 vanish .
the result of @xcite translates into the following : the elements of @xmath7 $ ] suffice to cut out the null cone set - theoretically .
that is , @xmath117 is exactly the set of points at which all elements of @xmath7 $ ] vanish .
now suppose that @xmath119 is the ideal of elements vanishing on the null cone . by the above proposition
, @xmath74 is the radical closure in @xmath47 of the ideal generated by @xmath7 \subset r(\lambda)$ ] . recall that @xmath32 is the ring of polynomials in @xmath47 that are invariant under the @xmath8-twisted action of @xmath12 .
since @xmath12 is linearly reductive , there is a canonical projection @xmath120 , called the _ reynolds operator _ , which is @xmath10-linear . following hilbert ( cf .
3.1 ) ) , we have the following result . the invariant ring @xmath10 is a finitely - generated module over the subalgebra generated by @xmath7 \subset r(\lambda,\mu)$ ] .
let @xmath121 and @xmath122 be the ideal and subalgebra , respectively , generated by @xmath7 $ ] in @xmath47 . then , since @xmath123 , we have that @xmath124 for some @xmath125 . since @xmath12 is linearly reductive , the invariant ring @xmath10 is finitely generated .
thus , there exist homogeneous @xmath126 such that @xmath127 generate @xmath10 .
suppose that @xmath128 span @xmath7 $ ] .
we have that each @xmath129 belongs to the ideal @xmath121 , and so @xmath130 for some homogeneous @xmath131 .
now we apply the reynolds s operator @xmath132 to obtain @xmath133 .
each coefficient @xmath134 is a homogeneous invariant of degree less than @xmath129 .
it follows that @xmath10 is generated as an @xmath122-module by monomials @xmath135 , where each @xmath136 .
there are only a finite number of such monomials , proving the claim .
we now have the following : ( cf .
proposition 3.2 of @xcite ) the degree - one piece @xmath7 $ ] of @xmath10 contains a homogeneous system of parameters .
we already know from the previous proposition that @xmath10 is a finitely - generated module over the subalgebra generated by its degree - one piece @xmath7 $ ] .
let @xmath101 be the krull dimension of @xmath137 .
it is easy to show that @xmath101 generic elements @xmath138 in @xmath7 $ ] are algebraically independent and that @xmath10 is an integral extension of @xmath139 $ ] .
the weight ring @xmath10 is cohen macaulay .
the argument is the same as that given in ( * ? ? ?
* corollary 14.25 ) to show that @xmath47 is cohen macaulay .
we know that @xmath10 has a grbner degeneration to a semigroup algebra of gelfand tsetlin patterns ( see section [ sec : gensoftoricring ] ) .
such semigroup algebras are the invariant subrings of polynomial rings by the action of a torus ( cf .
@xcite ) . by the theorem of hochster @xcite ,
the subring of torus invariants in a polynomial ring is cohen macaulay .
a general principle regarding grbner degenerations is that any good property of the special fiber is shared by the general fiber .
this is true in particular for the cohen
macaulay property ( * ? ? ?
* corollary 8.31 ) .
if @xmath140 is a homogeneous system of parameters for @xmath10 , then @xmath10 is a free @xmath141$]-module .
since @xmath140 are algebraically independent , the ring @xmath141 $ ] is regular , and so this proposition follows from ( * ? ? ?
* proposition 2.2.11 ) . for
a graded module @xmath142 , let @xmath143 denote the _
hilbert series _ of @xmath142 .
it is well known that , if @xmath142 is finitely generated , then @xmath144 is a rational function in @xmath145 .
let @xmath146 be the degree of @xmath144 as a rational function .
the number @xmath146 is called the _ @xmath15-invariant _ of @xmath142 . fix a homogeneous system of parameters @xmath147 $ ] for @xmath10 .
let @xmath148 $ ] be the subalgebra generated by the @xmath149 .
for brevity of notation , we will write @xmath150 . let @xmath151 denote the @xmath101-tuple @xmath140 . by theorems 13.37(5 ) and 13.37(6 ) of @xcite , @xmath152 is a free @xmath122-module , and @xmath153 but we can easily compute @xmath154 . suppose that @xmath155 .
let @xmath156 .
now , @xmath154 is the polynomial @xmath157 , where @xmath158 is the number of @xmath159 such that @xmath160 .
therefore , we have proved the following .
[ weakbound ] the ring @xmath10 is generated in degree @xmath161 .
[ strongbound ] the @xmath15-invariant of @xmath10 is negative .
let @xmath150 .
the dimension of the @xmath162-th graded piece @xmath163 of @xmath152 is equal to the number of semistandard tableaux of shape @xmath164 and content @xmath165 ; this coincides with the number of integer lattice points in the @xmath162-th dilate of the rational polytope @xmath166 ( see definition [ def : gtpolytope ] below ) . as a result of the theory of lattice point enumeration for rational polytopes ( see , _
e.g. _ , ( * ? ? ?
* chapter 4 ) ) , we may conclude that the hilbert series @xmath167 is a rational function of negative degree .
in fact , in all types , given a pair of weights @xmath21 with @xmath6 dominant , the dimension of the @xmath165-weight space in the irreducible representation @xmath168 with highest weight @xmath164 equals the number of integer lattice points in the @xmath162-th dilate of a certain polytope ( for example the string polytope associated with the reduced word for the longest weyl element ) . and
so , in all types , the @xmath15-invariant of weight rings is negative .
the above propositions imply our first theorem : [ thm : krulldimensionbound ] the algebra @xmath10 is generated in degree strictly less than the krull dimension of @xmath10 .
this follows immediately from proposition [ weakbound ] and proposition [ strongbound ] . finally , we point out that the krull dimension of @xmath10 is one more than the dimension of the git quotient of the flag variety by @xmath12 .
this is at most the dimension of the flag variety itself , which is @xmath169 . in the case of @xmath3 points in projective space @xmath170 , where @xmath6 is a multiple of the @xmath171-th fundamental weight @xmath172 for @xmath5 , the krull dimension of @xmath10 is at most @xmath173 .
a _ gelfand tsetlin pattern _ , or gt pattern , is a triangular array @xmath174 of real numbers satisfying the _ interlacing inequalities _ @xmath175 .
we express @xmath176 as a triangular array by arranging the entries as follows : @xmath177 given a semistandard tableaux @xmath71 with entries from @xmath68 through @xmath3 , let @xmath178 be the tableau obtained from @xmath71 by deleting all boxes containing indices strictly larger than @xmath179 .
hence , @xmath180 . let @xmath181 denote the shape of @xmath178 .
one obtains an integral gt pattern @xmath182 by letting @xmath183 . if @xmath71 has shape @xmath6 and content @xmath184 , then the resulting gt pattern @xmath185 has top row @xmath6 , and , for @xmath186 , @xmath187 we denote this assignment by @xmath188 .
it is easy to see that it is a bijection from semistandard tableaux of shape @xmath6 and content @xmath8 to integral gt patterns with top row @xmath6 and row sums equal to the partial sums of @xmath8 .
the gt patterns with a fixed top row and fixed row sums constitute a rational polytope .
[ def : gtpolytope ] the _ gt polytope _
@xmath166 is the set of _ real _ gt patterns @xmath189 with top row @xmath6 and with row sums @xmath190 for @xmath191 .
let @xmath192 denote the graded semigroup of integer gt patterns ( under addition ) that lie in @xmath193 for some nonnegative integer @xmath55 .
gonciulea and lakshmibai have described a grbner degeneration of the ring @xmath194 to a semigroup algebra @xmath195 as the special fiber @xcite .
it was shown in @xcite ( and also in ( * ? ? ?
* corollary 14.24 ) ) that this semigroup is isomorphic to the semigroup of integral gt patterns with top row @xmath196 for some nonnegative integer @xmath55 .
this construction also applies to the subring @xmath10 by restricting to @xmath12-invariants , as we now describe .
see @xcite for details .
the resulting degenerated ring @xmath17 has the same underlying graded vector space as @xmath10 .
the semistandard tableaux of shape @xmath196 and content @xmath197 , @xmath198 , index a basis for @xmath199 .
let @xmath200 denote the basis element indexed by @xmath71 .
the basis element @xmath201 is the leading term of @xmath202 for a certain filtration of @xmath10 ( see @xcite and @xcite ) .
the filtration has the special property that , if @xmath203 are any two semistandard tableaux , and if @xmath204 , where the sum is over semistandard tableaux , then the term @xmath205 appears on the right - hand side with coefficient equal to @xmath68 .
furthermore , all other terms @xmath206 have strictly smaller filtration level .
thus , in @xmath17 , the multiplication rule becomes @xmath207 therefore @xmath17 is isomorphic to @xmath208 $ ] , the semigroup algebra of gt patterns under addition of patterns .
given an @xmath171-tuple of rational numbers @xmath209 , define @xmath210 to be the least positive integer @xmath55 such that @xmath211 for each @xmath65 , @xmath212 .
we call this the _ denominator _ of the @xmath171-tuple . now , if some vertex @xmath176 of the polytope @xmath166 has denominator @xmath213 , then the integer point @xmath214 is an essential generator of the semigroup @xmath215 $ ] , since @xmath176 can not be written as a sum of other integral patterns in @xmath192 . in the next section
we show the existence of such a vertex with large denominator for the case where @xmath3 is a multiple of @xmath23 , @xmath216 , and @xmath217 .
suppose that @xmath22 , where @xmath218 is an integer .
let @xmath219 be a multiple of the third fundamental weight for @xmath5 .
thus , as a partition , @xmath220 . now let @xmath8 be the `` democratic '' weight dominated by @xmath6 .
that is , we represent @xmath8 by the composition @xmath221 . with this choice of @xmath6 and @xmath8 , the projective variety
@xmath222 is the moduli space of equally weighted @xmath223-tuples of points in projective space @xmath224 ( see @xcite for more details ) .
we now construct a gt pattern that we claim will be a vertex of @xmath225 .
define the sequences @xmath226 and @xmath227 by the coupled recurrence relations @xmath228 solving this system of recurrence relations yields the closed - form expressions @xmath229 let @xmath230 .
we will construct a triangular array @xmath176 by filling in the entries of @xmath176 in blocks from the upper left to the lower right using the values @xmath231 and @xmath232 .
begin by filling the entries in the upper left of the triangular array as follows .
@xmath233{$x_{1n}$ } & & \makebox[\xij]{$x_{2n}$ } & & \makebox[\xij]{$x_{3n}$ } & \\ & & & & & \\ & \makebox[\xij]{$x_{1,n-1}$ } & & \makebox[\xij]{$x_{2,n-1}$ } & & \makebox[\xij]{$x_{3,n-1}$ } \\ & & & & & \\ & & \makebox[\xij]{$x_{1,n-2}$ } & & \makebox[\xij]{$x_{2,n-2}$ } & \\ & & & & & \\ & & & \makebox[\xij]{$x_{1,n-3}$ } & & \end{matrix } \qquad = \quad \begin{matrix } \makebox[\xij]{$k$ } & & \makebox[\xij]{$k$ } & & \makebox[\xij]{$k$ } & \\ & & & & & \\ & \makebox[\xij]{$k$ } & & \makebox[\xij]{$k$ } & & \makebox[\xij]{$k-1 $ } \\ & & & & & \\ & & \makebox[\xij]{$k$ } & & \makebox[\xij]{$k - \frac{1}{2}$ } & \\ & & & & & \\ & & & \makebox[\xij]{$k$ } & & \end{matrix}\ ] ] we then proceed from the upper left to the lower right of the triangular array by filling in blocks of entries as follows . for @xmath234 ,
let @xmath235{$x_{3 , n - 2j}$ } & \\ & & & \\ & \makebox[\xij]{$x_{2 ,
n - 2j - 1}$ } & & \makebox[\xij]{$x_{3 , n - 2j - 1}$ } \\ & & & \\
\makebox[\xij]{$x_{1 , n - 2j - 2}$ } & & \makebox[\xij]{$x_{2 , n - 2j - 2}$ } & \\ & & & \\ & \makebox[\xij]{$x_{1 , n - 2j - 3}$ } & & \end{matrix } \qquad = \qquad \begin{matrix } & & \makebox[\xij]{$t^{(1)}_{j}$ } & \\ & & & \\ & \makebox[\xij]{$t^{(1)}_{j}$ } & & \makebox[\xij]{$t^{(1)}_{j}$ } \\ & & & \\
\makebox[\xij]{$t^{(2)}_{j}$ } & & \makebox[\xij]{$t^{(1)}_{j}$ } & \\ & & & \\ & \makebox[\xij]{$t^{(1)}_{j}$ } & & \end{matrix}\ ] ] if @xmath109 is even , the final entries at the bottom of the array are filled in as follows .
@xmath236{$2 - t^{(1)}_{n}$ } & & t^{(1)}_{n } & \\ & & & \\ & 1 & & \end{matrix}\ ] ] on the other hand , if @xmath109 is odd , then the final entries are filled in as follows : @xmath237{$3 - 2t^{(1)}_{n}$ } \\ & & & \\ & t^{(1)}_{n } & & \makebox[\tij]{$2 - t^{(1)}_{n}$ } & \\ & & & \\ & & 1 & & \end{matrix}\ ] ] all the remaining entries of the triangular array are assigned the value 0 . these are all straightforward consequences of the closed - form expressions and for @xmath231 and @xmath232 , respectively , so @xmath176 is a gt - pattern .
thus , to show that @xmath239 , we need only establish that the row - sums of @xmath176 are correct .
this amounts to showing that @xmath241 for @xmath242 .
these equalities may be shown using induction and the recursive definition of @xmath231 and @xmath232 .
it is clear from equation that @xmath231 has denominator @xmath243 when written as a reduced fraction .
hence , @xmath176 has denominator @xmath238 , as claimed .
it remains only to show that @xmath176 is a vertex of @xmath225 .
we prove this by showing that , for any triangular array @xmath244 , if @xmath245 , then @xmath246 .
this is most easily seen by partitioning the entries of @xmath176 so that entries that are equal and adjacent are grouped together .
we call each group of entries in this partition a _ tile_. see figure [ fig : tiling ] for a depiction of the case when @xmath109 is even .
each tile is labeled with the value shared by the entries that it contains .
suppose that @xmath247 .
note that , after the addition of @xmath248 , the entries in each tile must still share a value , and the row - sums must be unchanged .
we prove inductively that the entries in each tile can not have changed , proceeding from the upper left to the lower right .
the entries in the tile labeled @xmath249 can not have changed because the top row is fixed .
for the same reason , the @xmath250 entries in @xmath176 are also fixed .
proceeding by induction , the entries in the tile labeled @xmath231 can not have changed because there is a row on which this is the only tile besides the tile labeled @xmath251 and the tile of @xmath250s , which have already been fixed .
hence , the entries in all the tiles labeled @xmath231 , @xmath252 , are fixed under the addition of @xmath253 .
finally , for @xmath254 , the tile labeled @xmath232 lies on a row in which the other entries , @xmath231 , @xmath255 , and 0 , have been shown to be fixed , so the entry in this tile is also fixed under the addition @xmath248 .
therefore , we conclude that @xmath246 , so that @xmath176 is a vertex , as claimed .
r. p. stanley , _ enumerative combinatorics . vol . 1 _ , cambridge studies in advanced mathematics , vol .
49 , cambridge university press , cambridge , 1997 , with a foreword by gian - carlo rota , corrected reprint of the 1986 original . | a weight ring in type a is the coordinate ring of the git quotient of the variety of flags in @xmath0 modulo a twisted action of the maximal torus in @xmath1 .
we show that any weight ring in type a is generated by elements of degree strictly less than the krull dimension , which is at worst @xmath2 . on the other hand , we show that the associated semigroup of gelfand tsetlin patterns can have an essential generator of degree exponential in @xmath3 . |
in the current framework of galaxy evolution , galaxies and black holes are intimately coupled in their formation and evolution .
the masses of galactic bulges and their central black holes ( bhs ) in the local universe follow a tight relation ( e.g. * ? ? ?
* ) with only 0.3 dex scatter .
currently it is not clear how this relation comes about and if and how it evolved over the last 13 gyrs , but basically all semi - analytic models now include feedback from active galactic nuclei ( agn ) as a key ingredient to acquire consensus with observations ( e.g. * ? ? ?
* ; * ? ? ?
in these models it is assumed that black hole growth by accretion and energetic re - emission from the ignited agn back into the galaxy can form a self regulating feedback chain .
this feedback loop can potentially regulate or possibly also truncate star formation and in this process create and maintain the red / blue color magnitude bimodality of galaxies . in this light
, any galaxy with an abnormal deviation from the @xmath10@xmath11-relation will be an important laboratory for understanding the coupling mechanisms of black hole and bulge growth .
it will set observational limits for these models , and constrain the time - lines and required physics involved . since the early work by @xcite on qso host galaxies with the _ hubble space telescope ( hst ) _ and the subsequently resolved dispute about putatively `` naked '' qsos @xcite , no cases for qsos without surrounding host galaxies were found when detection limits were correctly interpreted .
only recently the qso he04502958 renewed the discussion , when @xcite made a case for a 6@xmath12 too faint upper limit of the host galaxy of he04502958 with respect to the @xmath10@xmath11-relation . in light of a number of competing explanations for this , the nature of the he04502958 system
needs to be settled .
the qso he04502958 ( a.k.a .
iras 045052958 ) at a redshift of @xmath0 was discovered by @xcite as a warm iras source .
he04502958 is a radio - quiet quasar , with a distorted companion galaxy at 15 ( = 6.5 kpc ) distance at the same redshift , likely in direct interaction with the qso @xcite .
the combined system shows an infrared luminosity of an ultraluminous infrared galaxy ( ulirg , @xmath13 l@xmath4 ) .
he04502958 was observed with the hubble space telescope ( hst ) and its wfpc2 camera @xcite in f702w ( = @xmath14 band ) and acs camera @xcite in f606w ( = @xmath15 band ) , both observations did not allow to detect a host galaxy centered on the quasar position within their limits ( figure [ fig : allwave ] , left column ) .
@xcite estimated an expected host galaxy brightness if he04502958 was a normal qso system that obeyed the @xmath10@xmath11-relation in the local universe and given a bh mass estimate or luminosity of the qso .
they concluded that the acs f606w detection limits were six times fainter than the expected value for the host galaxy , which qualified he04502958 to be very unusual .
@xcite sparked a flurry of subsequent papers to explain the undetected host galaxy to black hole relation . over time three
different alternative explanations have been put forward and were substantiated : 1 .
he04502958 is a normal qso nucleus , but with a massive black hole residing in an under - massive host galaxy .
the system is lying substantially off the local @xmath10@xmath11-relation ; the host galaxy possibly hides just below the f606w detection limit @xcite .
the host galaxy is actually absent , he04502958 is a truly `` naked '' qso , by means of a black hole ejection event in a gravitational three body interaction or gravitational recoil following the merger of he04502958 with the companion galaxy @xcite .
the original black hole mass estimate was too high @xcite and is in fact @xmath210 times lower . with comparably narrow ( @xmath21500 km / s fwhm ) broad qso emission lines the qso could be the high - luminosity analog of the class of narrow - line seyfert 1 galaxies ( nlsy1 ) .
the host galaxy could be normal for the black hole mass and be absolutely consistent with the acs upper limits .
in this article we present new data initially motivated by the still undetected host galaxy and by the possibility that the host galaxy might be obscured by substantial amounts of dust .
we want to investigate the overall cool and warm dust properties of the system , using new near infrared ( nir ) and mid infrared ( mir ) images .
the f606w acs band is strongly susceptible to dust attenuation , and dust could have prevented the detection of the host galaxy in the optical . with new nir data we look at a substantially more transparent wavelength . at the same time
the new infrared data is meant to localize the source(s ) of the ulirg emission .
three components are candidates for this : the agn nucleus , the host galaxy , and the companion galaxy .
our nir data allow to trace star formation and the mir image traces the hot dust in the system .
we present the new data and interpret it in the view of the so far collected knowledge from x - ray to radio - wavelengths that was built up since the article of @xcite . throughout we
will use vega zero - points and a cosmology of @xmath16 , @xmath17 , and @xmath18 , corresponding to a distance modulus of 40.84 for @xmath0 and linear scales of 4.312 kpc/.
up to now the only existing infrared observations on he04502958 were from the 2mass survey in the near infrared @xmath19 , @xmath20 , and @xmath21 bands at @xmath24resolution , and in the mir from the iras mission @xcite at 12 , 25 , 60 , and 100 @xmath1 m with about 4 resolution .
both surveys do not resolve the different individual components of the system ( qso , companion galaxy , foreground star ) .
de grijp et al .
( @xcite ) noted that the he04502958 system is showing the mir / fir luminosities of a ulirg system , but it was not clear which components of the system are responsible for this emission due to the coarse iras resolution .
we want to localize the dust emission in two ways : ( a ) a direct observation of the hot dust component at 8.9@xmath1 m ( rest - frame ) with the visir imager at the eso vlt .
( b ) a localization of dust in general by combining new hst near infrared and the existing acs optical data . for this purpose we obtained hst nic2 imaging in the rest - frame @xmath19-band at @xmath21.3@xmath1 m . in the near and mid
infrared the he04502958 system clearly has a spectral energy distribution ( sed ) that is composed of more than a single component : in figure [ fig : iras_sed ] we model the iras and visir flux densities with a composite sed of a quasar plus a star forming component .
for the quasar we test the median and 68 percentile reddest quasar sed from @xcite , and for the star forming component an arp220-like starbursting sed , but we also tried a medium star formation m82 sed , both from @xcite .
the median quasar sed plus arp220 can reproduce the data at all wavlengths , except at observed 25@xmath1 m , where it leaves a small mismatch .
the 68 percentile reddest sed on the other hand creates a perfect match also there . for both cases the flux predicted for the companion galaxy at 11.3@xmath1
m lies below the detection limit as observed .
milder , m82-like star formation can be ruled out on the same basis , as it predicts a detection of the companion also at 11.3@xmath1 m both with the information from the mid infrared , as well as when extrapolating the observed @xmath20-band flux .
@xcite match a simple model of two black - body emission curves to the four iras points , yielding a cool dust component heated by star formation and a warm dust component which can be attributed to intense agn emission ( see section [ sec : discussion_ulirg ] ) .
while it is not possible to spatially resolve the system at fir wavelengths with current telescopes , we aim for the highest wavelength where this is currently possible , in order to localize the warm emission component and test whether this comes solely from the ( optically visible ) qso or from extra sources .
the sed of he04502958 in the mid - infrared : shown are the iras flux density measurements from @xcite _
( open circles ) _ , our visir data point _ ( filled circle ) _ and upper limit on the companion galaxy _ ( arrow ) _ and overlaid composite agn plus starburst seds _ ( lines)_. for the quasar nucleus
we use the median _
( green dashed line ) _ and 68 percentile reddest seds _ ( black dashed line ) _ from @xcite , the starburst _ ( red solid line ) _ is a model for arp220 by @xcite . the median quasar plus arp220 sed _ ( green solid line ) _ can explain the data except for a slightly too low value at the observed 25@xmath1 m point , but with the 68 percentile sed _
( black solid line ) _ the match is perfect . the predicted flux of the companion galaxy where the star formation of the system is located _ ( bar ) _ lies below our detection limit , consistent with the data .
milder star formation templates as e.g. m82 can be ruled out , since they predict too high fluxes for the companion also from the observed @xmath20-band data which should be visible in the visir image . ]
the observations were performed using visir , the eso / vlt mid - infrared imager and spectrograph mounted on unit 3 of the vlt ( melipal ) .
visir gives a pixel size of 0075 and a total field - of - view of 192 .
the diffraction limited resolution is 035 fwhm .
standard `` chopping and nodding '' mid - infrared observational technique was used to supress the background dominating at these wavelength .
all the observations were interlaced with standard star observations of hd 29085 ( 4.45 jy ) and hd 41047 ( 7.21 jy ) .
the estimated sensitivity was 4 mjy/10@xmath22/1h .
imaging data were obtained on the 12th of december 2005 in service observing mode , through the pah2 filter centered on 11.3 @xmath1 m having a half - band width of 0.6 @xmath1 m .
weather conditions were very good , optical seeing was below 1 , and the object was observed always at an airmass of 1.15 , which resulted in a diffraction limited image of 035 resolution .
chopping / nodding parameters were 8/8and 0.25 hz/0.033 hz .
the total time spent on - source was 1623 s. the data were reduced using a dedicated pipeline written in idl , which does the chopping / nodding correction and removes the spurious stripes due to detector instabilities @xcite .
the reduced data were finally flux - calibrated using the two reference stars as photometric calibrators . the error on the photometry due to variations of the atmospheric transmission are estimated to be less than 2% ( 3@xmath22 ) .
the acs @xmath15-band is too blue to penetrate any substantial amount of dust . with the scenario of a dust enshrouded host galaxy in mind , we acquired new hst nicmos data ( nic2 with 0075 plate scale ) in the f160w @xmath20-band ( program # 10797 , cycle 15 ) to reduce the dust attenuation by a factor of 3.5 in magnitude space .
a total of 5204s integration on target was forcedly split into two observation attempts due to telescope problems , and carried out in july 2006 and 2007 .
these yielded two sets of data with 2602 s integration each , but slightly different orientations . in order to minimize chromatic effects
, we also observed a point spread function ( psf ) calibrator star ( eis j033259.33274638.5 ) with the sed - characteristics over the f160w filter bandpass similar to a mean qso template .
we do not know the actual sed of he04502958 itself , as no nir imaging or spectroscopic data of the system with high enough spatial resolution exist to date . as the stellar type yielding the likely most similar psf we found k4iii , by comparing the psfs predicted by the tinytim package @xcite .
the only cataloged stars faint enough to not immediately saturate were observed by the eso imaging survey ( eis , * ? ? ?
* ) located in the e - cdfs , and had to be observed at 6 months distance in time to he04502958 . since we also want to minimize the psf variation due to differences in observing strategy , we applied the same dither patterns for both he04502958 and the psf star . due to the absolute pointing accuracy of hst the centroid location of the star relative to the chip
is shifted about 15 pixels ( 11 ) from the qso centroid towards the companion galaxy .
data reduction and combination of the individual frames were carried out using a mix of stsci pipeline data products , pyraf , and our own procedures in midas and fortran .
the resulting image is shown together with the analysis in figure [ fig : dataimages]a .
two parts of the team analyzed the combined images in complementary ways , by decomposition of the components using two - dimensional modeling and by image deconvolution . in order to detect a putative faint host galaxy underneath the bright qso nucleus we require a precise knowledge of the psf .
the psf will vary spatially , with the energy distribution in the filter as well as temporally , with a changing effective focus of the telescope due to changing thermal history .
we opt for a double approach : first , we observe the separate psf star with the properties described in the last section ( and see fig .
[ fig : dataimages]b ) .
secondly , we also have the foreground star available that is located at 18 distance from the qso to the north - west .
it is classified as a g star @xcite .
its on - chip distance to the qso will leave only room for small spatial variations , but its sed in the @xmath20-band likely will not perfectly match the sed of the qso .
it is difficult to assess the psf uncertainty at the position of the qso . in principle
we have a combined effect of color , spatial , and temporal variation , but only one bit of information : the difference between the foreground star and the psf star .
we thus model the expected difference in the shape of these two stars with tinytim and then compare their actual observed shapes .
this shows that the foreground star should be slightly narrower than the observed psf star , which is consistent with psf star s later , redder spectral type and an increase of psf width with wavelength .
we observe this effect also in the data , however somewhat stronger . a temporal variation
can thus not be separated and ruled out . in any case
we conclude that the psf star is wider and thus will yield more conservative ( = fainter ) estimates for a qso host galaxy , while in case of a non - detection the foreground star will yield brighter upper limits . for two - dimensional modeling of the system we use galfit @xcite . in order to quantify the psf uncertainty for this process , we first let galfit fit a single point source , represented by the psf star , to the foreground star . in this process
we use an error map created from the data itself and we add the sky as a free parameter .
we minimize the influence of the nearby qso on the foreground star by first fitting the former with a single point source as well , removing its modeled contribution , and mask out the remaining residuals starting at 09 from the star .
the psf created in this way is shown in figure [ fig : dataimages]c .
this image is fed into the modeling process of the psf star , or later the qso / host / companion system .
the residual flux in this process is of the order of 3% of the total , inside the 05 radius aperture where most apparent residuals are located , the absolute value of the residuals in the same region is 14% .
this means that it will be generally impossible to detect any host galaxy of less than 3% of the total flux of the qso , and it will even be difficult to isolate a somewhat brighter smooth galaxy in the non - smooth residuals .
this level of residuals is consistent with experience from the hst acs camera , where we find that due to psf uncertainties 5% of the total flux are approximate detection limits for faint host galaxies ( * ? ? ?
* jahnke et al . in prep . ) .
including the structured psf residuals we will only consider a host galaxy component as significant if it has clearly more than 35% of residual flux inside an 05 radius of the qso , or that shows up as a non co - centric structure above the noise outside this region . in absolute magnitudes and related to the qso these limits
correspond to the following : inside an 05 radius of the qso we can hide a galaxy co - centric with the qso of at least @xmath23 ( for the 3% case ) or @xmath24 ( for 5% ) .
we detect a single unresolved point source in the visir field - of - view with a flux density 62.5 mjy at observed 11.3@xmath1 m ( figure [ fig : allwave ] , right column ) .
this compares to 69.3 mjy in the iras 12 @xmath1 m channel .
there is no second source detected in the field down to a point source sensitivity of at least 3 mjy at the 5@xmath22 level .
extended sources of the visual size of the companion galaxy have a 5@xmath22 detection limit of 5.5 mjy . with only one source in the total 192 visir field three optical sources
have in principle to be considered as potential counterparts : the qso nucleus , the companion galaxy , and the foreground star .
however , the star is a g spectral type and can thus be safely ruled out .
we find that the initial position of the mir point source as recorded in the visir image header comes to lie between the qso and the companion , somewhat closer to the qso . to clarify this we conducted an analysis of the pointing accuracy of visir testing the astrometry of a number of reference stars observed with visir at different epochs .
the two results are : ( 1 ) in all cases the offset between targeted and effective ra , dec is less than 1 rms , but ( 2 ) there is a systematic offset of 0.15s in ra recorded in the fits header , so the true positions need to be corrected by 0.15s in ra .
this correction places the mir point source exactly onto the locus of the qso in the hst acs images .
it is thus clearly the qso nucleus that is responsible for all of the 11.3 @xmath1 m emission . to extract information on the host galaxy , we use three different methods to remove the flux contribution from the qso nucleus
first , we make a model - independent test for obvious extended emission : in a simple peak subtraction we remove a psf from the qso , scaled to the total flux inside two pixels radius around the qso center .
this is a robust approach that is independent of specific model assumptions and quite insensitive to the noise distribution in the image @xcite . as a result
, the peak subtracted image shows no obvious extended residual , i.e. host galaxy , centered on the qso , when using the psf star as psf . as a second step we use on the one hand galfit to model the 2-dimensional light distribution of the he04502958 system and decompose it into different morphological components . on the other hand we use the mcs deconvolution method @xcite to mathematically
deconvolve the system to a well defined and narrower psf .
the procedure we follow is based on the one described in @xcite .
for galfit we use the two empirical psfs , for mcs deconvolution we construct a number of combinations of empirical psf and tinytim models including very red dust - like sed components . while these two approaches are complementary in method , their results agree as can be seen in figure [ fig : dataimages ] : the inner part of the qso inside of 05 radius is consistent with a point source within the psf uncertainties , but there is extra flux present outside of this radius .
the structure of the psf removal or deconvolution residuals points to a substantial mismatch between shape of the qso nucleus and the separately observed psf star , but also to too simple models of tinytim . in order to remove obvious residual psf structure a very red sed needs to be assumed , which at this point can not be discriminated from a marginally resolved red component on top of the agn point source .
however , in light of the non - average properties of this qso , a mean qso sed is also not expected . in the following we present our results in more detail and
focus on the galfit results , since it allows a more direct estimate of the significance of detected structures .
a comparison of the original and point source - removed images in the optical and nir , and the mir image are shown in figure [ fig : allwave ] .
we use galfit to perform a number of different model fits . in all of them the companion star and qso nucleus are described by a pure point source , while the companion galaxy is fit with one or two srsic .
it reverts to an exponential disk profile typical for spiral galaxies for @xmath25 and for @xmath26 it becomes a de vaucouleurs profile found for many elliptical galaxies . ]
components with free axis ratio , or left unmodelled .
we also attempt to add another srsic component for the putative host galaxy .
we always leave the srsic parameter @xmath27 free , although the companion galaxy is too complex and the putative host galaxy too faint for @xmath27 to be interpreted physically . with the psf star used as
psf galfit finds a result consistent with the peak subtraction .
a positive residual of @xmath28 inside a @xmath21 radius aperture has a flux below 2% of the 13.05 mag of the qso itself ( see figure [ fig : dataimages]e+f ) . even though we choose an aperture larger than in our calculation in section [ sec : psfuncertainty ] , we receive a value far below our significance limit , so no significant co - centered host galaxy is seen in this way . if we use the foreground star as psf ( fig .
[ fig : dataimages]g ) we find as expected a residual flux that is slightly higher than before , and consistent values for two different approaches : first , for a pure psf fit to the qso location , integrating the flux of the residual within a 1 radius aperture around the qso , except along the se nw - axis where we expect residual flux from foreground star and companion galaxy .
secondly , we get a similar flux for a fitted additional host galaxy srsic component .
these two approaches yield a magnitude of @[email protected] and 16.2 , respectively , for the host , @xmath21.5mag brighter than for the psf star fit .
@xmath20@xmath216 corresponds to @xmath26% of the 13.05 mag of the qso nucleus .
again the qso residual shows substantial structure as reported in section [ sec : psfuncertainty ] .
it consists of nested rings of positive and negative flux , typical signs of a close but different width between the psf we use and the actual one .
the bulk of structure is contained in the innermost 05 radius and contains 2/3 of the residual flux .
the remaining residual of 2% of the total flux outside this radius is again insignificant , and no main body of the host galaxy co - centered with the quasar is found which satisfies our significance criterion .
going back to the psf residuals that we quantified earlier on , we detect no co - centered host galaxy at a level above 3% of the flux of the quasar nucleus , corresponding to an upper limit of @xmath29
. however , after removal of the point source , a feature becomes clearer , what we dub the `` ne - extension '' .
this faint structure extends from the qso to the n e , and it can be traced starting at the edge of the strong psf residuals at 06 ( 2.5 kpc ) n e of the nucleus ( figures [ fig : dataimages ] and [ fig : ne - extension ] ) .
some signs of it are already visible in the optical , when going back to the the f606w image ( * ? ? ?
* see also fig .
[ fig : allwave ] ) , but it is much more pronounced in the new @xmath20-band data compared to the @xmath15-band .
the ne - extension is possibly part of a tidal arm similar to the arm towards the south of the companion , already described by @xcite , but our @xmath20-band image shows it to be clearly disjoint from the companion galaxy .
due to its proximity it is very likely associated with the qso , even though it is clearly not its main body .
it is unlikely that the ne - extension is just a gas cloud with star formation induced by the radio jet in the system , since it lies at least 50@xmath30 from the jet direction @xcite .
it is also unlikely a chance superposition of a gas cloud with emission line gas , as seen by @xcite , since the observed @xmath20-band does not contain any strong enough line .
the ne - extension contains non - negligible flux far above the noise of the background and is unaffected by qso residuals and independent of the psf used .
we estimate its brightness at @xmath31 using an aperture encompassing all visible extension outside the qso nucleus residual .
the same region in the acs image has @xmath32 , so @xmath33 a slight zoom into the inner region of he04502958 to show the newly found `` ne - extension '' of the qso .
we removed the star and the qso using the psf star as psf .
an extension to the n
e is visible ( marked with _
red brackets _ ) at a distance of 0615 that is clearly not due to psf residuals a very similar result is seen when using the narrower foreground star as psf ( fig .
[ fig : dataimages]e ) , or mcs deconvolution ( fig .
[ fig : dataimages]d ) .
this structure is disjoint from the companion galaxy so very likely belongs to the qso host galaxy itself .
the estimated brightness is @xmath31 .
the image size is 45 on the side . ] in summary , we detect no significant host galaxy that is co - centered with the qso .
we conclude this from the size and shape of the residuals underneath the qso in comparison to the `` psf star minus foreground star '' subtraction residuals we discussed above .
the ne - extension , however , that can be seen outside of the residuals of the qso nucleus , is a real and significant emission structure and it is very likely associated with the main part of the host galaxy . in the acs
@xmath15-band the companion galaxy located 15 to the s
e appears clumpy , with several bright knots as well as lower surface brightness in the center . @xcite
even call the companion a `` collisional ring galaxy '' . with the nicmos @xmath20-band
we get a substantially different picture .
the galaxy at @xmath34 is still asymmetric , with tidal extensions , but contrary to the visible wavelengths it is smooth and shows a pronounced center : clear signs for substantial dust , distributed not smoothly but unevenly and clumpily , with concentration towards the center that only shows up in the optical ( fig .
[ fig : companion ] ) .
the complexity of the companion is manifested in that there is no good description with neither one or two sersic components , when the azimuthal shape is restricted to ellipses .
the sersic index of the companion is around @xmath35 for a single sersic component , and @xmath36 if two sersic components are used .
taken at face value , both cases point to a more disk- than bulge - like companion , but a substantial fraction in flux is containted in the non - symmetric distorted part of the companion and this should be the main description of the companion .
more complex descriptions were put forward , with either a proposed additional faint agn hosted by the companion galaxy @xcite , explosive quasar outflows @xcite , or quasar - induced star formation @xcite .
zoom on the companion galaxy . as a difference to the @xmath37-band ( fig .
[ fig : allwave ] ) the galaxy has a pronounced peak of emission and no ring . the light in the optical is obviously attenuated by dust , very strongly in the center where the dust is optically thick , less in the outer regions .
image size is 3 on the side . ]
there was substantial confusion about the source for the ulirg - strength iras mir and fir emission in the literature . from the uncorrected [ oii ]
line flux a star formation rate ( sfr ) of only 1 m@xmath4/yr can be inferred @xcite .
@xcite still assign the ulirg emission to the companion galaxy due to its balmer decrement which yields non - negligible dust extinction , while @xcite note that the corrected sfr would still be below 10 m@xmath4/yr .
this number is in strong disagreement with a sfr up to @xmath2800 m@xmath4/yr inferred from total ir - luminosity or 370 m@xmath4/yr from co @xcite .
the new nicmos images show that the stars in the companion galaxy are not distributed in a ring , but smoothly ( fig .
[ fig : companion ] ) and that an optically thick dust creates the ring - like structure in the optical acs images ( fig .
[ fig : allwave]a+b ) .
this means that at optical wavelengths only information from the less extincted outer regions of the galaxy as well as the surface of the strongly extincted central regions is seen .
uv - based sfrs must therefore dramatically underestimate the true sfrs when corrected with dust extinction estimated from ( also optical wavelength ) balmer - decrements .
the actual scale of the uncertainty in @xmath38 , the optical extinction correction , can be estimated by comparing @xmath38 estimates from balmer lines and paschen / bracket lines in other ulirgs .
@xcite studied five ulirgs for which they estimated @xmath38 both from h@xmath39/h@xmath40 as well as from pa@xmath41/br@xmath42 .
nir - derived values for a@xmath43 were in every case significantly larger , ranging from factors of @xmath21.16 to @xmath210 ( mean 4.0 ) times higher .
as this factor does not scale in any way with the optical @xmath38 estimate , but only with the nir estimate , we can not determine a correction for he04502958 .
when starting out with the redshifted [ oii]-line at @xmath443727 and @xmath45 like @xcite , and the correction factors from @xcite , a huge range of possible star formation rates arises .
an average of the two _ lowest _ correction values from @xcite of 1.16 and 1.65 means @xmath46 , @xmath47 or corrected sfrs of 21 m@xmath4/yr .
using their mean correction factor of @xmath24 would lead to @xmath48 or @xmath495000 m@xmath4/yr . so already a number below the mean correction ( @xmath50 ) would make these numbers consistent with fir - emission based sfr estimates .
this directly shows that optical / uv line - emission based sfrs as used by @xcite can not at all be used to constrain the true sfr of ulirgs and does not provide an argument against strong star formation in the companion .
@xcite approximate the iras ir sed with a 2-component black - body model and find a cool component @xmath51k , dust mass @xmath52 m@xmath4 and @xmath53 l@xmath4 , and a warm component with @xmath54k , @xmath55 m@xmath4 and @xmath56 l@xmath4 .
we can now for the first time spatially localize the warm component from the detection of the single 11.3@xmath1 m point source with visir to be coincident with the position of the qso nucleus .
since the measured flux density is consistent with a warm component having the previously known 12@xmath1 m iras flux density , we conclude that the qso nucleus itself already is a ulirg - level emitter , but with a warmer component compared to star formation . for localizing star formation in the system , there are two recent new datasets available , radio data from @xcite and the co maps by @xcite . while the radio maps do not set strong constraints when trying to exploit the radio fir relation to assign a location for the fir emission , the co data are more powerful : at least the bulk , possibly all of molecular gas and thus star formation activity is located in the companion galaxy .
we can add two further constraints from our nicmos and visir images .
both the mid infrared sed of the system ( fig .
[ fig : iras_sed ] ) as well as an extrapolation from the @xmath20-band are consistent with an arp220-like star formation , while ruling out milder , m82-like conditions . in the latter case
the companion would have to be visible in our observed 11.3@xmath1 m image , but it is absent ( fig . [
fig : allwave ] ) .
together with the dense and clumpy dust geometry of the companion when comparing optical and nir morphology , it becomes clear that the companion is responsible for most , if not all , of the 370 m@xmath4/yr star formation .
if we follow the 5:1 co detection significance for the companion given by @xcite , this means that as a minimum 5/6=83% of co are located in the companion and thus also @xmath57% of the star formation and fir emission .
this number converts to an integrated ir luminosity of @xmath58 l@xmath4 , so the companion also qualifies as a ulirg . while the presence of very strong star formation in the companion is clear now , its trigger is a priori not so clear .
the most probably solution is merger induced sf , so the system would be a classical ulirg just with a non - standard geometry
but there is room for a radio jet induced effect as well .
one of the lobes of the jets from the qso is located directly at the companion position . if and how much this contributes to star formation in the companion still needs to be quantified . with
the companion identified as the main star - former , we get limits from the co that less than 1/6 of the total cool dust is located within the putative host galaxy . thus 1/6 of the fir - inferred sfr by @xcite of @xmath59 m@xmath4/yr correspond to an upper limit of 62 m@xmath4/yr .
this leaves room for a non - negligible amount of sf in the host galaxy , but is also an upper limit qsos ( also 47 k , * ? ? ?
* ) , and can be composed of intrisically cooler dust ( 2030 k ) plus agn heating .
this temperature could thus be a hint that indeed a part of this cool dust component is located in the qso host galaxy and not in the companion . ] .
if we assume the host galaxy to have a mix of old and young stellar population as we find for other qso host galaxies at these redshifts
@xcite , we can convert this to an expected @xmath20-band flux . if the host galaxy had the same population mix as @xcite modelled for the companion galaxy
95.5% of a 10 gyr old population with 5 gyr e - folding sfr timescale plus 4.5% of a 128 myr young population , this sfr upper limit would translate to an expected nir magnitude of 1.75 mag fainter than the companion or @xmath60 .
the combined color and @xmath21-correction term is @xmath61 , and changes by only about @xmath620.3mag for a pure old ( 10 gyr ) or young ( 100 myr ) population .
so they are rather insensitive to the exact choice of stellar population . however , this limit will get brighter if the host galaxy contained less dust by about 0.3 mag per magnitude decrease in @xmath38 . with that in mind ,
this limit is not more stringent than the limit from nicmos itself : no significant main host galaxy body is found after psf removal ( section [ results : host ] ) and so an upper limit from the nir decomposition of @xmath29 applies for a host galaxy co - centered with the quasar nucleus .
we therefore conclude that the current upper limit from nicmos lies at around @xmath63 .
this is consistent with the co / fir limits . how do these numbers relate to the current upper limit for a co - centered host galaxy from the optical hst data ?
we convert our @xmath20-band limit to absolute @xmath15-band magnitudes with again the assumption of the host galaxy having the same stellar population mix as the companion . in the conversion to @xmath64
we assume two different values for dust extinction , ( a ) @xmath65 , motivated by the nearly dust - free line of sight to the qso nucleus , and ( b ) a moderate @xmath66 ( corresponding to @xmath67 ) .
this yields host - galaxy upper limits of @xmath68 and @xmath69 , for the cases ( a ) and ( b ) respectively .
if we convert the @xcite upper limits to our @xmath70 cosmology and assume the same stellar population and dust properties , we receive @xmath71 and @xmath72 , respectively .
we note here that this corresponds to a detection limit of only 1.5% of the total quasar flux in the optical .
this factor of two is owed to the better determined psf in the acs images .
this allows @xcite to set somewhat stricter upper limits for a nucleus co - centered host galaxy component , particularly if a low dust extinction is present .
concerning lower limits to the host galaxy , the ne - extension ( figure [ fig : ne - extension ] ) is a structure of real emission that can be traced towards the qso from @xmath215 to a radius of 06 , where the region of substantial psf residuals begins .
we can not say for sure whether it continues further inward from this position .
signs of this structure are visible in the acs @xmath15-band ( see figure [ fig : allwave ] , left column ) but it is not clear whether the more compact region only @xmath202 n e of the nucleus in acs image is real or an artefact of the deconvolution process .
we measured the @xmath73-color to be 2.8 outside this region , which is consistent with a stellar population of intermediate age . in the dust - free case
this color corresponds to a @xmath22.1 gyr old single stellar population ( * ? ? ?
* solar metallicity ) , for @xmath74 to an age of 800 myr .
this is consistent with stellar material from a host galaxy , e.g. tidally ejected disk stars .
we conclude that with its spatial detachment from the companion galaxy this ne - extension is likely a part of the host galaxy , possibly as a tidal extension , but its vicinity to the qso makes other interpretations less likely . with this interpretation
, we receive an @xmath75 _ lower _
limit for the host , corresponding to @xmath76 ( @xmath65 ) or @xmath77 ( @xmath66 ) .
if we include this off - center emission to the upper limit of a co - centered host galaxy , we obtain a total host galaxy upper limit of @xmath78 and @xmath79 .
we thus bracket the host galaxy luminosity in the @xmath15-band by 0.8 and 1.3 mag or factors of @xmath80 and @xmath81 , respectively .
formally , the co detection significance and nicmos give the same limit on a star formation rate of up to @xmath82 m@xmath4/yr .
if we take into account the stricter acs @xmath15-band limits of @xmath71 and @xmath72 , depending on dust cases ( a ) and ( b ) , these are fainter by 1.3 and 0.6mag than the co predicted magnitures .
inversely , these reduce the upper limits on star formation to 18 and 35 m@xmath4/yr , respectively . beyond @xmath83mag
the co and nicmos limits again become the most stringent .
this means that we can not rule out dust obscuration in the host galaxy . at the same time
the dust - free line of sight to the quasar nucleus is a strong argument against large amounts of dust , unless a very special geometrical configuration is invoked , while the warm ulirg emission from the qso points to dust in the very central few 100 pc
. only better co limits or a detection of the host galaxy in the nir will be able to finally resolve this matter .
black hole mass estimates for he04502958 vary significantly through the literature .
the original 8@xmath1210@xmath84 @xmath6 @xcite were revised later to a substantially lower value of 4@xmath1210@xmath85 @xcite . both values are virial estimates based on h@xmath86 width , but
while narrow and broad components were separately measured in the former study , the fwhm of the whole line was used in the latter .
this revised value is consistent with the independent virial estimate of 69@xmath1210@xmath85 by @xcite , and even with an estimate from x - ray variability , @xmath87@xmath1210@xmath85 @xcite .
since the virial estimates agree now , we will adopt the range 49@xmath1210@xmath85 @xmath6 for the black hole mass .
@xcite noted the rather narrow broad emission lines of he04502958 and suggested that it should actually be viewed not as a standard qso but as a higher-@xmath88 analog of local nlsy1s .
if we compare he04502958 with estimates from the literature @xcite , we find that he04502958 is consistent with the high black hole mass end of the known nlsy1 distribution and does not need to constitute a new `` higher-@xmath88 nlsy1 analog '' class of its own .
but is it consistent regarding other properties as well ?
morphologically , nlsy1 are mostly spirals , often barred , mostly not strongly disturbed @xcite .
since galaxies have increasing bulge mass with increasing black hole mass it is not clear which structural properties to expect and if a merging system like this is consistent with the properties of the local , lower mass nlsy1 population .
there is even a debate on how different nlsy1 actually are from normal seyferts .
recent studies show smaller bh mass differences between normal broad - line sy1 and nlsy1 when using line dispersions instead of fwhm @xcite , although a difference might remain .
if galaxies with potentially core outflow - affected lines are considered separately , nlsy1 share the same @xmath89-relation with blsy1 , but their accretion rates are confirmed as lying often close to the eddington limit @xcite . if we compute the he04502958 accretion rate as derived from the @xmath15-band absolute magnitude of the quasar nucleus ( @xmath90 , recomputed from the hst / acs data with updated agn color and @xmath21-correction ) and a bolometric correction of @xmath91 @xcite in relation to its eddington accretion rate , we obtain from @xmath92 @xmath6 a super - eddington accretion rate of @xmath5 .
this is consistent with high eddington ratios observed for nlsy1 @xcite .
@xmath10@xmath93-relation for inactive galaxies in the local universe as presented by @xcite , with data from @xcite , @xcite and @xcite _ ( black lozenges and lines)_. overplotted are the upper limits for the host galaxy of he04502958 for the dust - free case by @xcite from @xmath15-band imaging _ ( small blue arrow ) _ and with an @xmath66 added _ ( small red arrow ) _ , with their original black hole estimate , converted to our cosmology .
the _ blue and red rectangles _ show the range for black hole mass estimates and our new lower limits for the ( total ) galaxy luminosity from nicmos and new upper limits based on the ( still better constrained ) optical hst data . note : here we combined the off - center flux lower limit ( nicmos @xmath20-band ) with the upper limit for a co - centered host galaxy ( acs @xmath15-band ) for a total upper limit .
the arrows to the bottom right show the conversion of our @xmath94 limits to @xmath93 limits for bulge - to - disk ratios of 1:2 and 1:4 .
both the dust - free as well as the @xmath66 dust case show a galaxy that is absolutely consistent with the black hole mass , even if the bulge - to - disk ratio is accounted for . ] with the new data and an explicit assumption / interpretation that the ne - extension is indeed associated with the host galaxy , we can for the first time present a black hole mass for he04502958 and bracketing limits for its host galaxy luminosity .
we can thus place he04502958 on the @xmath10@xmath93-relation of active and inactive galaxies , with more than just an upper limit for galaxy luminosity . in figure
[ fig : m_m ] we show data from @xcite and others , as collected by @xcite .
we overplotted the limits on he04502958 for the two assumptions of dust attenuation strength ( sec .
[ sec : hostgalaxydetection ] ) .
this shows that even when applying a sensible conversion factor of 1 to 1/4 ( up to 1.5 mag ) to convert from total to bulge luminosity , the host of he04502958 will be a perfectly normal galaxy in this parameter space , with a luminosity around the knee of the galaxy luminosity function , @xmath95 .
contrary to the claim by @xcite it does not deviate substantially from the local @xmath10@xmath93-relation for normal inactive local massive galaxies , mainly due to the revised mass estimate for the black hole .
however , this also means that he04502958 does not show a @xmath10/@xmath93 different from local broad - line agn , consistent with being a nlsy1-analog if the @xcite result is taken as a base . with the normal @xmath10/@xmath93-ratio and the fact that we can now rule out huge amounts of obscuring dust around the qso nucleus ,
the most likely explanation for the evasive host galaxy is indeed a high @xmath96 accretion rate system a nlsy1 at the high mass end of the normal nlsy1 population . with the current evidence occam
s razor favors this explanation over more exotic scenarios as the ejection of the qso s black hole in a 3-body interaction or a gravitational recoil event involving the companion galaxy ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
however , these scenarios are formally not ruled out even if the upper limit can be pushed down by another @xmath25 magnitudes .
all evidence combined is consistent with a system of a qso with ulirg - size ir emission , residing in an @xmath97 host galaxy that is in the process of colliding with a substantially more luminous and possibly more massive companion ulir - galaxy .
much deeper high - resolution nir imaging with a well controlled psf are the best way to finally find and trace the here predicted host galaxy ( bulge ) component of he04502958 co - centered with the qso nucleus and to estimate its luminosity and mass directly . given the black hole mass and eddington ratio the accretion rate of the bh is 1.4 m@xmath4/yr . at the same time
@xcite derive a star formation rate from co of 370 m@xmath4/yr , predominantly in the companion galaxy . applying
a correction factor of 0.5 for mass returned to the interstellar matter by stellar winds , the stellar mass growth of the whole he04502958 system from star formation is 185 m@xmath4/yr . the ratio of black hole accretion and stellar mass growth is then 12/185=6.5% , which is substantially higher than the @xmath10/@xmath11 relation for local galaxies of 0.14% @xcite .
we can conclude the following : if activity timescales are identical for star formation and bh accretion , this system grows in black hole mass much more rapidly than the bulge is required to grow to keep the system on the @xmath10/@xmath11 relation .
this is not possible , since the star formation is taking place in the companion and not the host galaxy .
so in any case a potential maintainance of the relation for this system , if actually true , needs to be seen as an integral over more than several 10@xmath84 yrs . on the other hand , a gas consumption timescale of 9.5@xmath1210@xmath85 yrs
if we divide the h@xmath98 masses and sf rates derived by @xcite and account for 50% mass recycling is possibly longer than the luminous quasar accretion phase .
this would add to the requirement , that processes like the tidal forces of the galaxy interaction redistribute mass , adding stars to the bulge of the host galaxy .
these were to the larger extent already preexisting in the host galaxies disk or the companion before the interaction and not created only now .
the `` coevolution '' of the host galaxy and its black hole in he04502958 is clearly a two - part process : the build - up of stellar mass and the build - up of black hole and bulge mass .
the former will take place on timescales of @xmath491 gyr through star formation , the latter two can `` coevolve '' if seen as an average over timescales of longer than the bh accretion lifetime , and a few dynamical timescales for redistribution of stellar orbits of , say , @xmath9500 myrs .
he04502958 is an unusual object .
agn in ulirgs are common , but agn right next to ulirgs are not , particularly not luminous qsos with inconspicuous host galaxies next to extreme starformers .
so is he04502958 one of a kind or was it just the scarceness of ir imaging with 1 resolution and high - resolution co maps that prevents us from finding similar objects en masse ? in the higher redshift universe
there was a recent report of a very similar system @xcite .
lh850.02 at @xmath99 is the brightest submm galaxy in the lockman hole . using the submillimeter array ,
the authors find two components of which one is a ulirg with intense star formation , while the other component likely harbors an agn . at @xmath100
however , objects like this might be quite common , since merging rates and gas reservoirs were much larger than today .
if there existed a substantial number of similar systems at low redshifts , this would allow to study mechanisms of the high - redshift universe at much lower distances .
we try to estimate the frequency of such systems in the local universe using the three morphologically best studied samples of quasars at @xmath101 .
we deliberately use optically selected quasars only , as they have no bias with respect to frequency of merger signatures or extreme sfrs as ir - selected samples have by construction . in this way statements about the general population
are possible .
@xcite investigated a volume - limited and complete sample of 19 luminous qsos out to @xmath102 . while at least five of these qsos are seen in intermediate and late stages of major mergers , only one , he12540934 ,
is a likely ulirgm or assume the flux to be zero . ] , as determined from its iras fluxes .
it is also among the most distorted systems , with a companion at @xmath21 distance from the qso nucleus .
the companion is more luminous than the host galaxy , and shows a substantial tidal tail .
it looks remarkably similar to he04502958 .
the two other samples are not volume - limited samples , so the selection function is unclear except that these quasars stem from either optical or radio surveys , but not the ir .
@xcite studied the morphologies of two intermediate- and high - luminosity samples of ten radio - quiet and seven radio - loud quasars at @xmath103 , using hst - imaging data .
only one of their 17 quasars shows a distorted geometry similar to he04502958(1237040 at @xmath104 ) but there exists no information about the total ir emission or star - formation rates .
the iras flux limits of 200mjy is equivalent to upper limits of @xmath105 @xmath106 at @xmath107 .
ulirg - strength emission for 1237040 could have gone unnoticed by iras . a recent study by @xcite determined the morphologies of 45 hst - archived quasars at @xmath108 .
it has one object in common with @xcite and three objects with @xcite . of their sample , three other objects
( he03545500 , pg1613 + 658 , pks234901 ) are clearly merging with a nearby companion , and are likely ulirgs as judged from their iras fluxes .
however , only in the case of he03545500 the quasar and companion are still well separated and their envelopes have not yet merged into a common halo .
the two other cases are in a very late merger state and star - formation will likely occur all over the system .
this adds up to only @xmath1093/77 qsos to possibly be he04502958-like in the three samples combined . at @xmath110%
such systems are indeed rare in the local universe .
these three quasars however should be investigated in more detail .
it needs to be tested how strong their star - formation actually is , where in the system it is localized , and if the separated companion is in any way connected to the agn - fuelling .
if a similar situation as for he04502958is found , the result can set strong constraints on the ulirg
agn evolutionary scenario @xcite and the creation mechanisms of agn at high redshifts .
it can contribute to answering the question whether sf - ulirg activity in agn systems is an indicator of a specific mechanism of agn fuelling . or , if these are just the most gas - rich merger - triggered agn systems at the top end of sfrs , with a continuous sequence towards less gas - rich merger - triggered agn systems .
the merging
agn fuelling mechanism could be identical from ulirgs down to the seyfert regime , where at some point secular mechanisms become more dominant .
lower sfr systems could just be the consequence of lower gas mass , but this might only mildly impact on the much smaller agn fuelling rate .
with new nir and mir images to spatially resolve the he04502958 system , and in the light of previously existing data , we find : 1 .
the companion galaxy is covered in optically thick and unevenly distributed dust .
this makes it appear as a collisional ring galaxy in the optical , but intrinsically it is smooth and has smooth nir emission increasing towards a pronounced center .
the star formation in the companion is similar to the strong starburst arp220 , while softer m82-like star formation is ruled out .
this can reconcile the sfr estimates from the optical and fir .
the companion is a star - formation powered ulirg .
2 . our mir image confirms a single warm dust point source at the location of the qso nucleus .
this supports a two component dust sed with the warm component fully associated with the qso nucleus , which is an agn - powered ulirg .
3 . a dust - free line of sight to the quasar nucleus is evidence that the host galaxy is not obscured by large amounts of dust .
however , the ulirg - strength warm ir emission by the nucleus and the upper limit on star formation in the host galaxy of substantial 60 m@xmath4/yr leave room for dust .
4 . with @xmath111
the current nicmos images do not set stronger upper limits on the host galaxy of he04502958 .
the @xmath15-band , @xmath20-band , and co - constraints give @xmath112 to @xmath113 depending on the assumed dust masses .
flux in the ne - extension of @xmath31 is likely associated with the qso s host galaxy .
it corresponds to a first lower limit of @xmath76 for the host galaxy . with a black hole of @xmath114 m@xmath4 , an accreting rate of 12 m@xmath4/yr equal to super - eddington accretion , @xmath5 ,
the host galaxy is consistent with the @xmath10@xmath11-relation for normal galaxies .
it is also consistent with he04502958 being a nlsy1 at the high end of the known black hole mass distribution .
the reason for the high accretion rate is unclear but could be connected to he04502958 being in an early stage of merging with its gas - rich companion .
a more exotic explanation for the system is currently not required by any data , but can in the end only be ruled out with much deeper , high - resolution nir images to find the main body and bulge of the host galaxy . 6 .
if host galaxy and black hole in he04502958 are co - evolving according to the local @xmath10@xmath11 relation , it has to occur over longer timescales ( @xmath109500 myr ) and/or the mass growth for the bulge is predominantly not caused by the current star formation in the system , but by redistribution of preexisting stars .
. a constellation as in the he04502958 system with separate locations of qso nucleus and strongly star forming ulirg companion might be common at @xmath100 where gas masses and merger rates were higher , but at a fraction of @xmath1094% it is extremely rare in the local universe . the authors would like to thank e. f. bell , a. martnez sansigre , h.dannerbauer , e. schinnerer , k. meisenheimer , f. courbin , p. magain and h .-
klckner for very fruitful discussions and helpful pointers . based on observations made with eso telescopes at the paranal observatory under programme
i d 276.b-5011 .
also based on observations made with the nasa / esa hubble space telescope , obtained at the space telescope science institute , which is operated by the association of universities for research in astronomy , inc .
, under nasa contract nas 5 - 26555 .
these observations are associated with program # 10797 .
this research has made use of the nasa / ipac extragalactic database ( ned ) .
kj acknowledges support through the emmy noether programme of the german science foundation ( dfg ) with grant number ja 1114/3 - 1 .
ab is funded by the deutsches zentrum fr luft- und raumfahrt ( dlr ) under grant 50 or 0404 .
vc , research fellow , thanks belgian funds for scientific research .
this work was also supported by prodex experiment arrangement 90312 ( esa and pps science policy , belgium ) .
jahnke , k. , snchez , s. f. , wisotzki , l. , barden , m. , beckwith , s. v. w. , bell , e. f. , borch , a. , caldwell , j. a. r. , huler , b. , heymans , c. , jogee , s. , mcintosh , d. h. , meisenheimer , k. , peng , c. y. , rix , h .- w . , somerville , r. s. , & wolf , c. 2004 , apj , 614 , 568 , j. d. , dunlop , j. s. , peck , a. b. , ivison , r. j. , biggs , a. d. , chapin , e. l. , clements , d. l. , dye , s. , greve , t. r. , hughes , d. h. , iono , d. , smail , i. , krips , m. , petitpas , g. r. , wilner , d. , schael , a. m. , & wilson , c. d. 2008 , mnras , 387 , 707 | the luminous @xmath0 quasar he04502958 is interacting with a companion galaxy at 6.5 kpc distance and the whole system radiates in the infrared at the level of an ultraluminous infrared galaxy ( ulirg ) .
a so far undetected host galaxy triggered the hypothesis of a mostly `` naked '' black hole ( bh ) ejected from the companion by three - body interaction .
we present new hst / nicmos 1.6@xmath1 m imaging data at 01 resolution and vlt / visir 11.3@xmath1 m images at 035 resolution that are for the first time resolving the system in the near- and mid - infrared .
we combine these data with existing optical hst and co maps .
( i ) at 1.6@xmath1
m we find an extension n - e of the quasar nucleus that is likely a part of the host galaxy , though not its main body . if true , a combination with upper limits on a main body co - centered with the quasar brackets the host galaxy luminosity to within a factor of @xmath24 and places he04502958 directly onto the @xmath3-relation for nearby galaxies .
( ii ) a dust - free line of sight to the quasar suggests a low dust obscuration of the host galaxy , but the formal upper limit for star formation lies at 60 m@xmath4/yr .
he04502958 is consistent with lying at the high - luminosity end of narrow - line seyfert 1 galaxies , and more exotic explanations like a `` naked quasar '' are unlikely .
( iii ) all 11.3@xmath1 m radiation in the system is emitted by the quasar nucleus .
it has warm ulirg - strength ir emission powered by black hole accretion and is radiating at super - eddington rate , @xmath5 , or 12 @xmath6/year .
( iv ) the companion galaxy is covered in optically thick dust and is not a collisional ring galaxy .
it emits in the far infrared at ulirg strength , powered by arp220-like star formation ( strong starburst - like ) .
an m82-like sed is ruled out .
( v ) with its black hole accretion rate he04502958 produces not enough new stars to maintain its position on the @xmath3-relation , and star formation and black hole accretion are spatially disjoint .
this relation can either only be maintained averaging over a longer timescale ( @xmath7500 myr ) and/or the bulge has to grow by redistribution of preexisting stars .
( vi ) systems similar to he04502958 with spatially disjoint ulirg - strength star formation and quasar activity might be common at high redshifts but at @xmath8 we only find @xmath94% ( 3/77 ) candidates for a similar configuration . |
in classical chemical kinetics , the evolution of concentrations of various components in a reaction is described by ordinary differential equations .
such a description turns out to be unsatisfactory in a number of applications , especially in biology ( see @xcite ) .
there are several ways to construct a more adequate mathematical model .
if the reaction is fast enough , one should take into account that the concentration is not constant in space in the volume where the reaction takes place .
then , the change of concentration due to spatial transport should be taken into account and the system of odes should be replaced by a system of pdes of reaction - diffusion type . in some cases , one should also take into account random changes in time of the rates of reaction .
then , the ode is replaced by a stochastic differential equation .
if the rates change randomly not just in time but also in space , then evolution of concentrations can be described by a system of spdes . on the other hand ,
the rates of chemical reactions in the system and the diffusion coefficients may , and as a rule do , have different orders .
some of them are much smaller than others and this allows one to apply various versions of the averaging principle and other asymptotic methods , thereby eventually obtaining a relatively simple description of the system . in this paper , we study the case where the diffusion rate is much larger than the rate of reaction and we show that in this case , it is possible to replace spdes of reaction - diffusion type by suitable sdes .
such an approximation is valid , in particular , if the reaction occurs only on the boundary of the domain ( this means that the nonlinearity is included in the boundary conditions ) .
this replacement is a result of _ averaging _ in the fast spatial transport .
we would like to stress that our approach allows us also to calculate the main terms of deviations of the solution of the original problem from the simplified model .
notice , moreover , that the case where the diffusion coefficients and some of the reaction rates are large compared with other rates can be considered in a similar way .
more precisely , we are dealing with the following class of equations : @xmath0 for some @xmath1 .
these are reaction - diffusion equations perturbed by a noise of multiplicative type , where the diffusion term @xmath2 is multiplied by a large parameter @xmath3 and a noisy perturbation is also acting on the boundary of the domain @xmath4 . here , @xmath4 is a bounded open subset of @xmath5 , with @xmath6 , having a regular boundary ( for more details , see section [ sec2 ] ) and , in the case @xmath7 , we take @xmath8 $ ] .
@xmath2 is a uniformly elliptic second order operator and @xmath9 is the corresponding conormal derivative .
this is why the same constant @xmath3 , which is in front of the operator @xmath2 , is also present in front of the conormal derivative @xmath10 . in what follows
, we shall denote by @xmath11 the realization in @xmath12 of the differential operator @xmath2 , endowed with the conormal boundary condition .
the coefficients @xmath13 are assumed to be measurable and satisfy a lipschitz condition with respect to the third variable , uniformly with respect to the first two variables , and the mapping @xmath14 is bounded with respect to the space variable .
the noisy perturbations are given by two independent cylindrical wiener processes , @xmath15 and @xmath16 , defined on the same stochastic basis @xmath17 , which take values on @xmath12 and @xmath18 , respectively , and have covariance operators @xmath19 and @xmath20 , respectively . , we denote by @xmath21 the banach space of bounded linear operators on @xmath22 and by @xmath23 the subspace of nonnegative and symmetric operators . ] in space dimension @xmath7 , we can take @xmath24 equal to the identity operator so that we can deal with space time white noise .
moreover , as @xmath25 , in space dimension @xmath7 , we do not assume any condition on @xmath26 .
stochastic partial differential equations with a noisy term also acting on the boundary have been studied by several authors ; see , for example , da prato and zabczyk @xcite , freidlin and wentzell @xcite and sowers @xcite
. the last two mentioned papers also deal with some limiting results with respect to small parameters appearing in front of the noise .
however , the limiting results which we are studying in the present paper seem to be completely new and we are not aware of any previous results dealing with the same sort of multiscaling problem , even in the simpler case of homogeneous boundary conditions ( i.e. , @xmath27 ) .
as mentioned above , our interest is in studying the limiting behavior of the solution @xmath28 of problem ( [ eq1 ] ) as the parameter @xmath29 goes to zero , under the assumption that the diffusion @xmath30 associated with the operator @xmath2 , endowed with the conormal boundary condition [ this corresponds to a diffusion @xmath30 on some probability space @xmath31 which reflect on the boundary of @xmath4 ] , admits a unique invariant measure @xmath33 and a spectral gap occurs .
that is , for any @xmath34 , @xmath35 for some constant @xmath36 . this can be expressed in terms of the semigroup @xmath37 associated with the diffusion @xmath30 , by saying that @xmath38 moreover , as shown in remark [ rem21 ] , the space @xmath12 is continuously embedded into @xmath39 .
our aim is to prove that equation ( [ eq1 ] ) can be replaced by a suitable one - dimensional stochastic differential equation , whose coefficients are obtained by averaging the coefficients and the noises in ( [ eq1 ] ) with respect to the invariant measure @xmath33 .
more precisely , for any @xmath34 , we define @xmath40 and for any @xmath34 , @xmath41 and @xmath42 , we define @xmath43 and @xmath44(x ) \mu(dx),\qquad t\geq0,\ ] ] where @xmath45 is the neumann map associated with @xmath2 and @xmath46 is a suitable constant ( see section [ sec2 ] , @xcite and @xcite for definitions ) .
we prove that for any @xmath47 , the mappings @xmath48 and @xmath49 are both well defined and lipschitz continuous , and @xmath50 , so that the stochastic ordinary differential equation @xmath51 admits , for any @xmath52 and @xmath53 , a unique strong solution @xmath54))$ ] which is adapted to the filtration of the noises @xmath15 and @xmath16 .
notice that ( [ eq1bis ] ) is a one - dimensional stochastic equation , in the sense that the space variables have disappeared . in section [ sec4 ]
, we show that it can be rewritten as @xmath55 where @xmath56 is a standard brownian motion and the diffusion coefficient @xmath57 is explicitly given in terms of @xmath24 , @xmath58 , @xmath26 and @xmath59 .
when we say that equation ( [ eq1 ] ) can be replaced by ( [ eq1bis ] ) , we mean that the solution @xmath28 of ( [ eq1 ] ) can be approximated by the solution @xmath60 of ( [ eq1bis ] ) in the following sense : @xmath61}\biggl|\int_d @xmath62 and @xmath63 . in order to prove ( [ finale ] )
, we first have to prove that for any fixed @xmath64 , equation ( [ eq1 ] ) admits a unique adapted mild solution in @xmath65;l^2(d)))$ ] , that is , there exists a unique adapted process @xmath28 such that @xmath66 where @xmath67 is the boundary term ( the _ stochastic boundary convolution _ ) @xmath68,\qquad t\geq0\ ] ] ( here , and in what follows , @xmath69 and @xmath58 denote the composition / multiplication operators associated with @xmath70 and @xmath71 , resp . ) .
in particular , we have to show that the above term is well defined in @xmath65;l^2(d)))$ ] . concerning the notion of mild solutions and existence and uniqueness results for spdes like ( [ eq1 ] ) , with fixed @xmath64
, we refer to da prato and zabczyk @xcite .
however , we would like to stress that in the present paper , we are not imposing the hilbert schmidt condition on the covariance operators @xmath24 and @xmath26 , and this makes the treatment of the stochastic convolution and of the stochastic boundary convolution more complicated , in view also of the a priori estimates with respect to @xmath64 . actually , once we have a unique adapted mild solution @xmath28 for ( [ eq1 ] ) , we prove an a priori estimate of the following type : @xmath72}{\mathbb e}|u_\varepsilon ( t)|_{c([0,t];l^2(d))}^p\leq c_{t , p}\bigl(1+|u_0|_{l^2(d)}^p\bigr).\ ] ] due to ( [ dubbio ] ) , this allows us to proceed to the proof of ( [ finale ] ) . after we have proven ( [ finale ] ) , in the final section , we study the fluctuations of @xmath28 from @xmath60 .
namely , we introduce the random field @xmath73 and show that , under the assumption that the noisy perturbation in ( [ eq1 ] ) is of additive type ( i.e. , the diffusion coefficient @xmath71 is independent of @xmath74 ) , for any @xmath75 , @xmath76 where @xmath77 is the gaussian random field taking values in @xmath39 for any @xmath75 , defined by @xmath78(x)-\langle\hat{\sigma}(t ) , dw^b(s)\rangle_{l^2(\partial d)}\bigr).\end{aligned}\ ] ] the random field @xmath77 is well defined in @xmath39 because of the spectral gap inequality ( [ dubbio ] ) and , in the case where the coefficients @xmath71 and @xmath79 do not depend on @xmath80 , @xmath77 also does not depend on @xmath80 so that the weak limit of @xmath81 as @xmath82 depends only on the space variable @xmath83 and is constant in time for any @xmath75 .
let @xmath4 be a bounded domain in @xmath5 , with , satisfying the extension and exterior cone properties , and let @xmath84 be the outward normal at @xmath85 .
we assume that @xmath85 is a @xmath86 manifold and @xmath4 is locally only on one side of @xmath87 . in the case
@xmath7 , @xmath4 is a bounded interval @xmath88 .
we define @xmath89 and @xmath90 and , for any @xmath91 , we define @xmath92 and @xmath93 ( in particular , @xmath94 and @xmath95 ) .
we assume that @xmath2 is a second order differential operator , @xmath96 satisfying the uniform ellipticity condition @xmath97 for some @xmath98 . the coefficients @xmath99 and @xmath100 are assumed to be smooth [ for simplicity , we take them to be in @xmath101 . in what follows , we shall denote by @xmath11 the realization in @xmath102 of the operator @xmath2 , endowed with the boundary condition @xmath103 namely , @xmath104 as is well known , the operator @xmath11 generates an analytic semigroup @xmath105 in @xmath102 which is also strongly continuous .
moreover , @xmath106 and @xmath107 ( for proofs , see @xcite and @xcite , resp . ) .
if , for any @xmath108 , we denote by @xmath109 the realization in @xmath110 of the operator @xmath2 , endowed with the boundary condition ( [ boundary ] ) , it can be proven that @xmath109 generates a strongly continuous analytic semigroup @xmath111 in @xmath110 .
notice that all of these semigroups are consistent , so , in what follows , we shall denote them all by @xmath37 . as proved in , for example , @xcite , theorem 2.4.4 , since @xmath112 is uniformly elliptic and the domain @xmath4 has the extension property , the semigroup @xmath37 admits an integral kernel @xmath113 .
due to the boundary condition , the kernel satisfies @xmath114 for some constant @xmath115 , almost everywhere in @xmath116 . as a consequence of our assumptions on @xmath2 and @xmath4 , it is possible to prove that there exists some @xmath117 such that for any @xmath118 and @xmath119 , the elliptic boundary value problem @xmath120 admits a unique weak solution @xmath121 , which we will denote by @xmath122 . the application @xmath123 is known as the _
neumann map _ associated with the operator @xmath2 .
it is well known that @xmath124 maps @xmath125 into @xmath102 as a bounded linear mapping . moreover , according to elliptic theory for domains with smooth boundaries ( for a proof , see @xcite , theorem 7.4 of volume i ) , we have @xmath126 in what follows , we shall assume that @xmath37 has the following long - time behavior .
[ h1 ] the semigroup @xmath37 , @xmath47 , admits a unique invariant measure @xmath33 and there exists some @xmath36 such that , for any @xmath127 , @xmath128 in what follows , we shall set @xmath129 and @xmath130 [ rem21 ] 1 .
if @xmath2 is a divergence - type operator , that is , @xmath131 for any @xmath132 , then the operator @xmath11 is self - adjoint in @xmath102 .
this implies that it is possible to fix a complete orthonormal system @xmath133 in @xmath102 and an increasing sequence of nonnegative real numbers @xmath134 such that @xmath135 let @xmath136 be the constant eigenfunction corresponding to the @xmath137 eigenvalue and let @xmath138 be the first positive eigenvalue .
it is immediate to check that @xmath139 and , in particular , that @xmath140 , with equivalence of norms .
moreover , as for any @xmath141 , we have @xmath142 and @xmath143 for any @xmath144 , it is immediate to check that @xmath145 so the constant @xmath146 in ( [ ergo ] ) coincides with @xmath138 .
2 . if @xmath11 is self - adjoint , as above , for any @xmath147 and @xmath148 it holds that @xmath149 actually , for any @xmath119 , we have @xmath150 now , if we assume that @xmath151 , according to ( [ sobolev ] ) , we have that @xmath152 and then , due to the gauss green formula and to ( [ neumann ] ) , we obtain @xmath153 this implies that @xmath154 so that @xmath155 as @xmath156 is dense in @xmath125 , we can conclude that ( [ adjoint ] ) holds .
3 . as @xmath157 and @xmath158
, we have @xmath159 due to the invariance of @xmath33 , this implies that for any @xmath160 , @xmath161 so @xmath37 acts on @xmath162 as a contraction , that is , @xmath163 4 .
we have that @xmath102 is continuously embedded into @xmath162 .
actually , due to the invariance of @xmath33 and to the kernel representation of @xmath37 , for any @xmath164 , we have @xmath165 then , thanks to ( [ kernel ] ) , we have @xmath166 5 . as a matter of fact , there exists a nonnegative function @xmath167 such that @xmath168 actually , let @xmath169 , with @xmath170 fulfilling the boundary condition ( [ boundary ] ) .
integrating by parts , we obtain @xmath171 where @xmath172 hence , the operator @xmath173 , endowed with the boundary condition @xmath174 is the formal adjoint of the operator @xmath2 , endowed with the boundary condition ( [ boundary ] ) . + now , the function @xmath175 is a nonzero solution of the problem @xmath176 then , by the fredholm alternative , there exists a nonzero weak solution @xmath177 to the adjoint problem @xmath178 by elliptic regularity results ( cf .
@xcite , chapter 3 ) , as the boundary of @xmath4 and the coefficients of @xmath2 ( and hence of @xmath173 ) are of class @xmath86 , we have that @xmath179 is a classical solution to the adjoint problem . hence , if @xmath180 is the adjoint of @xmath11 in @xmath102 , for any @xmath181 sufficiently large , we have @xmath182 and by taking the inverse laplace transform , we obtain @xmath183 for any @xmath47 . +
now , due to the positivity of the semigroup @xmath37 ( and hence of the semigroup @xmath184 ) and to the fact that @xmath37 is conservative , we have that the set @xmath185 is a lattice , that is , @xmath186 for any @xmath187 . therefore , if we set @xmath188 we have that @xmath189 for any @xmath47 and hence @xmath190 is a probability measure and is invariant for @xmath37 . as @xmath33 is the unique invariant measure for @xmath37 ,
we are done . concerning the coefficients @xmath70 , @xmath71 and @xmath79 we assume the following conditions . [ h2 ] 1 .
the mappings @xmath191 are measurable and the mappings @xmath192 are lipschitz continuous , uniformly with respect to @xmath193\times d$ ] , for any @xmath52 .
namely , for any @xmath194 @xmath195\times d}}|f(t , x,\xi)-f(t , x,\eta)|&\leq & l_{t , f } |\xi-\eta|,\\ { \sup_{(t , x ) \in[0,t]\times d}}|g(t , x,\xi)-g(t , x,\eta)|&\leq & l_{t , g } |\xi-\eta|.\end{aligned}\ ] ] 2 .
the mapping @xmath196 is measurable and for any @xmath52 , @xmath197}}|\sigma(t,\cdot)|_{l^\infty(\partial d)}=:c_{t,\sigma } < \infty.\ ] ] in what follows , for any @xmath47 and @xmath198 , we shall define @xmath199 and @xmath200(x):=g(t , x , h_1(x))h_2(x),\qquad x \in d.\ ] ] due to hypothesis [ h2 ] , we have that @xmath201 , @xmath202 and @xmath203 are all lipschitz continuous , uniformly with respect to @xmath204 $ ] , for any @xmath52 .
notice that the same is true for the mappings @xmath205 , @xmath206 and @xmath207 .
analogously , if , for any @xmath47 and @xmath208 , we set @xmath209(x):=\sigma(t , x)z(x),\qquad x \in\partial d,\ ] ] then we have that @xmath210 is a bounded linear operator on @xmath125 and for any @xmath52 , @xmath211.\ ] ] finally , concerning the noisy perturbations @xmath212 and @xmath213 , we assume that they are two independent cylindrical wiener processes defined on the same stochastic basis @xmath214 , taking values in @xmath102 and @xmath125 , respectively , with respective covariance operators @xmath215 and @xmath216 .
namely , @xmath217 where @xmath218 is the orthonormal basis of @xmath102 which diagonalizes @xmath24 , with eigenvalues @xmath219 , @xmath220 is the orthonormal basis of @xmath125 which diagonalizes @xmath26 , with eigenvalues @xmath221 , and @xmath222 and @xmath223 are two sequences of independent standard brownian motions , both defined on the stochastic basis @xmath214 .
notice that the two sequences above are not convergent in @xmath102 and @xmath125 , but in any hilbert spaces @xmath224 and @xmath225 which contain @xmath102 and @xmath125 , respectively , with hilbert
schmidt embedding .
moreover , in the case @xmath7 , we have @xmath226 and hence @xmath227 where @xmath228 and @xmath229 is a two - dimensional standard brownian motion . in what follows , we shall assume the following summability conditions on the eigenvalues @xmath230 and @xmath231 and the sup - norm of the corresponding eigenfunctions .
[ h3 ] 1 .
if @xmath232 , then there exists @xmath233 such that @xmath234 2 . if @xmath232 , then there exists @xmath235 such that @xmath236 [ rem22 ] 1 . from the proofs of lemmas [ lemma33 ] , [ mar4 ] and [ tax10 ] , it is possible to see that if the mapping @xmath237\times d\times{\mathbb r}\to{\mathbb r}$ ] is uniformly bounded for any @xmath52 , then we do not need to require that the sequence @xmath238 is contained in @xmath239 and condition ( [ lambda ] ) can be replaced by @xmath240 2 . as both @xmath241 and @xmath242 are strictly greater than @xmath243 , neither @xmath24 nor @xmath26 are required to be hilbert
schmidt operators in general .
moreover , in space dimension @xmath7 , we have no conditions on the eigenvalues @xmath244 and we can take @xmath245 .
this means that we can deal with space time white noise .
in this section , we are concerned with uniform bounds for the @xmath246th moments of the @xmath247;h)$]-norm of the mild solution @xmath28 of ( [ eq1 ] ) .
we first recall some general facts about the linear parabolic equation with nonhomogeneous boundary conditions @xmath248 where @xmath60 is a @xmath125-valued function .
if @xmath249 is twice continuously differentiable and there exists @xmath250 such that @xmath251 for @xmath252 , then the solution of problem ( [ detlin ] ) is given by @xmath253 ( for a proof , see , e.g. , @xcite , proposition 13.2.1 ) .
such a formula can be extended by continuity to less regular functions @xmath60 .
in particular , for each @xmath64 , we can consider the problem @xmath254 where @xmath16 is the cylindrical wiener process defined in @xmath125 , introduced in section [ sec2 ] .
in analogy to formula ( [ solution ] ) , by taking @xmath255 and @xmath256 , we say that for any @xmath257 $ ] , the process @xmath68,\qquad t\geq0,\ ] ] is a _ mild solution _ to problem ( [ stoclin ] ) .
the process @xmath67 can be interpreted as a _ boundary ornstein uhlenbeck _ process and can be written as the infinite series @xmath258 \,d\hat { \beta}_k(s),\qquad t\geq0.\ ] ] as proved in the next lemma , such a series is well defined in @xmath259;h))$ ] for any @xmath52 and @xmath53 .
moreover , a uniform estimate with respect to @xmath260 $ ] holds .
[ lemma31 ] under part 2 of hypothesis [ h3 ] , the process @xmath261 belongs to @xmath259;h))$ ] for any @xmath52 , @xmath53 and @xmath260 $ ] , and @xmath262}{\mathbb e}|w^\varepsilon _ { a , b}|^p_{c([0,t];h)}=:c_{t , p}<\infty.\ ] ] as a consequence of the stochastic fubini theorem and of the elementary identity @xmath263 we have the factorization formula @xmath264 where @xmath265\ ] ] ( for a proof , see @xcite ) . by the hlder inequality
, this implies that for any @xmath266 , @xmath267 } } |w^\varepsilon_{a , b}(t)|_h^p\nonumber\\ & & \qquad\leq c_{t , p,\alpha}\int_0^t { \mathbb e}|y_{\varepsilon,\alpha}(s)|_h^p \,ds\nonumber\\[-8pt]\\[-8pt ] & & \qquad\leq c_{t , p,\alpha}\int_0^t\biggl(\int_0^s(s - r)^{-2\alpha}\nonumber\\ & & \qquad\quad\hspace*{63.4pt}{}\times\sum_{k \in { \mathbb n}}\theta_k^2 \bigl|(\delta_0-a)e^{(s - r)a/\varepsilon}n_{\delta_0 } [ \sigma(r)f_k]\bigr|_h^2 \,dr\biggr)^{p/2 } \,ds,\nonumber \ ] ] the last inequality following from the burkholder davis gundy inequality
. now , assume that @xmath268 ( the case @xmath7 is simpler ) .
according to ( [ theta ] ) , we have @xmath269\bigr|_h^2\nonumber\\ & & \qquad\leq\kappa_b^{2/\beta}\biggl(\sum_{k \in{\mathbb n}}\bigl|(\delta _ 0-a)e^{(s - r)a/\varepsilon}n_{\delta_0 } [ \sigma(r)f_k ] \bigr|_h^2\biggr)^{1/\zeta}\\ & & \qquad\quad\hspace*{0pt}{}\times \sup_{k \in{\mathbb n}}\bigl|(\delta_0-a)e^{(s - r)a/\varepsilon}n_{\delta_0 } [ \sigma(r)f_k]\bigr|_h^{{2(\zeta-1)}/\zeta},\nonumber \ ] ] where @xmath270 .
thanks to ( [ dominio ] ) and ( [ sobolev ] ) , for any @xmath271 , we have @xmath272 hence , for any @xmath64 and @xmath273 , due to ( [ sigma ] ) , we have @xmath274\bigr|_h^2\nonumber\\ & & \qquad= \sum_{k \in{\mathbb n}}\bigl|e^{{(s - r)}/2 a/\varepsilon}(\delta _ 0-a)^{({1+\rho})/4}e^{{(s - r)}/2 a/\varepsilon}s_\rho \sigma(r)f_k\bigr|_h^2\nonumber\\ & & \qquad=\sum_{k \in{\mathbb n } } \sum_{h \in{\mathbb n}}\bigl|\bigl\langle f_k,\sigma ( r)s_\rho^\star \bigl[(\delta_0-a)^{({1+\rho})/ 4}e^ { { ( s - r)}/2 a/\varepsilon}\bigr]^\star e^{{(s - r)}/2{a^\star } /\varepsilon } e_h\bigr\rangle_z\bigr|^2\\ & & \qquad=\sum_{h \in{\mathbb n}}\bigl|\sigma(r)s_\rho^\star \bigl[(\delta _ 0-a)^{({1+\rho})/4}e^{{(s - r)}/2 a/\varepsilon}\bigr]^\star e^{{(s - r)}/2{a^\star}/\varepsilon } e_h\bigr|_z^2\nonumber\\ & & \qquad\leq c_{t,\rho } \biggl[\biggl(\frac\varepsilon{s - r}\biggr)^ { ( { 1+\rho})/2}+1\biggr]\sum_{h \in{\mathbb n}}\bigl|e^{{(s - r)}/2 { a^\star}/\varepsilon } e_h\bigr|_h^2.\nonumber \ ] ] as the semigroup @xmath37 admits an integral kernel @xmath113 , that is , @xmath275 we have @xmath276 this implies @xmath277\\[-8pt ] & = & \sum_{h \in{\mathbb n}}\int _
d\bigl|\bigl\langle k_{({s - r})/({2\varepsilon})}(\cdot , y),e_h\bigr\rangle_h\bigr|^2 \,dy\nonumber\\ & = & \int_d\bigl|k_{({s - r})/({2\varepsilon})}(\cdot , y)\bigr|_h^2 \,dy.\nonumber \ ] ] now , due to ( [ kernel ] ) , for any @xmath75 and @xmath278 , we have @xmath279 and hence @xmath280 this implies that for any @xmath64 , @xmath281,\ ] ] so , thanks to ( [ somma ] ) , we have @xmath282\bigr|_h^2\biggr)^{1/\zeta}\nonumber\\[-8pt]\\[-8pt ] & & \qquad\leq c_{t,\rho } \biggl[\biggl(\frac\varepsilon{s - r}\biggr)^{({d+1+\rho } ) /({2\zeta})}+ 1\biggr].\nonumber\end{aligned}\ ] ] next , by proceeding as in ( [ somma ] ) , we have @xmath283\bigr|_h^{{2(\zeta-1)}/\zeta}\nonumber\\[-8pt]\\[-8pt ] & & \qquad\leq c_{t,\rho } \biggl[\biggl(\frac\varepsilon{s - r}\biggr)^{{(1+\rho)(\zeta -1)}/({2\zeta})}+1\biggr].\nonumber\end{aligned}\ ] ] therefore , thanks to ( [ 1 ] ) , ( [ 2 ] ) , ( [ quadrato ] ) and ( [ 3 ] ) , we can conclude that for any @xmath260 $ ] , @xmath284 } |w^\varepsilon_{a , b}(t)|_h^p\leq c_{t , p,\alpha , \rho}\biggl(\int_0^t\bigl[s^{-(2\alpha+({d+\zeta})/({2\zeta})+ { \rho}/2)}+ 1\bigr ] \,ds\biggr)^{p/2}.\ ] ] now , as in hypothesis [ h3 ] , we are assuming that @xmath285 , so we have @xmath286 .
this means that we can fix @xmath287 and @xmath288 such that @xmath289 and then , for any @xmath290 we obtain @xmath72 } { \mathbb e}\sup_{t \in[0,t ] } |w^\varepsilon _ { a , b}(t)|_h^p\leq c_{t , p}.\ ] ] the estimate for general @xmath53 follows from the hlder inequality .
next , we pass to ( [ eq1 ] ) .
let @xmath52 and @xmath53 .
an adapted process @xmath291,@xmath292;h))$ ] is a _ mild _ solution of ( [ eq1 ] ) if , for any @xmath204 $ ] , @xmath293 where , for any @xmath294;h))$ ] , we define @xmath295 as is well known , @xmath296 is the unique mild solution of the problem @xmath297 where @xmath15 is the cylindrical wiener process with values in @xmath102 , introduced in section [ sec2 ] . as for @xmath261
, we show that @xmath298 satisfies a bound in @xmath259;h))$ ] which is uniform with respect to @xmath260 $ ] .
[ lemma33 ] assume hypothesis [ h2 ] and part 1 of hypothesis [ h3 ] .
then , @xmath298 is lipschitz continuous from @xmath299;h))$ ] into itself for any @xmath52 and @xmath53 , and @xmath300}{\mathbb e}|w^\varepsilon _ { a , q}(u)|^p_{c([0,t];h)}\leq c_{t , p}\biggl(1+{\mathbb e}\int_0^t|u(s)|_h^p \,ds\biggr).\ ] ] the proof of the lipschitz continuity of @xmath298 in @xmath259 $ ] ; @xmath301 is classical and can be found in , for example , @xcite . concerning estimate ( [ aprioriq ] ) , as in the proof of lemma [ lemma31 ]
, we use a factorization argument and , for any @xmath266 , we get @xmath302 } |w^\varepsilon_{a , q}(t)|_h^p\\ & & \qquad\leq c_{t , p,\alpha } { \mathbb e}\int_0^t\biggl(\int_0^s(s - r)^{-2\alpha}\sum_{k \in{\mathbb n}}\lambda_k^2 \bigl|e^{(s - r ) a/\varepsilon } [ g(r , u(r))e_k]\bigr|_h^2 \,dr\biggr)^{p/2 } \,ds.\end{aligned}\ ] ] according to ( [ lambda ] ) , if we set @xmath303 , then we have @xmath304\bigr|_h^2\nonumber\\ & & \qquad\leq\kappa_q^{2/\rho}\biggl(\sum_{k \in{\mathbb n } } \bigl|e^{(s - r ) a/\varepsilon}[g(r , u(r))e_k]\bigr|_h^2\biggr)^{1/ \zeta}\\ & & \qquad\quad{}\times\sup_{k \in{\mathbb n}}\bigl|e^{(s - r)a/\varepsilon}[g(r , u(r))e_k ] \bigr|_h^{{2(\zeta-1)}/\zeta}|e_k|_\infty^{-4/\rho}.\nonumber \ ] ] as in the proof of ( [ sommabis ] ) , we have @xmath305\bigr|_h^2=\int_d \bigl|k_{({s - r})/\varepsilon}(x,\cdot)g(r,\cdot , u(r))\bigr|_h^2 \,dx.\ ] ] now , thanks to ( [ kernel ] ) , for any @xmath75 , @xmath306 and @xmath164 , we have @xmath307 and this is meaningful since @xmath37 is well defined in @xmath308 . in particular , for any @xmath64 , @xmath309\bigr|_h^2\\ & & \qquad\leq c\biggl[\biggl(\frac\varepsilon{s - r } \biggr)^{d/2}+ 1\biggr]\int_d e^{(s - r)a/\varepsilon}g^2(r,\cdot , u(r))(x ) \,dx\\ & & \qquad = c \biggl[\biggl(\frac\varepsilon{s - r}\biggr)^{d/2}+ 1 \biggr]\bigl|e^{(s - r)a/\varepsilon}g^2(r,\cdot , u(r))\bigr|_{l^1(d)}\\ & & \qquad\leq c \biggl[\biggl(\frac\varepsilon{s - r}\biggr)^{d/2}+ 1\biggr]|g(r,\cdot , u(r))|_h^2\end{aligned}\ ] ] and , due to the linear growth of @xmath71 , @xmath310\bigr|_h^2\biggr)^{1/\zeta}\nonumber\\[-8pt]\\[-8pt ] & & \qquad\leq c_t \biggl[\biggl(\frac\varepsilon{s - r}\biggr)^{d/({2\zeta})}+ 1 \biggr]\bigl(1+|u(r)|_h^{2/\zeta}\bigr).\nonumber\end{aligned}\ ] ] by analogous arguments , we have @xmath311\bigr|_h^ { { 2(\zeta-1)}/\zeta}|e_k|_\infty^{-4/\rho}\leq c_t \bigl(1+|u(r)|_h^{{2(\zeta-1)}/\zeta}\bigr)\ ] ] and then , thanks to ( [ 22 ] ) , ( [ 33 ] ) and ( [ 33bis ] ) , we get , for any @xmath260 $ ] , @xmath302 } |w^\varepsilon_{a , q}(t)|_h^p\\ & & \qquad\leq c_{t , p,\alpha } { \mathbb e}\int_0^t\biggl(\int_0^s\biggl[\biggl(\frac1 { s - r}\biggr)^{2\alpha+{d}/({2\zeta})}+1\biggr]\bigl(1+|u(r)|_h^2\bigr)\ , dr \biggr)^{p/2}\ , ds.\nonumber\end{aligned}\ ] ] as we are assuming @xmath233 , we can find @xmath287 such that @xmath312 . due to the young inequality , this implies ( [ aprioriq ] ) for all @xmath313 and hence for all @xmath53 .
according to lemmas [ lemma31 ] and [ lemma33 ] , we have the following result . [ sol ] under hypotheses [ h1 ] , [ h2 ] and [ h3 ]
, for any @xmath52 and @xmath53 , and for any @xmath314 and @xmath64 , equation ( [ eq1 ] ) admits a unique adapted mild solution @xmath315;h))$ ] .
moreover , @xmath316}{\mathbb e}|u_\varepsilon|_{c([0,t];h)}^p\leq c_{t , p}(1+|u_0|_h^p).\ ] ] as both @xmath201 and @xmath317;h))\to l^p(\omega$ ] ; @xmath247;h))$ ] are lipschitz continuous and @xmath318;h))$ ] , we have that the mapping @xmath319 defined by @xmath320 is lipschitz continuous from the space of adapted processes in @xmath259 $ ] ; @xmath301 into itself .
therefore , by a classical fixed point argument , equation ( [ eq1 ] ) admits a unique adapted mild solution @xmath321;h))$ ] .
next , for any @xmath64 , we have @xmath322 and then , according to ( [ apriorib ] ) and ( [ aprioriq ] ) , we conclude that @xmath284 } |u_\varepsilon(t)|_h^p\leq c_{t , p } \biggl(1+|u_0|_h^p+\int_0^t{\mathbb e}\sup_{r \in[0,s]}|u_\varepsilon(r)|_h^p \,ds\biggr).\ ] ] the gronwall lemma allows us to obtain ( [ epsilon ] ) .
in this section , we show that for any @xmath62 and @xmath53 , the sequence @xmath323}$ ] converges in @xmath324;h_\mu))$ ] to the solution of a suitable one - dimensional stochastic differential equation . in what follows ,
we first introduce the limiting equation by constructing the coefficients and by describing a situation in which they are given by a simple expression . in the second part of this section
, we prove the convergence result .
we start with the drift term . for each @xmath47 and @xmath164
, we define @xmath325 where @xmath326 is the unique invariant measure associated with the semigroup @xmath37 ( see section [ sec2 ] and hypothesis [ h1 ] ) . according to hypothesis [ h2 ] , for any @xmath52 and @xmath327 , we have @xmath328\times d,\ ] ] so that @xmath329 is lipschitz continuous , uniformly with respect to @xmath204 $ ] , for any @xmath52 .
notice that , as @xmath330 , this implies that @xmath331 is also lipschitz continuous .
next , we construct the term arising from the stochastic convolution@xmath332 . for each @xmath47 and @xmath164 , we introduce the linear mapping @xmath333 as @xmath102 is continuously embedded into @xmath162 , for any @xmath52 , we have
@xmath334 this means that there exists @xmath335 such that @xmath336 moreover , since for any @xmath337 and @xmath52 , @xmath338 we have that the mapping @xmath339 is lipschitz continuous , uniformly with respect to @xmath204 $ ] , for any @xmath52 .
this , in particular , implies that the mapping @xmath340 is also lipschitz continuous , both in @xmath102 and in @xmath162 , uniformly for @xmath204 $ ] .
finally , we construct the term arising from the boundary convolution @xmath67 . for each fixed @xmath47
, we introduce the mapping @xmath341,\mu\rangle=\delta_0\int_d n_{\delta_0}[\sigma(t,\cdot)h](x ) \mu(dx ) \in{\mathbb r}.\ ] ] as @xmath45 is a bounded linear operator from @xmath125 into @xmath102 , @xmath342 is bounded and linear in @xmath125 and @xmath102 is continuously embedded in @xmath162 , such a mapping is bounded and linear from @xmath125 into @xmath343 and then , for any @xmath47 , there exists @xmath344 such that for any @xmath119 , we have @xmath345,\mu \rangle = \delta_0\int_d n_{\delta_0}[\sigma(t,\cdot)h](x ) \mu(dx).\ ] ] we can now introduce the limiting equation .
it is the one - dimensional stochastic differential equation @xmath346 as the mappings @xmath347 and @xmath348 are both lipschitz continuous , uniformly with respect to @xmath204 $ ] , for any @xmath52 , equation ( [ eq2 ] ) admits a unique strong solution @xmath349;{\mathbb r}))$ ] for any @xmath53 and @xmath52 , that is , there exists a unique adapted process in @xmath259;{\mathbb r}))$ ] which is adapted to the filtration @xmath350 such that @xmath351 where @xmath352 notice that both @xmath353 and @xmath354 are @xmath355-martingales having zero mean .
moreover , we have @xmath356 and @xmath357 in particular , as @xmath15 and @xmath16 are independent , we have that @xmath358 is an @xmath355-martingale having zero mean and covariance @xmath359 so that there exists some brownian motion @xmath56 defined on some stochastic basis @xmath360 such that the solution of problem ( [ eq2 ] ) coincides in law with the solution of the problem @xmath361 where @xmath362 as shown in remark [ rem21 ] , in the case where the operator @xmath11 is self - adjoint , we have @xmath363 so that , due to the definition of @xmath364 , we get @xmath365(x)|^2 \,dx.\ ] ] concerning the boundary term , due to ( [ adjoint ] ) , we have @xmath366,1\rangle_h|^2\\ & = & \sum_{k\in{\mathbb n } } \frac{\delta_0 ^ 2}{|d|^2}|\langle[\sigma(t)bf_k],n_{\delta_0}^\star 1\rangle_z|^2\\ & = & \sum_{k \in{\mathbb
n } } \frac{1}{|d|^2}|\langle f_k , b\sigma(t,\cdot)\rangle_z|^2=\frac 1{|d|^2}|b\sigma(t,\cdot)|_z^2\\ & = & \frac1{|d|^2}\int_{\partial d}|[b \sigma(t,\cdot)](\eta)|^2\ , d\eta.\end{aligned}\ ] ] therefore , in the self - adjoint case , we have @xmath367(x)|^2 \,dx+\int_{\partial d}|[b \sigma(t,\cdot)](\eta)|^2 \,d\eta \biggr)^{1/2}.\ ] ] now that we have described the candidate limit equation , we prove that @xmath28 in fact converges to its solution .
[ mar8 ] assume hypotheses [ h1 ] , [ h2 ] and [ h3 ] .
then , for any @xmath314 , @xmath53 , @xmath52 and @xmath368 , and for any @xmath147 , we have @xmath369}|u_\varepsilon(t)-v(t)|_{h_\mu}^p&\leq & c_{t , p,\theta } ( \varepsilon+\varepsilon^{{p\theta}/2 } ) ( 1+|u_0|_{h_\mu}^p)\nonumber\\[-8pt]\\[-8pt ] & & { } + e^{-{\gamma p \delta}/\varepsilon}|u_0|_{h_\mu}^p,\nonumber\end{aligned}\ ] ] where @xmath60 is the solution of the one - dimensional problem ( [ eq2 ] ) .
in particular , @xmath370}|u_\varepsilon ( t)-v(t)|_{h_\mu}^p=0.\ ] ] we have @xmath371 where @xmath372\\[-8pt ] & & { } + w^\varepsilon_{a , q}(u_\varepsilon)(t)-\hat{w}_{a , q}(u_\varepsilon ) ( t)+w^{\varepsilon}_{a , b}(t)-\hat{w}_{a , b}(t ) .
\nonumber \ ] ] this yields @xmath373 due to the lipschitz continuity of @xmath374 , for any @xmath375 , we have @xmath376 } \int_0^s |\hat{f}(r , u_\varepsilon ( r))-\hat{f}(r , v(r))|^p \,dr\nonumber\\[-8pt]\\[-8pt ] & & \qquad\leq c_{t , p}\int_0^t{\mathbb e}|u_\varepsilon ( r)-v(r)|_{h_\mu}^p \,dr.\nonumber \ ] ] analogously , due to the lipschitz continuity of @xmath339 and the burkholder
gundy inequality , for any @xmath375 , we easily obtain @xmath377 } \biggl|\int_0^s \langle \hat { g}(r , u_\varepsilon(r))-\hat{g}(r , v(r)),dw^{q}(r)\rangle_h\biggr|^p\nonumber\\[-8pt]\\[-8pt ] & & \qquad\leq c_{t , p}\int_0^t{\mathbb e}|u_\varepsilon(r)-v(r)|_{h_\mu}^p \,dr.\nonumber \ ] ] then , thanks to condition ( [ ergo ] ) , for any @xmath375 , @xmath378}|r_\varepsilon(t)|_{h_{\mu}}^p+\int_0^t{\mathbb e}|u_\varepsilon ( s)-v(s)|_{h_\mu}^p \,ds\biggr)\end{aligned}\ ] ] and , by comparison , this yields @xmath379}|r_\varepsilon ( t)|_{h_{\mu}}^p\bigr).\ ] ] in view of ( [ f100 ] ) , thanks to ( [ f1 ] ) and ( [ f2 ] ) , for any @xmath62 , we obtain @xmath380 } |u_\varepsilon(t)-v(t)|_{h_\mu}^p\\ & & \qquad\leq c e^{-{\gamma p \delta}/\varepsilon}|u_0|_{h_\mu}^p+
c_{t , p}\int_0^t{\mathbb e}|u_\varepsilon(s)-v(s)|_{h_\mu}^p \,dt\\ & & \qquad\quad{}+c_{t , p } { \mathbb e}\sup_{t \in[0,t]}|r_\varepsilon(t)|_{h_{\mu}}^p.\end{aligned}\ ] ] therefore , if we show that , for any @xmath52 , @xmath53 and @xmath381 , @xmath382}|r_\varepsilon(t)|_{h_\mu}^p\leq c_{t , p,\theta } \varepsilon^{{p\theta}/2 } ( 1+|u_0|_h^p),\ ] ] then we can conclude that ( [ mar9 ] ) holds .
due to ( [ repsilon ] ) , in order to prove ( [ f3 ] ) and hence complete the proof of theorem [ mar8 ] , we need the following three lemmas . [ mar7 ] assume hypotheses [ h1 ] , [ h2 ] and [ h3 ] .
then , for any @xmath52 and @xmath53 , and for any @xmath260 $ ] , we have @xmath383}\biggl|\int_0^t e^{(t - s ) a/\varepsilon } f(s , u_\varepsilon(s ) ) \,ds-\int_0^t\hat{f}(s , u_\varepsilon(s ) ) \,ds \biggr|_{h_\mu}^p\nonumber\\[-8pt]\\[-8pt ] & & \qquad\leq c_{t , p } ( 1+|u_0|_h^p ) \varepsilon^p.\nonumber\end{aligned}\ ] ] due to hypothesis [ h1 ] , for any @xmath204 $ ] , we have @xmath384 this implies that , for any @xmath204 $ ] , @xmath385 so that , thanks to ( [ epsilon ] ) , for any @xmath260 $ ] , we obtain @xmath302}\biggl|\int_0^t e^{(t - s ) a/\varepsilon } f(s , u_\varepsilon(s ) ) \,ds-\int_0^t\hat{f}(s , u_\varepsilon(s ) ) \,ds \biggr|_{h_\mu}^p\\ & & \qquad\leq c_{t , p}(1+|u_0|^p_h)\varepsilon^p.\end{aligned}\ ] ] [ mar4 ] assume hypotheses [ h1 ] , [ h2 ] and [ h3 ] , and fix @xmath52 , @xmath53 and @xmath368 .
then , there exists some constant @xmath386 such that for any @xmath260 $ ] , @xmath387}|w^\varepsilon_{a , q}(u_\varepsilon)(t)-\hat { w}_{a , q}(u_\varepsilon)(t)|_{h_\mu}^p\leq c_{t , p,\theta } \varepsilon^{{p\theta}/2 } ( 1+|u_0|_h^p).\ ] ] as in the proofs of lemmas [ lemma31 ] and [ lemma33 ] , we use a factorization argument . since @xmath388 , for any @xmath47 and @xmath389 , we have @xmath390 where @xmath391 and , for any @xmath392 , @xmath393 hence , due to ( [ mar10 ] ) , @xmath37 is a contraction in @xmath162 for any @xmath47 , and by proceeding as in the proofs of lemmas [ lemma31 ] and [ lemma33 ] , for @xmath394 , we obtain @xmath302}\biggl|w^\varepsilon_{a , q}(u_\varepsilon ) ( t)-\int_0^t\langle\hat{g}(s , u_\varepsilon(s)),dw^q(s)\rangle_h \biggr|^p_{h_\mu}\\ & & \qquad\leq c_{t , p,\alpha } { \mathbb e}\int_0^t\biggl(\int_0^s(s - r)^{-2\alpha}\sum_{k \in{\mathbb n}}\lambda_k^2 \bigl|e^{(s - r ) a/\varepsilon}\psi ( r , u_\varepsilon(r))e_k\bigr|_{h_\mu}^2 \,dr\biggr)^{p/2}\,ds.\end{aligned}\ ] ] due to the invariance of @xmath33 and condition ( [ ergo ] ) , we have @xmath395-\langle \hat{g}(r , u_\varepsilon(r)),e_k\rangle_h\bigr|_{h_\mu}\\ & & \qquad=\bigl|e^{({s - r})/2 a/\varepsilon}\bigl(e^{({s - r})/2 a/\varepsilon}[g(r , u_\varepsilon(r))e_k]\bigr)\\ & & \hspace*{61.6pt}{}-\bigl\langle e^ { ( { s - r})/2 a/\varepsilon}[g(r , u_\varepsilon(r))e_k],\mu\bigr\rangle \bigr|_{h_\mu}\\ & & \qquad\leq c e^{-{\gamma(s - r)}/({2\varepsilon})}\bigl|e^ { ( { s - r})/2 a/\varepsilon}[g(r , u_\varepsilon(r))e_k]\bigr|_{h_\mu}\end{aligned}\ ] ] so that @xmath396}\biggl|w^\varepsilon_{a , q}(u_\varepsilon ) ( t)-\int_0^t\langle \hat{g}(s , u_\varepsilon(s)),dw^q(s)\rangle_h \biggr|^p_{h_\mu}\nonumber\\ & & \qquad\leq c_{t , p,\alpha } { \mathbb e}\int_0^t \biggl(\int_0^s(s - r)^{-2\alpha } e^{-{\gamma(s - r)}/{\varepsilon}}\\ & & \qquad\quad\hspace*{70.5pt}{}\times\sum_{k \in{\mathbb n}}\lambda_k^2 \bigl|e^{({s - r})/2 a/\varepsilon}[g(r , u_\varepsilon(r))e_k ] \bigr|_{h_\mu}^2 \,dr\biggr)^{p/2 } \,ds.\nonumber \ ] ] using the same arguments that were used in the proof of lemma [ lemma33 ] [ see ( [ 33 ] ) and ( [ 33bis ] ) ] , for any @xmath397 , we get @xmath398\bigr|_{h_\mu}^2\leq c_t \biggl[\biggl(\frac\varepsilon{s - r}\biggr)^ { d/({2\zeta})}+ 1\biggr ] \bigl(1+|u_\varepsilon(r)|_h^{2}\bigr),\ ] ] with @xmath399 if @xmath268 and with @xmath400 if @xmath7 . thanks to ( [ epsilon ] ) , this yields @xmath302}\biggl|w^\varepsilon_{a , q}(u_\varepsilon ) ( t)-\int_0^t\langle \hat{g}(s , u_\varepsilon(s)),dw^q(s)\rangle_h \biggr|^p_{h_\mu}\\ & & \qquad\leq c_{t , p,\alpha}(1+|u_0|_h^p)\varepsilon^{-\alpha p}\biggl(\int _ 0^t\biggl[\biggl(\frac\varepsilon
t\biggr)^{(2\alpha+{d}/({2\zeta}))}+ 1\biggr]e^{-{\gamma t}/{\varepsilon } } \,dt\biggr)^{p/2}.\end{aligned}\ ] ] now , according to the first condition in hypothesis [ h3 ] , we have @xmath401 so that , for any @xmath368 , we can fix @xmath287 such that @xmath402 then , with a change of variable , we easily obtain @xmath403}\biggl|w^\varepsilon_{a , q}(u_\varepsilon)(t)-\int _ 0^t\langle \hat{g}(s , u_\varepsilon(s)),dw^q(s)\rangle_h\biggr|^p_{h_\mu}\leq c_{t , p,\theta } \varepsilon^{{p\theta}/2}(1+|u_0|_h^p)\ ] ] for any @xmath290 . by the hlder inequality , we obtain an analogous estimate for any @xmath53 and ( [ limg ] ) then follows . [ mar6 ] assume hypotheses [ h1 ] , [ h2 ] and [ h3 ] , and fix any @xmath52 , @xmath53 and @xmath368 . then , there exists some constant @xmath386 such that for any @xmath260 $ ] , @xmath404}|w^\varepsilon_{a , b}(t)-\hat{w}_{a , b}(t ) \leq c_{t , p,\theta } \varepsilon^ { { p\theta}/2}.\ ] ] notice that @xmath405 for any @xmath47 .
then , as in lemma [ lemma31 ] , by factorization , we obtain @xmath406 where @xmath407 and for any @xmath119 , @xmath408-\frac1{\delta_0}\langle \hat{\sigma } ( r),h\rangle_z.\ ] ] hence , according to ( [ mar10 ] ) , by arguing as in the proofs of lemmas [ lemma31 ] and [ lemma33 ] , for any @xmath409 , we obtain @xmath410 } \biggl|w^\varepsilon_{a , b}(t)-\int_0^t \langle \hat{\sigma}(s),dw^b(s)\rangle_z\biggr|_{h_\mu}^p\\ & & \qquad \leq c_{t , p,\alpha}\int_0^t\biggl(\int_0^s(s - r)^{-2\alpha}\\ & & \qquad\quad\hspace*{63.1pt}{}\times\sum_{k \in { \mathbb n}}\theta_k^2 \bigl|(\delta_0-a)e^{(s - r ) a/\varepsilon}[\psi ( r)f_k]\bigr|_{h_\mu}^2 \,dr\biggr)^{p/2 } \,ds.\end{aligned}\ ] ] due to the invariance of @xmath33 and to condition ( [ ergo ] ) , we have @xmath411\bigr|_{h_\mu } \\ & & \qquad=\bigl|(\delta_0-a)e^{(s - r ) a/\varepsilon}n_{\delta_0}[\sigma ( r)f_k]-\delta _ 0\langle n_{\delta_0}[\sigma(r)f_k],\mu\rangle\bigr|_{h_\mu}\\ & & \qquad=\bigl|e^{({s - r})/{2 } a/\varepsilon}\bigl((\delta_0-a)e^ { ( { s - r})/2 a/\varepsilon}n_{\delta_0}[\sigma(r)f_k]\bigr)\\ & & \qquad\quad\hspace*{27.37pt } { } - \bigl\langle ( \delta _ 0-a)e^{({s - r})/2 a/\varepsilon}n_{\delta_0}[\sigma(r)f_k],\mu \bigr\rangle\bigr|_{h_\mu}\\ & & \qquad\leq c e^{-{\gamma(s - r)}/({2\varepsilon})}\bigl|(\delta _ 0-a)e^{({s - r})/2 a/\varepsilon}n_{\delta_0}[\sigma(r)f_k ] \bigr|_{h_\mu}.\end{aligned}\ ] ] this implies that @xmath302 } \biggl|w^\varepsilon_{a , b}(t)-\int_0^t \langle \hat{\sigma}(s),dw^b(s)\rangle_z\biggr|_{h_\mu}^p\\ & & \qquad\leq c_{t , p,\alpha}\int_0^t\biggl(\int_0^s(s - r)^{-2\alpha}e^{- { \gamma(s - r)}/{\varepsilon}}\\ & & \qquad\quad\hspace*{64.3pt}{}\times\sum_{k \in{\mathbb n}}\theta_k^2 \bigl|(\delta _ 0-a)e^{({s - r})/2 a/\varepsilon}n_{\delta_0}[\sigma(r)f_k ] \bigr|_{h_\mu}^2 \,dr\biggr)^{p/2 } \,ds\end{aligned}\ ] ] and , hence , by proceeding as in the proof of lemma [ lemma31 ] , we conclude that @xmath302 } \biggl|w^\varepsilon_{a , b}(t)-\int_0^t \langle \hat{\sigma}(s),dw^b(s)\rangle_z\biggr|_{h_\mu}^p\\ & & \qquad\leq c_{t , p,\alpha,\rho } \varepsilon^{-\alpha p}\biggl(\int_0^t\biggl [ \biggl(\frac\varepsilon s\biggr)^{2\alpha+({d \operatorname{sign } ( d-1)+\zeta } ) /({2\zeta})+{\rho}/2}+ 1\biggr]e^{-{\gamma s}/{\varepsilon}}\ , ds\biggr)^{p/2},\end{aligned}\ ] ] where @xmath412 is a positive constant to be chosen and where @xmath413 if @xmath268 and @xmath400 if @xmath7 .
now , as we are assuming @xmath235 when @xmath232 , for any @xmath368 , we can fix @xmath414 and @xmath415 both positive such that @xmath416 then , with a change of variable , for any @xmath313 , @xmath284 } \biggl|w^\varepsilon_{a , b}(t)-\int_0^t\langle \hat { \sigma}(s),dw^b(s)\rangle_z\biggr|_{h_\mu}^p\leq c_{t , p } \varepsilon ^{{p \theta}/2}\ ] ] and this implies ( [ limsigma ] ) for any @xmath53 . 1 .
notice that from ( [ f1000 ] ) , we have @xmath417 so that @xmath418 2 .
if we take @xmath419 , then , for any @xmath53 , @xmath52 and @xmath368 , we have the stronger estimate @xmath420}|u_\varepsilon(t)-v(t)|_{h_\mu}^p\leq c_{t , p,\theta } \varepsilon^{{p\theta}/2}(1+|u_0|^p).\ ] ] 3 . from the proofs of lemmas [ mar4 ] and [ mar6 ]
, we easily see that for any @xmath52 and @xmath53 , @xmath421}{\mathbb e}|w^\varepsilon_{a , q}(u_\varepsilon)(t)-\hat { w}_{a , q}(t)|_{h_\mu}^p\leq c_{t , p } \varepsilon^{{p}/2 } ( 1+|u_0|_h^p)\ ] ] and @xmath422}{\mathbb e}|w^\varepsilon_{a , b}(t)-\hat { w}_{a , b}(t)|_{h_\mu}^p \leq c_{t , p } \varepsilon^{{p}/2}.\ ] ] then , for any @xmath52 and @xmath53 ,
@xmath423}{\mathbb e}|r_\varepsilon(t)|_{h_\mu}^p\leq c_{t , p } \varepsilon^{p/2}(1+|u_0|_{h_\mu}^p ) , \qquad\varepsilon \in(0,1].\ ] ] then , by repeating the arguments used in the proof of theorem [ mar8 ] , we have @xmath424}{\mathbb e}|u_\varepsilon(t)-v(t)|_{h_\mu}^p\leq c_{t , p } ( \varepsilon+\varepsilon^{{p}/2})(1+|u_0|_{h_\mu } ^p)+e^{-{\gamma p \delta}/\varepsilon}|u_0|_{h_\mu}^p.\ ] ] moreover , if @xmath419 , as in ( [ mar100 ] ) , we have @xmath425}{\mathbb e}|u_\varepsilon(t)-v(t)|_{h_\mu}^p\leq c_{t , p } \varepsilon^{{p}/2}(1+|u_0|^p).\ ] ]
in this section , we analyze the fluctuations of the motion @xmath426 around the averaged motion @xmath60 . more precisely , we will study the limiting behavior of the random field @xmath427 as the parameter @xmath29 goes to zero . in what follows , in addition to hypothesis [ h2 ]
, we shall assume that the coefficients @xmath70 and @xmath71 satisfy the following conditions .
[ h4 ] 1 .
the mapping @xmath428 is of class @xmath429 , with lipschitz continuous derivative , uniformly with respect to @xmath306 and @xmath204 $ ] , for any @xmath52 .
2 . the mapping @xmath71 does not depend on the third variable , that is , @xmath430 for any @xmath47 , @xmath306 and @xmath431 .
3 . for any @xmath306 , the mappings @xmath432 and @xmath433 are hlder continuous of exponent @xmath434 and @xmath435_{c^\alpha([0,+\infty))}&=&l_g<\infty,\nonumber\\[-8pt]\\[-8pt ] \sup_{\eta \in\partial d}[\sigma(\cdot,\eta)]_{c^\alpha([0,+\infty ) ) } & = & l_\sigma<\infty.\nonumber\end{aligned}\ ] ] from hypothesis [ h4 ] , we easily obtain that the mapping @xmath436 is frchet differentiable and , for any @xmath47 and @xmath437 , we have @xmath438 moreover , @xmath439 is lipschitz continuous , uniformly for @xmath204 $ ] .
[ teo5.1 ] assume hypotheses [ h1][h4 ] .
then , for any @xmath75 , @xmath440 in @xmath162 , where @xmath77 is the gaussian random field defined for any @xmath75 and @xmath306 by @xmath441\\[-8pt ] & & { } + \int_0^\infty\pi ( \delta_0-a)e^{sa}n_{\delta_0}[\sigma(t)\,dw^b(s)](x).\nonumber\end{aligned}\ ] ] [ for any @xmath442 , we have set @xmath443 .
notice that , due to the invariance of @xmath33 , @xmath444\ ] ] we now define @xmath445 and @xmath446.\ ] ] before proceeding with the proof of theorem [ teo5.1 ] , it is important to see that the two terms @xmath447 and @xmath448 are both well defined in @xmath449 for any @xmath47 .
[ fluc5 ] under hypotheses [ h1][h3 ] , @xmath450 due to the invariance of @xmath33 , we have @xmath451-\langle g(t)e_k,\mu\rangle\bigr)\ , d\beta_k(s)\ ] ] so that , by proceeding as in the proof of lemma [ lemma33 ] , thanks to ( [ lambda ] ) , we have @xmath452-\langle g(t)e_k,\mu\rangle|_{h_\mu}^2 \,ds\nonumber\\ & \leq & c \int_0^\infty\biggl ( \sum_{k=1}^\infty\bigl|e^{sa } \bigl([g(t)e_k]-\langle g(t)e_k,\mu\rangle\bigr)\bigr|_{h_\mu}^2\biggr)^{1/\zeta } \\ & & \hspace*{24.2pt}{}\times\sup_{k \in{\mathbb n}}|e^{sa}[g(t)e_k]-\langle g(t)e_k,\mu \rangle|_{h_\mu}^{{2(\zeta-1)}/\zeta}|e_k|_\infty^{- 4/\rho}\,ds,\nonumber \ ] ] where @xmath453 and @xmath412 is the constant appearing in ( [ lambda ] ) .
due to ( [ ergo ] ) and the invariance of @xmath33 , we have @xmath454-\langle g(t)e_k,\mu\rangle\bigr)\bigr|_{h_\mu}^2\leq e^{-\gamma s}|e^{s/ 2 a } [ g(t)e_k]|_{h_\mu}^2\ ] ] so that , according to ( [ 33 ] ) , we have @xmath455-\langle g(t)e_k,\mu \rangle\bigr)\bigr|_{h_\mu}^2\biggr)^{1/\zeta}\leq c_t e^{-{\gamma s}/\zeta } \bigl(s^{- d/({2\zeta})}+1\bigr).\ ] ] analogously , according to ( [ 33bis ] ) , we have @xmath456-\langle g(t)e_k,\mu\rangle c_t e^{-{\gamma(\zeta-1)s}/\zeta}\ ] ] and hence , in view of ( [ fluc3 ] ) , we conclude that @xmath457 as far as @xmath458 is concerned , we have the following , analogous , result .
[ fluc6 ] under hypotheses [ h1][h3 ] @xmath459 due to the invariance of @xmath33 , we have @xmath460-\delta_0\langle n_{\delta_0}[\sigma ( t)f_k],\mu\rangle\bigr ) \,d\hat{\beta}_k(s).\ ] ] using the same arguments used in lemma [ mar6 ] , due to ( [ ergo ] ) and the invariance of @xmath33 , we have @xmath461-\delta_0\langle n_{\delta _ 0}[\sigma ( t)f_k],\mu\rangle|_{h_\mu}^2\\ & & \qquad\leq c e^{-\gamma s}|(\delta _ 0-a)e^{s/2 a}n_{\delta_0}[\sigma(t)f_k]|_{h_\mu}^2\end{aligned}\ ] ] and then , as in the proof of lemma [ lemma31 ] , due to ( [ theta ] ) , we get @xmath462|_{h_\mu}^2\biggr)^{1/\zeta}\\ & & \hspace*{23.8pt}{}\times\sup_{k \in{\mathbb n}}|(\delta_0-a)e^ { s/2 a}n_{\delta _ 0}[\sigma(t)f_k]|_{h_\mu}^{{2(\zeta-1)}/\zeta } \,ds.\end{aligned}\ ] ] by using ( [ quadrato ] ) and ( [ 3 ] ) , this allows us to conclude that for some @xmath288 such that @xmath463 , @xmath464 it is immediate to check that for any @xmath47 , @xmath465 where @xmath466z_\varepsilon(s ) \,ds \,d\theta\\ & = & \!:\sum_{i=1}^3 r_{\varepsilon , i}(t)\end{aligned}\ ] ] and @xmath467 due to ( [ dubbio ] ) , we have @xmath468 for @xmath469 , with a change of variables , due to ( [ ergo ] ) , we have , for any @xmath470 $ ] , @xmath471}|u_\varepsilon(s)|_{h_\mu}\bigr)\int_0^{t}e^{-{\gamma s}/\varepsilon } \,ds\\ & \leq & c_t { \sqrt{\varepsilon}}\bigl(1+\sup_{s \in [ 0,t]}|u_\varepsilon(s)|_{h_\mu}\bigr)\end{aligned}\ ] ] and then , thanks to ( [ epsilon ] ) , we get @xmath472}|r_{\varepsilon,2}(t)|_{h_\mu}\leq c_t \sqrt { \varepsilon } ( 1+|u_0|_{h_\mu}).\ ] ] finally , for @xmath473 , due to the lipschitz continuity of @xmath474 , uniform with respect to @xmath475 $ ] , and estimate ( [ notice ] ) with @xmath476 and @xmath477 , we get @xmath478\\[-8pt ] & \leq & c_t \varepsilon^{\theta-1/2 } ( 1+|u_0|^2).\nonumber\end{aligned}\ ] ] therefore , collecting together ( [ prr1 ] ) , ( [ i1 ] ) and ( [ fluc4 ] ) , we can conclude that for any @xmath52 and @xmath260 $ ] , @xmath479.\hspace*{-36pt}\ ] ] next , for any @xmath64 , we introduce the problem @xmath480 where @xmath481 is the process introduced in ( [ tax5 ] ) . for any @xmath64 , we denote by @xmath482 its unique solution .
we have the following result , whose proof is postponed to the end of this section . [ tax10 ] under hypotheses [ h1][h4 ] , for any @xmath75
, we have @xmath483 in @xmath162 , where @xmath484 is the @xmath162-valued gaussian vector field defined in ( [ tax7 ] ) . for any @xmath306 and @xmath75 , we have @xmath491 then
, if we multiply both sides above by @xmath492 and integrate in @xmath83 with respect to the measure @xmath33 , we get @xmath493 where @xmath494 and @xmath495 it is then immediate to check that @xmath496 so that @xmath497 where @xmath498 _ step _ 1 .
we show that for any @xmath47 , @xmath499 due to ( [ tax5 ] ) and the stochastic fubini theorem , we have @xmath500\biggr\rangle_{h_\mu}\,ds \,d\beta _ k(\sigma)\\ & & \qquad\quad{}+\frac1{\sqrt{\varepsilon}}\sum_{k=0}^\infty\int_0^t\int_\sigma ^th(t , s)\biggl\langle \frac{\partial f}{\partial\xi}(s,\cdot , v(s)),\\ & & \qquad\quad\hspace*{115.1pt}(\delta _ 0-a)e^{(s-\sigma)a/\varepsilon}\\ & & \qquad\quad\hspace*{115.1pt}{}\times\pi[n_{\delta_0}(\sigma(\sigma ) bf_k)]\biggr\rangle_{h_\mu } \,ds \,d\hat{\beta}_k(\sigma).\end{aligned}\ ] ] then , as @xmath15 and @xmath16 are independent and @xmath501 is uniformly bounded , we get @xmath502\bigr|_{h_\mu } \,ds\biggr)^2 \,d\sigma\\ & & \qquad\quad{}+\frac{\kappa_t}\varepsilon\int_0^t \sum_{k=0}^\infty\biggl(\int _ \sigma^t e^{\kappa_t(t - s ) } \bigl|(\delta_0-a)e^{(s-\sigma ) a/\varepsilon}\\ & & \qquad\quad\hspace*{110.7pt}{}\times\pi[n_{\delta_0 } ( \sigma(\sigma)bf_k)]\bigr|_{h_\mu } \,ds\biggr)^2 \,d\sigma\\ & & \qquad=:\frac{\kappa_t}\varepsilon\int_0^t\bigl(j_{\varepsilon,1}(t,\sigma ) + j_{\varepsilon,2}(t,\sigma)\bigr ) \,d\sigma.\end{aligned}\ ] ] for the first term @xmath503 , in view of ( [ ergo ] ) , for any @xmath504 , we have @xmath505\bigr|_{h_\mu } \,ds\biggr)^2\\ & & \qquad\leq\biggl(\int_\sigma^t e^{\kappa_t(2-\alpha)(t - s)}e^{-{\gamma ( 2-\alpha)(s-\sigma)}/({2\varepsilon } ) } \,ds\biggr)^{2/({2-\alpha})}\\ & & \qquad\quad{}\times\sum _ { k=0}^\infty\lambda_k^2 \biggl(\int_\sigma^t \bigl|e^{(s-\sigma ) a/({2\varepsilon})}[g(\sigma)e_k]\bigr|_{h_\mu } ^{({2-\alpha})/({1-\alpha } ) } \,ds\biggr)^{{2(1-\alpha)}/({2-\alpha})}\\ & & \qquad\leq c_t \varepsilon^{2/({2-\alpha } ) } \sum_{k=0}^\infty\lambda_k^2 \biggl(\int_\sigma^t \bigl|e^{(s-\sigma ) a/({2\varepsilon})}[g(\sigma)e_k]\bigr|_{h_\mu } ^{({2-\alpha})/({1-\alpha } ) } \,ds\biggr)^{{2(1-\alpha)}/({2-\alpha})}.\end{aligned}\ ] ] then , if we set @xmath399 , by using the hlder inequality for infinite series , we get @xmath506\bigr|_{h}^{2\zeta } \\ & & \qquad\quad\qquad\quad\hspace*{65.8pt}{}\times|e_k|_\infty^{-4/({\rho-2})}\biggr)^{1/\zeta({2-\alpha } ) /({2(1-\alpha ) } ) } \,ds\biggr)^{{2(1-\alpha)}/({2-\alpha } ) } \ ] ] and , by proceeding as in the proof of lemma [ lemma33 ] , we conclude that for @xmath260 $ ] , @xmath507 now , in view of hypothesis [ h3 ] , we have @xmath401 and can fix @xmath508 such that @xmath509 and then @xmath510 , t\geq0.\ ] ] the same arguments can be repeated for the term @xmath511 , so we can find some @xmath512 such that @xmath513 , t\geq0.\ ] ] this , together with ( [ tax20 ] ) , implies that @xmath514 , t\geq0,\ ] ] where @xmath515 so ( [ tax11 ] ) follows .
_ step _ 2 .
we show that for any fixed @xmath75 , @xmath516 with a change of variable , we have @xmath517\\ & & \hspace*{22.4pt}{}+\int_0^t ( \delta_0-a)e^{(t - s ) a/\varepsilon}\pi[n_{\delta_0}(\sigma ( s)\,dw^b(s))]\biggr)\\ & = & \int_0^{t/\varepsilon } e^{r a}\pi[g(t-\varepsilon r)\,dw_{\varepsilon , t}^q(r)]\\ & & { } + \int_0^{t/\varepsilon } ( \delta_0-a)e^{r a}\pi\bigl[n_{\delta_0}\bigl(\sigma(t-\varepsilon r)\,dw_{\varepsilon , t}^b(r)\bigr)\bigr],\end{aligned}\ ] ] where @xmath518 this means that for any @xmath64 and @xmath75 , @xmath519 where @xmath520\\ & & { } + \int_0^{t/\varepsilon } ( \delta_0-a)e^{r a}\pi\bigl[n_{\delta_0}\bigl(\sigma(t-\varepsilon r)\,dw^b(r ) \bigr)\bigr].\end{aligned}\ ] ] thus , in order to obtain ( [ tax21 ] ) , it is sufficient to prove @xmath521\\ & & { } + \int_0^{t/\varepsilon } ( \delta_0-a)e^{r a}\pi\bigl[n_{\delta_0}\bigl(\bigl(\sigma(t-\varepsilon r)-\sigma(t)\bigr)\,dw^b(r)\bigr)\bigr]\\ & & { } -\int_{t/\varepsilon}^\infty e^{r a}\pi[g(t)\,dw^q(r ) ] \\ & & { } -\int_{t/\varepsilon}^\infty ( \delta_0-a)e^{r a}\pi[n_{\delta_0}(\sigma(t)\,dw^b(r))]\\ & = & \!:\sum _ { i=1}^4 j_{\varepsilon , i}(t).\end{aligned}\ ] ] with the same arguments used several times throughout the paper , we have @xmath522 then , due to hypothesis [ h4 ] , we have @xmath523 analogously , we have @xmath524 concerning @xmath525 , we have @xmath526 so that @xmath527 in an identical way , we can show that @xmath528}{\mathbb e}|j_{\varepsilon , 4}(t)|_{h_\mu}^2=0\ ] ] and this , together with ( [ fluc9 ] ) , ( [ fluc92 ] ) and ( [ fluc10 ] ) , implies ( [ tax22 ] ) .
we would like to thank the two anonymous referees who read the first version of our paper for their interesting and useful remarks and suggestions . | we consider a class of stochastic reaction - diffusion equations also having a stochastic perturbation on the boundary and we show that when the diffusion rate is much larger than the rate of reaction , it is possible to replace the spde by a suitable one - dimensional stochastic differential equation .
this replacement is possible under the assumption of spectral gap for the diffusion and is a result of _ averaging _ in the fast spatial transport .
we also study the fluctuations around the averaged motion . and . |
two - dimensional ( 2d ) quantum dots ( qd s ) created at semiconductor interfaces with refined control of their size , shape and number of electrons are often referred @xcite to as `` artificial atoms '' .
this analogy suggests that the physics of electrons in such man - made nanostructures is closely related to that underlying the traditional description @xcite of natural atoms ( pertaining particularly to electronic shells and the aufbau principle ) , where the electrons are taken @xcite to be moving in a spherically averaged effective central mean field ( cmf ) .
however , using as paradigms impurity - doped ( as well as the limiting case of clean ) two - electron - qd ( @xmath0 qd ) analogs to he - like atoms , we show that the interplay between the interelectron repulsion ( @xmath1 ) and the electronic zero - point kinetic energy ( @xmath2 ) may lead , for a broad range of impurity parameters , to spontaneous symmetry - breaking ( sb ) of the self - consistent mean field , resulting at zero magnetic field ( @xmath3 ) in formation of a singlet ground - state electron molecule .
such sb is beyond the cmf picture and , while negligible in the ground state of the he atom ( whose study was central to the development of the quantum theory of matter due to the failure of the bohr - type models @xcite ) , it is similar in nature to the sb found @xcite in the 1970 s in doubly excited he , where formation of an @xmath4-he@xmath5-@xmath4 `` triatomic '' molecule has been invoked .
furthermore , we show that symmetry breaking at @xmath3 of the self - consistent field associated with the triplet excited state of the @xmath0 qd originates from a jahn - teller distortion of the cmf , similar to that underlying formation of nonspherical open - shell nuclei @xcite and metal clusters .
@xcite along with a unification of concepts pertaining to spontaneous sb in a variety of finite fermionic systems ( from nuclei , metal clusters and natural atoms , to 2d qd s ) , we demonstrate the ability to control the orientation and to manipulate ( i.e. , to enhance , but also to counteract and even to void ) the formation of the electron molecules in 2e qd s via impurity and/or magnetic - field effects .
the ability to control the orientation of the electron molecule may in principle open new possibilities for designing `` on - off '' ( i.e. , switching between two discrete states ) devices , which eventually may be employed in applications of qd s as nanoscale logic gates @xcite ( the effect of impurities on the structure of multi - electron molecules in qd s with more than two electrons is further illustrated in appendix a ) .
that electrons in extended media may undergo crystallization at low densities , when @xmath1 dominates over @xmath2 , has been predicted @xcite by wigner in 1934 . such wigner crystallization ( wc ) in clean qd s results in formation of electron molecules @xcite [ also referred to as wigner molecules ( wm s ) ] , which are associated with spontaneous sb , where the symmetry of the ground state , calculated at the mean - field ( self - consistent - potential ) level , is found to be lower than that of the exact hamiltonian describing the system .
@xcite in clean qd s , formation @xcite of wm s is controlled @xcite by the parameter @xmath6 . for a parabolic confinement ( with frequency @xmath7 ) at @xmath3 , it is customary to take @xmath8 and @xmath9 , where @xmath10 is the dielectric constant and @xmath11 is the spatial extent of the lowest state s wave function of an electron with an effective mass @xmath12 ; wm s occur @xcite for @xmath13 , corresponding to much higher electron densities @xcite than those predicted for wc in an infinite 2d medium .
@xcite the many - body hamiltonian for a qd with @xmath14 electrons can be expressed as a sum of a single - particle part and the two - particle interelectron coulomb repulsion , @xmath15 the contributions to the single - particle part are written as @xmath16 and they contain a term describing the motion of an electron in a 2d parabolic confinement , i.e. , @xmath17 , where @xmath7 is the frequency of the 2d isotropic harmonic confining potential .
magnetic - field effects are included in @xmath18/2m^ * + g^ * \mu_b { \bf b } { \bf \cdot } { \bf s}_i/\hbar$ ] , where the vector potential , @xmath19 , is taken in the symmetric gauge , and the last term is the zeeman interaction with an effective factor @xmath20 , @xmath21 is the electron spin , and @xmath22 the bohr magneton . to include the effect of ( coulombic ) impurities
, we added to @xmath23 the term @xmath24 ( @xmath25 ) , where @xmath26 is the charge of the @xmath27th impurity located at @xmath28 ; such impurities which in general may be situated out of the 2d plane of the qd ( that is @xmath29 ) may correspond to implanted atoms ( donors or acceptors ) or represent a fabricated , controllable voltage - gate .
a clean qd ( that is with @xmath30 in the above hamiltonian ) may be regarded as a realization of the thomson atom , as are jellium models of metal clusters , @xcite where the positive charge is uniformly distributed ; in the @xmath0 thomson qd ( tqd ) , the confinement to the 2d plane is modeled by a parabolic potential . on the other hand ,
a qd analog of the rutherford model of the atom ( rqd ) , where the positive charges are grouped in a single nucleus , can be achieved through the introduction of a central attractive impurity ( in addition to the harmonic confinement ) .
the presence of many free parameters in eqs .
( [ mbh ] ) and ( [ hi ] ) results in a plethora of possible case studies generated by varying the material dependent parameters ( @xmath10 and @xmath12 ) , the harmonic confinement @xmath7 and the magnetic - field @xmath31 , as well as the number of impurities @xmath27 , their charges @xmath26 , and positions @xmath32 . in this paper , we have chosen to discuss here the following three representative classes of cases at a specific value of @xmath33 , which is sufficientlly high so that the electrons form @xcite a wm in the case of a clean qd , thus allowing for systematic investigations of the effects of impurities on the formation , orientational pinning , and structural distortions of the electron molecules ( for the other parameter values used throughout this paper , see ref . ) .
the three representative classes , which we discuss , are : ( i ) a 2e tqd at @xmath3 in the presence of two off - centered impurities situated on both sides of the dot ( section ii ) ; ( ii ) a 2e rqd ( with a single attractive central impurity ) in the presence of an applied magnetic field ( for both the cases of a weak and a strong field , section iii ) ; and ( iii ) a qd with eight electrons at @xmath3 and a single impurity with varying strength , polarity , and location ( appendix a ) . for the case of a clean 2e qd , with the hamiltonian given by eqs.([mbh ] ) and ( [ hi ] ) with @xmath34 and @xmath35 , the exact solution can be found quite easily @xcite owing to the separability of the schrdinger equation in the center - of - mass and relative coordinates ; in the presence of impurities , separability is lost and finding an exact solution @xcite becomes significantly more complicated ( even for two electrons in the presence of a central impurity ) . therefore , and also in order to expound the physical principles underlying spontaneous symmetry breaking in qd s , we will mostly use in the following ( sections ii and iii and appendix a ) the self - consistent spin - and - space unrestricted hartree - fock ( ss - uhf ) method @xcite which , unlike the restricted hf ( rhf ) technique , @xcite allows for the formation of broken - symmetry states ( of lower energy than those obtained via the rhf ) .
this ss - uhf , which we introduced for studies of qd s in ref .
, employs @xmath14 ( where @xmath14 is the number of electrons ) orbital - dependent , effective ( mean - field ) potentials and it differs from the usual @xcite rhf in two ways : ( i ) it relaxes the double - occupancy requirement , namely , it employs different spatial orbitals for the two different ( i.e. , the up and down ) spin directions [ dods , thus the designation `` spin ( s ) unresricted '' ] , and ( ii ) it relaxes the requirement that the electron orbitals be constrained by the symmetry of the external confining field [ thus the designation `` space ( s ) unrestricted '' ] .
subsequent to our discussion of symmetry breaking and formation of electron molecules in the framework of the ss - uhf method , we elaborate in section iv on the connection between the symmetry - broken ss - uhf solution and the exact one in the case of a clean 2e qd with @xmath3 .
in particular , analysis of the exact solution using the conditional - probability - distribution ( cpd ) technique @xcite reveals the formation of an electron molecule in agreement with the ss - uhf result .
to introduce some of the principal physical and methodological issues pertaining to symmetry breaking and formation of wigner molecules in finite fermion systems , we discuss first the 2@xmath4 tqd at @xmath3 [ we remind the reader that a sufficiently high value of @xmath36 ( @xmath37 ) was chosen , such that the qd is in the regime where the two electrons form a wm ; for the other parameters used throughout this paper , see ref . ] .
the single - particle wave functions ( modulus square ) and total electron densities displayed in fig .
1 are taken from calculations for the 2@xmath4 tqd in the presence of two attractive impurities ( @xmath38 , represented in the figure by filled dots on the two sides of the qd ) located symmetrically about the center of the qd at @xmath39= @xmath40 nm with the strength and location of the impurities purposefully chosen such that they will not affect the nature of the electronic ground states ( for the same qd but without the impurities ) , except for orientational pinning in the case of symmetry - broken states ( see below ) .
constraining first the solution to maintain the symmetry of the hamiltonian , in conjunction with double - occupancy of the hf orbitals by electrons of opposite spins [ that is through the use of the rhf method with input trial electron densities satisfying the symmetry of the external potentials @xcite ] , the resulting symmetry - adapted ( sa ) self - consistent singlet ( @xmath41 ) orbitals and corresponding total density distributions exhibit , as expected , an almost circular symmetry with minimal elliptical distortions ( due to the impurities ) , see fig .
1(a ) ; without the impurities , the sa singlet is perfectly circularly symmetric .
however , increasing the variational freedom through removal of the spatial symmetry and double - occupancy constraints via the use of the ss - uhf method results in a symmetry - broken singlet ground state of lower energy , that is formation of a wm characterized by localized orbitals with the `` bond length '' ( distance between the maxima in the total electronic distribution ) equal to 29 nm [ see fig .
1(b ) ] ; the energy of this state is lower by 1.62 mev than that of the sa solution [ fig .
this lowering of the ground - state energy reflects gain in correlation energy ( for the definition of correlation energy , see section iv below ) .
note that the 2e wm is orientationally pinned along the interimpurity axis .
similar formation of a wm occurs also for a @xmath0 tqd without the pinning impurities ( with an energy gain of 1.32 mev compared to the corresponding sa solution ) .
however , in the absence of pinning , the formation of the wm is accompanied by orientational degeneracy @xcite ( that is there is an infinite manifold of rotationally degenerate ss - uhf ground states ) .
the formation of a fermionic molecule , associated with electron localization , in the ground - state of a qd under magnetic - field - free conditions does not have an analog within the framework of the traditional models of atomic structure .
@xcite however , the physics underlying this phenomenon , which is a manifestation of sb resulting from strong electronic correlations ( see also section iv below ) , resembles closely that found in doubly - excited two - electron atoms . indeed , spectroscopical studies on doubly - excited helium atoms revealed rovibrational bands which were interpreted , borrowing from models developed in the context of nuclear and molecular physics , by invoking the formation of a `` triatomic '' molecule comprised of the two localized electrons and the he@xmath5 nucleus ( @xmath42 particle ) , with the collinear configuration being of particular significance .
@xcite the first electronically excited state of the @xmath0 tqd is the triplet @xmath43 state ( with the spins of the two electrons parallel to each other ) whose total electron density distribution [ fig .
1(c ) , left ] resembles that of the ground - state singlet [ fig .
1(b ) , left ] .
however , the individual electronic wave functions in the @xmath43 state differ in character from those of the @xmath41 state [ compare right panels in fig .
1(c ) and 1(b ) ] , with the lower - energy one being @xmath44-like ( but elliptically distorted ) , and the other is a @xmath45-like orbital oriented by the impurities along the @xmath46-axis .
note that the @xmath43 state has the symmetry of the hamiltonian including the two pinning impurities ( that is , here the ss - uhf solution coincides with the symmetry - adapted one ) .
the same orbital characters are obtained also in the absence of the pinning impurities , but without a preferred orientation .
in this case , however , the character of the @xmath43 state is a result of spontaneous sb , with an energy gain of 0.09 mev compared to the corresponding sa ( circular ) solution .
underlying the type of spontaneous sb in the ( open - shell ) @xmath43 state of the @xmath0 tqd is the jahn - teller ( jt ) effect @xcite where lowering of the total energy is achieved via mixing of the two - fold degenerate @xmath47 ( @xmath48 ) amd @xmath49 ( @xmath50 ) angular momentum states , concomitant with a deformation of the self - consistent potential away from circular symmetry .
@xcite to distinguish such electron molecules from the wm discussed above for the closed - shell singlet state ( whose formation is driven by the dominance of the electron - electron repulsion ) , we refer to them as jt electron molecules ( jtem s ) .
such spontaneous sb via jt distortions is familiar from studies of the rotational spectra of open - shell nuclei @xcite and from investigations of shape deformations of open - shell metal clusters .
@xcite similar calculations for the @xmath0 tqd , but with repulsive pinning impurities ( that is @xmath51 ) yield for the singlet ground state qualitatively similar results ( with different values for the energies ) , but with an important distinction that now the pinned orientation of the wm is rotated by @xmath52 compared to the @xmath38 case ( i.e. , the `` intramolecular '' axis of the wm is oriented normal to the interimpurity axis ) . consequently , through variation of the sign ( polarity ) of the impurity gate voltages , one may `` flip '' the orientation of the wm , and with it the direction of the polarization of the electronic charge distribution in the qd . in this way , the formation of wm s in qd s and the ability to control their discrete orientations via pinning voltage gates may serve as a method for the creation of on - off information storage cells and nanoscale logic gates . @xcite
next , we examine the properties of a @xmath0 rutherford qd ( rqd ) , that is a @xmath0 qd with a central attractive impurity .
the ss - uhf singlet and triplet electronic orbitals corresponding to a @xmath0 rqd for @xmath3 with a single impurity ( @xmath53 ) located at ( 0,0,10 ) nm are shown in fig .
2(b ) and 2(c ) respectively .
they exhibit wm symmetry breaking and jt - distortion features similar to those found for the @xmath0 tqd [ compare fig .
1(b and c ) ] , but with a reduced wm bond length and a more compact triplet .
the `` strength '' of the sb depends of course on the impurity charge @xmath54 and/or its distance @xmath55 from the qd plane . for example , for @xmath56 and @xmath57 ( and for an arbitary position of the impurity inside the qd ) , no symmetry breaking was found by us due to the strong trapping by the impurity of the two electrons which occupy circularly symmetric orbitals , resembling the behavior of the ground - state of the natural he atom .
@xcite we also remark that for the case described in fig .
2(b ) and fig .
2(c ) both electrons are slightly trapped by the impurity potential for the @xmath41 state , while for the @xmath43 state the @xmath44-like electron is strongly trapped and the @xmath45-like electron occupies a much less bound orbital .
note that here , as with the case of a clean qd , the singlet wm and the jt - distorted triplet are free to rotate in the plane of the qd , since there are no off - centered pinning impurities .
the large physical size of qd s makes them ideally suited for investigations of magnetic - field effects and controlled manipulations .
to illustrate such effects , we display in fig .
2(a ) the magnetic - field induced variation of the total energies of the @xmath41 and @xmath43 states in the @xmath0 rqd ( qualitatively similar behavior is found also for the @xmath0 tqd ) .
as seen , the energy of the singlet state increases and that of the triplet state decreases with increasing @xmath31 ; for fields @xmath58 ( 0.2 t ) , the variation of the energy of the @xmath43 ( the slope of the curve ) is smaller than that for @xmath59 [ see inset in fig .
furthermore , at a critical value @xmath60 t [ marked by a down - arrow in fig .
2(a ) ] , the energies of the two states cross and from then on the triplet lies below the singlet .
for the singlet , the broken - symmetry wm state [ fig .
2(b ) ] maintains under the influence of the applied magnetic field in the range considered in fig .
2(a ) , with the increase in the magnetic - field strength leading to further shrinkage of the bond length of the wm accompanied by an overall increase of the energy of the @xmath41 state @xcite [ see fig .
2(a ) ] .
the influence of the magnetic field on the triplet state is more subtle . as aforementioned , at @xmath3 the symmetry of the @xmath43 state is broken by the jt effect involving mixing of the @xmath61 degenerate angular momentum orbitals ( see ref . ) . on the other hand ,
the magnetic field lifts the degeneracy of these @xmath48 and @xmath50 states ( without mixing them ) , and this effect competes with the jt distortion . for small enough fields [ @xmath58 ,
see inset in fig .
2(a ) ] , the jt effect prevails , and thus the orbitals and electron densities remain similar to those shown in fig .
2(c ) , and they maintain an orientational degeneracy in the plane .
at stronger fields ( @xmath59 ) , the lifting of the energetic degeneracy of the @xmath48 and @xmath50 states overcomes the jt effect , and the second electron populates the lower of these two orbitals . as a result ,
circular symmetry is recovered , as illustrated in fig .
the overall decrease with @xmath31 of the energy of the @xmath43 state relative to the @xmath41 state is due to enhanced stabilization by the ( parallel - spin ) exchange energy in the former , reduced coulomb repulsion between the electrons occupying @xmath44- and @xmath45-like orbitals , and quenching of the kinetic energy of the @xmath45-like orbital by the magnetic field .
@xcite this effect increases with @xmath31 , and at @xmath62 the @xmath43 state becomes the ground state [ see fig .
note that this transition is driven primarily by the interelectron repulsion and not by the interaction of the electrons moments with the magnetic field ( see ref . and references therein )
; for our system , the zeeman splitting energy is 0.026 mev / t . at even larger fields , the coulomb repulsion between the electrons increases ( due to the shrinking of the orbitals ) resulting in an ascending trend of the energy of the @xmath43 state , which remains , however , lower than the singlet state .
a similar scenario is found also for the tqd ( without pinning impurities ) . in light of previous findings
@xcite pertaining to formation of fully spin - polarized symmetry - broken states in clean qd s ( tqd s ) at high magnetic fields , it is pertinent to inquire whether the circular symmetry found for the @xmath43 state of the @xmath0 rqd for @xmath63 t will also be broken at higher fields .
for the clean @xmath0 tqd , we verified that indeed an orientationally degenerate electron molecule [ with the molecular orbitals of the electrons distributed about the two molecular centers , see fig .
2(e ) ] formed at sufficiently high @xmath31 ( e.g. , @xmath64 t ) .
such an electron molecule is akin @xcite to the jtem discussed above in the context of the triplet state for @xmath3 .
interestingly , such reemergence of a jtem structure does not occur at these conditions for the @xmath0 rqd studied here due to the enhanced gap between the @xmath48 orbital and the strongly trapped @xmath44 orbital .
this provides an additional venue for impurity - assisted manipulation and design of the electronic properties of qd s .
as mentioned earlier , for @xmath13 , the ss - uhf approach applied to qd s yields _ approximate _ ground - state solutions which violate the symmetries of the original many - body hamiltonian , e.g , the spontaneous breaking of rotational symmetry discussed in sections ii and iii for a circularly symmetric clean qd ( i.e. , a tqd ) or one with a central impurity ( i.e. , rqd ) . at a first glance , this situation may appear puzzling , but it is not unique in the context of many - body theory of finite fermionic systems . indeed , a similar situation was encountered in nuclear physics in the 1950 s , when it was discovered that open shell nuclei carried permanent quadrupole moments and that many of them exhibited well developed rotational spectra ( i.e. , they behaved like rigid rotors ) .
the explanation of these experimental findings was formulated in the framework of breaking of the rotational symmetry associated with nuclear deformations of the jahn - teller type , and it led to several celebrated models and semi - empirical methods , i.e. , the particle - plus - rotor model of bohr and mottelson , @xcite the modified - anisotropic - oscillator model of nilsson @xcite and the strutinsky shell - correction method .
@xcite in the language used by us here , this means that at the microscopic level the breaking of the rotational symmetry had to be accounted @xcite for via space ( s)-uhf methods ( that is allowing the spatial - orbital solutions of the hartree - fock equations to assume symmetries lower than those of the underlying many - body hamiltonian ) . starting with peierls and yoccoz @xcite ( see also peierls and thouless @xcite ) ,
numerous theoretical investigationss have addressed the connection of the broken - symmetry hf solution to the exact solution , and led to the theory of _ restoration - of - symmetry _ via projection techniques .
@xcite the central physical ingredient of the projection technique rests with the observation that a hf solution which breaks rotational symmetry is not unique , but belongs to an infinitely - degenerate manifold of states with different spatial orientation . a proper linear combination of the hf determinants in such a manifold yields multi - determinental states with good _ total _ angular momenta that are a better approximation to the exact solution . for a comprehensive review on restoration of symmetry in the context of nuclear many - body theory , we refer the reader to the book by ring and schuck mentioned in ref .
( see also ref .
, where the principles of restoration of rotational symmetry are discussed in the two - dimensional case )
. the restoration of symmetry via projection techniques in the case of ss - uhf solutions describing 2d - qd s will be presented in a future publication . in this section
, we found it more convenient to discuss the connection between the ss - uhf and the exact solutions by taking advantage of the simplicity of solving the exact problem at @xmath3 for two electrons interacting via the coulomb force and confined by an external parabolic confinement without impurities ( clean qd ) .
indeed , it is well known that in this case , the exact schrdinger equation for two interacting electrons is separable in the center - of - mass , @xmath65 ( with a corresponding mass @xmath66 ) , and relative , @xmath67 ( with a corresponding reduced mass @xmath68 ) , coordinates . to analyze the properties of the exact solutions , two quantities @xcite have customarily been extracted from the two - body wave function @xmath69 [ where the tilde denotes both spatial and spin variables , i.e. , @xmath70 @xmath71 , see appendix b ] : ( i ) the usual pair - correlation function , @xcite @xmath72 and ( ii ) the electron density ( ed ) @xmath73 however , for the exact @xmath74 in the case of a circularly symmetric confinement , both of these quantities turn out to be also circularly symmetric and thus they do not reveal the full physical picture , the ( possible ) formation of an electron molecule generated by electron correlations .
a more suitable quantity here is the conditional probability distribution ( cpd ) @xmath75 for finding one electron at @xmath76 given that the second electron is at @xmath77 .
this quantity has been extensively used @xcite in the analysis of electron correlations in doubly - excited helium - like atoms , and is defined as follows , @xmath78 [ for details concerning the calculation of the quantities in eqs.([pair ] ) @xmath79 ( [ cpd ] ) , see appendix b ] . in the left column of fig . 3
, we display the above three quantities for the exact ground state in the case of a parabolic confinement with the same parameters as used throughout this paper ( see ref . ) and @xmath3 . the pair - correlation function in fig .
3(a ) exhibits a well developed depression at @xmath80 , namely the two electrons on the average keep apart from each other at a distance @xmath81 nm . in fig .
3(b ) , we plot the cpd with @xmath82 ( marked by a cross ) .
it is seen that the maximum probability for finding the second electron occurs at the diametrically opposite point @xmath83 , in accordance with the picture of an electron molecule presented earlier within the ss - uhf approach . the exact electron density shown in fig .
3(c ) is circularly symmetric , as expected .
comparing figures 3(b ) and 3(c ) , the following interpretation ensues naturally , namely that the cpd in fig .
3(b ) describes the electron molecule in its _ intrinsic ( body - fixed ) _ frame of reference , while the electron density in fig .
3(c ) describes the electron molecule in the laboratory frame of reference where rotational and and center - of - mass displacements are superimposed upon the intrinsic probability density .
3(d ) displays the electron density for the corresponding ss - uhf ground state .
as discussed earlier , the ss - uhf electron density breaks the rotational symmetry and clearly exhibits the morphology of an electron molecule , unlike the exact one in fig .
it is apparent that the ss - uhf electron density corresponds to that in the intrinsic frame of the electron molecule .
restoration of the symmetry via projection techniques will bring the ss - uhf electron density closer to that of the exact solution .
as aforementioned , this interpretation is familiar in the context of nuclear and it is further supported by the cpd calculated with the ss - uhf ground - state [ i.e. , by using @xmath84 in eq .
( [ cpd ] ) , instead of the exact many - body @xmath74 , see appendix b ] and displayed in fig.3(e ) .
although naturally not identical , the two cpd s ( i.e. , the exact and the uhf ) are of similar nature and both illustrate graphically the correlation effect associated with electron localization and formation of an electron molecule .
we further illustrate this point by contrasting the exact and ss - uhf cpd s with the cpd of the symmetry - adapted rhf ground state , shown in fig .
3(f ) . in this latter case , @xmath85 , with @xmath86 being the 1@xmath44 orbital of the rhf ;
as a result , the probability of finding one electron at @xmath76 is independent of the position of the second electron and it is centered about the origin where it achieves its maximum value , as expected from an independent - particle description , i.e. , @xmath87 finally it is of interest to examine the energetic aspects of the symmetry breaking .
indeed the energy of the ground - state is @xmath88 mev for the rhf solution , @xmath89 mev for the ss - uhf solution , and @xmath90 mev for the exact one .
since the correlation energy is the difference between the rhf and the exact energies , @xcite one finds @xmath91 mev .
thus the symmetry breaking associated with the ss - uhf solution is able to capture @xmath92 44% of the correlation energy ; the remaining amount can be captured through improvements via projection techniques .
the dominance of interelectron interactions and correlations in quantum dots ( often referred to as `` artificial atoms '' ) results in spontaneous symmetry breaking of the self - consistent mean field , beyond the central - field picture on which the modern understanding of atomic structure has been founded . indeed , as shown in this paper through spin - and - space unrestricted hartree - fock @xcite many - body calculations for the singlet ground state of a 2e qd , such symmetry breaking and the resulting formation of wigner molecules do occur for a wide range of system parameters ( with or without impurities ) even in the most elementary case of two - electron qd s .
additionally , in the case of fully polarized dots ( e.g. , the triplet state of a 2e qd ) , the ss - uhf calculations show that similar electron molecules form due to symmetry breaking associated with jahn - teller - type distortions . in spite of the simplicity of the hamiltonian ,
correlations in two - electron qd s underlie a remarkably rich and complex physical behavior . as a result
, 2e qd s can serve as paradigms for a unification of concepts pertaining to spontaneous sb in various finite fermion aggregates , including nuclei , excited atoms , and clusters .
furthermore , impurity and/or magnetic - field effects allow for controlled manipulation of the formation and pinning of the discrete orientations of the electron molecules in 2e qd s . such ability may be employed in future applications of qd s as nanoscale logic cells and information storage elements .
impurity effects were also illustrated for the case of a quantum dot with more than two electrons .
confirmation of the formation in 2e qd s of electron molecules associated with symmetry breaking of the self - consistent field ( in the context of ss - uhf calculations ) was obtained through an analysis of the exact ground state via the conditional - probability - distribution technique .
this research is supported by the us d.o.e .
( grant no .
fg05 - 86er-45234 ) .
in the main body of this paper we limited ourselves to the case of 2e qd s , since focusing on this elementary case allowed us to better elucidate the intricate physical principles involved in the formation and impurity control of electron molecules in qd s , from both the perspective of the ss - uhf treatment and the exact solution . in this appendix , we will present an example of the many different structural possibilities that can arise when impurities are introduced in a qd with a larger number of electrons . in particular , we consider a qd with 8 electrons in the presence of a hydrogenic - like @xmath93 impurity of variable nominal charge @xmath94 placed at the center ( and in one case off - center ) of the qd ( for the other input parameters , we use same values as used throughout the text , see ref . ) .
since this appendix does not intend to present an exhaustive study of larger dots , but simply aims at presenting an illustrative example , we will consider only one spin configuration , i.e. , the ss - uhf solutions having 4 spin - up and 4 spin - down electrons .
however , @xmath95 will be allowed to take both positive ( repulsive ) and negative ( attractive ) values .
4(a ) displays the ss - uhf electron density for a repulsive central impurity [ located at ( 0,0,0 ) ] with @xmath96 .
it is seen that a wigner molecule consisting of a single ring with 8 electrons [ denoted as a ( 0,8 ) ring ] is formed in this case . for a slightly attractive central impurity with @xmath97 , however , a structural change takes place , namely one electron moves to the center of the dot [ see fig .
notice that this ( 1,7 ) structure agrees with the arrangement found in studies of classical point - charges in a purely @xmath98 harmonic confinement .
@xcite increasing the attractive nominal charge to @xmath99 [ see fig . 4(c ) ] results in a further structural change , namely the central impurity is now able to trap two electrons , thus leading to a ( 2,6 ) arrangement
. a further increase of the attractive charge of the impurity to the value @xmath100 does not produce any qualitative change in the ( 2,6 ) arrangement as long as the impurity remains at the center of the dot . by moving the impurity to an off - center position , however ,
various structural morphologies can arise , an example of which is presented in fig .
4(d ) for a @xmath100 impurity located at ( 20,0,0 ) nm , forming a highly distorted ( 2,6 ) wm ( here the first index denotes that two electrons are trapped by the impurity ) .
a further increase in the attractive charge @xmath54 leads to sequential trapping of the remaining six electrons and to progressive elimination of symmetry breaking , until all eight electrons have been captured by the impurity ( see also section iii ) .
in the case of two interacting electrons confined by a parabolic potential ( clean qd ) , one can perform a change of variables to center - of - mass ( cm ) , @xmath101 and @xmath102 , and relative - motion ( rm ) , @xmath103 and @xmath104 , coordinates .
then the two - electron hamiltonian separates into cm and rm contributions , @xmath105 with @xmath106 and @xmath107 where @xmath66 , @xmath108 , @xmath109 , and @xmath110 .
the center - of - mass motion associated with the coordinate @xmath111 obeys a schrdinger equation describing the motion of a particle of mass @xmath66 in a 2d isotropic harmonic potential of frequency @xmath7 . here
@xmath7 is the frequency of the original parabolic confinement , i.e. , the interelectron repulsion has no baring on the center - of - mass motion .
using dimensionless polar coordinates @xmath112 and @xmath113 , the center - of - mass wave function can be written as @xmath114 with the radial part given by @xmath115 where @xmath116 are the radial and azimuthal ( related to the angular momentum ) quantum numbers , respectively , @xmath117 , the normalization constant @xmath118^{1/2}$ ] , and @xmath119 are associated laguerre polynomials . since in the exact problem
the coulomb interaction preserves the rotational symmetry , the radial part of the wave function @xmath120 associated with the relative motion obeys the following one - dimensional schrdinger equation [ in dimensionless polar coordinates @xmath121 and @xmath122 , @xmath123 the @xmath124 term results from the interelectron repulsion . defining @xmath125 ( where @xmath126 is the spin of the @xmath127th electron and @xmath128 ) , the exact many - body ( here two - body ) wave function
is given by @xmath129 with @xmath130 and @xmath131 is the spin part .
the exact eigenvalues are given by @xmath132 where @xmath133 are the eigenvalues associated with the relative motion [ see eq .
( [ relmot ] ) ] , @xmath134 being the corresponding radial and azimuthal quantum numbers . [ [ pair - correlation - electron - density - and - conditional - probability - distribution - for - the - exact - and - ss - uhf - two - electron - wave - functions ] ] pair correlation , electron density , and conditional probability distribution for the exact and ss - uhf two - electron wave functions ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ the bracket notation in eqs .
( [ pair ] ) , ( [ spdens ] ) , and ( [ cpd ] ) imply integrations over both the spatial and spin variables . in the case of the exact wave function @xmath74 given by eq .
( [ exwf ] ) , the spin variables separate out , and thus @xmath135 , @xmath136 , and the cpd can be expressed as double integrals over the positions @xmath137 of the following two - body spatial probability density @xmath138 , where @xmath139 is the spatial part of @xmath74 .
in particular , one finds @xmath140 for the pair - correlation function , @xmath141 for the electron density , and @xmath142 for the conditional probability distribution . in the case of the ss - uhf singlet ground state ,
the two electrons occupy two different spatial orbitals @xmath143 and @xmath144 .
then the corresponding two - body wave function is the following single determinant , @xmath145~ , \label{uhfwf}\end{aligned}\ ] ] where @xmath42 and @xmath146 denote the spin - up and spin - down spinors , respectively .
integration of @xmath147 over the spin variables yields the following two - body spatial probability density , @xmath148 m.a .
kastner , phys . today *
46(1 ) * , 24 ( 1993 ) . r.c .
ashoori , nature * 379 * , 413 ( 1996 ) .
s. tarucha , d.g .
austing , t. honda , r.j .
van der hage , and l.p .
kouwenhoven , phys .
lett . * 77 * , 3613 ( 1996 ) .
e.u . condon and g.h .
shortley , _ the theory of atomic spectra _ ( cambridge univ . press , london , 1935 ) .
d.r . hartree , proc .
* 24 * , 89 ( 1928 ) .
the failure of bohr - type models , based on the orbiting of spatially correlated electrons in conjunction with the bohr - sommerfeld quantization rule , to yield a reasonable estimate of the ground state of the he atom signaled a looming crisis in physics in the 1920 s , which bohr himself , as well as others , had been keenly aware of , as summarized succintly by sommerfeld : `` all attempts made hitherto to solve the problem of the neutral helium atom have proved to be unsuccessful '' [ a. sommerfeld , rev .
* 7 , * 509 ( 1923 ) ] ; see also the 10th chapter entitled `` it was the spring of hope , it was the winter of despair '' in a. pais , _ niels - bohr s times _ ( clarendon , oxford , 1991 ) , the review by j.h .
van vleck , phil . mag . * 44 * , 842 ( 1922 ) , and m. born , _ the mechanics of the atom _ ( ungar , new york , 1927 ) . while , since , numerical solutions of the two - electron schrdinger equation provided a quantitative resolution to the problem , the first successful semiclassical treatment of the three - body coulomb system awaited till 1980 , see j.g .
leopold and i.c .
percival , j. phys .
b * 13 , * 1037 ( 1980 ) ; see also d. wintgen , k. richter and g. tanner , chaos * 2 * , 19 ( 1992 ) .
furthermore , based on rather general group - theoretical arguments arising from the observation of hierarchies with lower symmetry in the excited spectra , and motivated by ideas originating in nuclear - physics spectroscopy , it has been discovered in the late 1970 s and early 1980 s that electron correlations in doubly excited he lead to quantization of the spectrum much like in a linear triatomic molecule , @xmath4-he@xmath5-@xmath4 . this molecular picture , with near rigidity and separability , results in `` infinite sequences of vibrational levels , on each of which
is built an infinite sequence of rotational levels '' , see m.e .
kellman and d.r .
herrick , phys .
a * 22 * , 1536 ( 1980 ) ; see also g.s .
ezra and r.s .
berry , phys .
a * 28 * , 1974 ( 1983 ) , and s. watanabe and c.d .
lin , phys .
a * 34 , * 823 ( 1986 ) .
interestingly , as in the aforementioned semiclassical treatments , the collinear configuration plays a special role in the molecule - like model , serving perhaps as `` partial vindication '' of the geometry considered originally by bohr .
. bohr and b.r .
mottelson , _ nuclear structure _
( benjamin , reading , ma , 1975 ) .
w. nazarewicz , nucl .
a * 574 * , 27c ( 1994 ) . c. yannouleas and u. landman , phys .
b * 51 * , 1902 ( 1995 ) .
w. ekardt , _ metal clusters _ ch . 1 & ch . 4 ( wiley , new york , 1999 ) .
in modeling of practical `` on - off '' ( i.e. , two - state ) devices , one will undoubtedly have to take into account additional potential terms in the many - body hamiltonian , such as those associated with image charges resulting from interfaces between different dielectric materials .
the feasibility of constructing nanoscale logic cells via controlled manipulation of the antipodal arrangements of two electrons in a larger device made out of four dots arranged in a square to form a quantum - dot supermolecule is currently under intense experimental investigation , see i. amlani , a.o .
orlov , g. toth , g.h .
bernstein , c.s .
lent , and g.l .
snider , science * 284 * , 289 ( 1999 ) .
e. wigner , phys . rev . * 46 * , 1002 ( 1934 ) . p.a .
maksym , phys .
b * 53 * , 10871 ( 1996 ) . h .- m .
m " uller and s.e .
koonin , phys .
b * 54 * , 14532 ( 1996 ) .
r. egger , w. h " ausler , c.h .
mak , and h. grabert , phys .
82 * , 3320 ( 1999 ) . c. yannouleas and u. landman , phys .
* 82 * , 5325 ( 1999 ) .
anderson , _ basic notions of condensed matter physics _ ( addison - wesley , reading , ma , 1984 ) .
thouless , _ the quantum mechanics of many - body systems _ ( academic press , new york , 1972 ) .
h. fukutome , int .
20 * , 955 ( 1981 ) . for studies pertaining to the arrangement of classical point - charges in a harmonic confinement ,
lozovik , usp .
nauk * 153 * , 356 ( 1987 ) [ sov .
usp . * 30 * , 912 ( 1987 ) ] and v.m .
bedanov and f.m .
peeters , phys .
b * 49 * , 2667 ( 1994 ) . that @xmath36 is the only relevant parameter in the formation of a wm in a clean qd can be seen immediately from the exact schrdinger equation describing the relative motion of two interacting electrons at @xmath3 in a parabolic confinement ( appendix b )
indeed , this equation when expressed in dimensionless variables as in eq .
( [ relmot ] ) contains @xmath36 as the sole free parameter .
the corresponding center - of - mass motion is independent of the interelectron repulsion . for how to extract information about the wigner molecule from the exact two - body solution ,
see section iv .
b. tanatar and d.m .
ceperley , phys .
b * 39 * , 5005 ( 1989 ) . in all calculations
, we used @xmath20(gaas)@xmath151 , @xmath12(gaas)@xmath152 , and @xmath153 mev , with @xmath33 ( that is @xmath154 ) ; the same value of @xmath36 is obtained by using @xmath10(gaas)@xmath155 and @xmath156 mev , yielding the similar results given here , but with different absolute energy values .
m. wagner , u. merkt , and a.v .
chaplik , phys .
b * 45 * , 1951 ( 1992 ) .
d. pfannkuche , v. gudmundsson , and p.a .
maksym , phys .
b * 47 * , 2244 ( 1993 ) .
m. taut , j. phys .
a * 27 * , 1045 ( 1994 ) .
m. taut , phys .
a * 48 * , 3561 ( 1993 ) .
m. rontani , f. rossi , f. manghi , and e. molinari , phys . rev .
b * 59 * , 10 165 ( 1999 ) . here
we differentiate between the exact solution of the hamiltonian specified by eqs .
( [ mbh ] ) and ( [ hi ] ) from `` exact diagonalization methods '' of hubbard - type hamiltonians used as models of qd s [ see e.g. , r. berkovits , phys .
* 81 * , 2128 ( 1998 ) ] .
the ss - uhf equations were solved using the pople - nesbet - roothaan formalism , see p. 214 in ref . ; see also j.a .
pople and r.k .
nesbet , j. chem . phys . * 22 * , 571 ( 1954 ) . apart from ref .
, the only previous use of a space - uhf method in qd s was in an investigation ( see ref . ) of wm s but only for fully polarized qd s ( i.e. , under high magnetic fields where the spin unrestriction is not at play ) .
a. szabo and n.s .
ostlund , _ modern quantum chemistry _
( mcgraw - hill , new york , 1989 ) .
gross , e. runge , and o. heinonen , _ many - particle theory _
( adam hilger , bristol , 1991 ) ch . 8 . from the point of view of the terminology adopted here , the simple designation hartree - fock ( hf ) in the literature most often refers to the rhf , in particular in atomic physics and the physics of the homogeneous electron gas . in nuclear physics
, however , the simple designation hf most often refers to a space ( s)-uhf .
see the article by ezra and berry in ref . ; see also r.s .
berry in _ structure and dynamics of atoms and molecules : conceptual trends _ , edited by j.l .
calais and e.s .
kryachko ( kluwer academic , dordrecht , 1995 ) p. 155 , and references therein .
the orientational degeneracy of the ss - uhf solutions describing an electron molecule in a cirularly symmetric qd is a residue of the rotational invariance of the exact hamiltonian and naturally it suggests the possible emergence of a ro - vibrational spectrum associated with the excitations of the electron molecule .
depending on the value of the parameter @xmath36 , the electron molecule may exhibit different degrees of `` rigidity '' , with electron molecules of high rigidity ( forming for large @xmath36 ) exhibiting well - developed ro - vibrational bands ; for the value of @xmath33 used by us here , the electron molecule is rather `` floppy '' [ c. yannouleas and u. landman , cond - mat/0003245 ] .
such ro - vibrational spectra can be described by several theoretical approaches , for example : ( i ) at a more phenomenological level , by a quantization of the collective classical motions associated with the molecule , such as rotations , and bending and stretching vibrations [ see kellman and herrick in ref . and
maksym in ref . ] ; ( ii ) at a microscopic level , by projecting the symmetry - broken hf determinant onto multi - determinental wave functions with good angular momenta , thus restoring the symmetry of the exact hamiltonian ; see p. ring , and p. schuck , _ the nuclear many - body problem _ , ch . 11 ( springer - verlag , new york , 1980 ) and p .- o .
l " owdin , in _ quantum theory of atoms , molecules , and the solid state _ , edited by p .- o .
l " owdin , pp .
601 - 623 ( academic press , new york , 1966 ) . furthermore
, starting from an exact solution for the electrons in the qd , rovibrational characteristics can be obtained from the many - body eigenvalue spectra evaluated for different total angular momenta and spins in conjunction with an analysis of the conditional - probability densities [ c. yannouleas and u. landman , cond - mat/0003245 ] .
we note that the _ discrete _ ro - vibrational collective spectra associated with symmetry - breaking in a qd may be viewed as finite analogs to the goldstone modes accompanying symmetry breaking transitions in extended media ( see ref . ) .
jahn and e. teller , proc .
london ser .
a * 161 * , 220 ( 1937 ) .
recall that for an electron moving in a 2d circularly symmetric potential the wave functions are given ( in polar coordinates ) by @xmath157 with corresponding energies @xmath158 , where @xmath159 is the number of radial nodes and @xmath160 is the angular momentum quantum number , with @xmath161 in the jt framework , the second electron of the ( open - shell ) triplet state of the @xmath0 qd occupies an orbital formed as a linear combination of the @xmath162 , @xmath61 states ( denoted by @xmath48 and @xmath50 ) , i.e. , @xmath163 or @xmath164 , with the one of lower energy being selected through the self - consistent distortion of the central mean field . as a consequence , the orbitals of both electrons in the @xmath0 qd exhibit noncircular symmetry .
the energy gain due to the jt distortion depends on the strength of the confinement potential , and it increases for weaker confinements . in natural atoms ,
the spontaneous jt effect is fairly small and perturbative methods are usually adequate ; see ch .
iii.2 in ref . .
this cancellation of crystallization due to a strong - attractive - impurity effect is illustrated in fig .
4(d ) where the attractive impurity is able to capture two of the eight electrons in the dot .
the two captured electrons form a singlet subunit and their orbitals are clearly circularly symmetric ( see appendix a below ) .
the effective confinement frequency @xmath165 in the darwin - fock model of a 2d harmonically - confined electron in the presence of a magnetic field @xmath31 ( where @xmath7 is the frequency of the harmonic potential and @xmath166 is the cyclotron frequency ) increases as a function of the magnetic field . in a magnetic field ,
the energy of a harmonically confined electron is given in the darwin - fock model by @xmath167 .
such jt - like electron molecules have been called wigner molecules in previous publications treating fully polarized qd s ( refs . and ) . however
, this naming is not appropriate since their formation is not driven by the dominance of the interelectron repulsion ( over the zero - point energy ) , which for large magnetic fields can be treated perturbatively ( refs . and ) .
. bohr and b.r .
mottelson , mat .
* 27 * ( no .
16 ) ( 1953 ) .
nilsson , mat .
* 29 * ( no .
16 ) ( 1955 ) .
strutinsky , nucl . phys . *
a95 * , 420 ( 1967 ) .
for an early review of the application of hartree - fock methods to the description of open - shell , deformed nuclei , see g. ripka , adv .
. phys . * 1 * , 183 ( 1968 ) .
peierls and j. yoccoz , proc .
( london ) * a70 * , 381 ( 1957 ) .
peierls and d.j .
thouless , nucl . phys . * 38 * , 154 ( 1962 ) .
the concept of restoration of symmetry is not restricted to spatial symmetry , but includes also restoration of spin symmetry ( see the paper by lwdin in ref . .
brown , _ unified theory of models and forces _
( north - holland , amsterdam , 1967 ) p. 64 .
for the purely parabolic confinement considered here , it turns out that the pair - correlation function reduces @xcite to the modulus square of the radial part of the wave function ( see appendix b ) describing the relative motion , i.e. , @xmath168 . see p. 620
of the article by lwdin mentioned in ref .
( in this article the rhf is referred to as _ conventional hartree - fock scheme _ ) ; this definition was also followed in ref . , where a rhf solution was contrasted to the exact ground - state solution for two interacting electrons . c. yannouleas and u. landman , to be published . | interelectron interactions and correlations in quantum dots can lead to spontaneous symmetry breaking of the self - consistent mean field resulting in formation of wigner molecules . with the use of spin - and - space unrestricted hartree - fock ( ss - uhf ) calculations ,
such symmetry breaking is discussed for field - free conditions , as well as under the influence of an external magnetic field . using as paradigms impurity - doped ( as well as the limiting case of clean ) two - electron quantum dots ( which are analogs to helium - like atoms )
, it is shown that the interplay between the interelectron repulsion and the electronic zero - point kinetic energy leads , for a broad range of impurity parameters , to formation of a singlet ground - state electron molecule , reminiscent of the molecular picture of doubly - excited helium .
comparative analysis of the conditional probability distributions for the ss - uhf and the exact solutions for the ground state of two interacting electrons in a clean parabolic quantum dot reveals that both of them describe formation of an electron molecule with similar characteristics .
the self - consistent field associated with the triplet excited state of the two - electron quantum dot ( clean as well as impurity - doped ) exhibits symmetry breaking of the jahn - teller type , similar to that underlying formation of nonspherical open - shell nuclei and metal clusters .
furthermore , impurity and/or magnetic - field effects can be used to achieve controlled manipulation of the formation and pinning of the discrete orientations of the wigner molecules .
impurity effects are futher illustrated for the case of a quantum dot with more than two electrons .
formation and control of electron molecules in artificial atoms : + impurity and magnetic - field effects + |
the pairing mechanism of a superconductor determines the symmetry of its order parameter , which in turn is connected to the symmetry of the superconducting gap .
superconductors , whose averaged order parameter over the entire fermi surface yields zero , are called unconventional . they attract much interest @xcite in modern condensed matter physics because a description of the underlying physics has to go beyond the standard bardeen - cooper - schrieffer ( bcs ) theory for an @xmath3-wave order parameter .
their gap contains nodes , whose existence and position can be detected by a variety of experimental probes . often , only a combination of results from different probes allows one to draw a conclusive picture .
the search and study of unconventional superconductors was triggered by the discovery of heavy - fermion superconductivity @xcite , the high-@xmath7 cuprates @xcite , and organic superconductors @xcite . by now , a lot more materials are believed to be unconventional superconductors . in this context , the filled - skutterudite compounds @xmath8 ( @xmath9 electropositive metal , @xmath10 transition metal , and @xmath11 usually a pnictogen ) attracted much attention with the discovery of pros@xmath0sb@xmath1 , which is the first pr - based heavy - fermion superconductor and believed to be of unconventional type @xcite .
it exhibits exotic properties probably connected to the quadrupole degrees of freedom @xcite .
the skutterudite family @xmath12 ( @xmath13 sr , ba , la , pr , th ) with a pt - ge framework are also superconductors @xcite .
the two members prpt@xmath6ge@xmath1 and lapt@xmath0ge@xmath1 show superconductivity at relatively high transition temperatures compared to other skutterudites , namely @xmath14 and @xmath15 @xcite .
prpt@xmath6ge@xmath1 seems to be a good candidate for unconventional superconductivity .
it is considered to be a moderately strong - coupling superconductor from the large specific - heat jump compared to the bcs value @xcite .
there are indications of point nodes from nmr @xcite , specific heat , and penetration depth @xcite .
furthermore , @xmath16sr measurements detected a time - reversal symmetry breaking below @xmath7 @xcite .
additionally , a number of investigations including photoemission , magnetization , penetration depth , and specific heat revealed multiband superconductivity in this compound @xcite . the continuous evolution of @xmath7 across the doping series ( pr@xmath4la@xmath17)pt@xmath0ge@xmath1 suggests compatible order parameters of the end members prpt@xmath6ge@xmath1 and lapt@xmath0ge@xmath1 @xcite . however , the few existing investigations on lapt@xmath0ge@xmath1 point towards a single isotropic gap : the specific - heat jump @xmath18 is only slightly above the bcs value suggesting a weaker coupling than in prpt@xmath6ge@xmath1 @xcite
. nmr and photoelectron - spectroscopy results for lapt@xmath0ge@xmath1 could be best explained by a single isotropic gap @xcite .
additionally , no time - reversal symmetry breaking is observed in lapt@xmath0ge@xmath1 by @xmath16sr @xcite .
only one recent study using @xmath16sr and tunnel diode spectroscopy for penetration - depth measurements reports indications of multiband superconductivity in lapt@xmath0ge@xmath1 @xcite .
in order to develop a conclusive picture , it is necessary to shed more light on the superconducting properties of lapt@xmath0ge@xmath1 . in particular
, it is important to clarify , if lapt@xmath0ge@xmath1 is a multiband superconductor . to this end , we performed specific - heat measurements at temperatures down to 0.4k for different magnetic fields . while the temperature dependence at zero field can be described with a single superconducting @xmath3-wave gap , the field dependence of the specific - heat @xmath19-coefficient shows sub - linear behavior , which is discussed in terms of fermi - surface anisotropies .
additionally , we use the specific - heat results to analyze the data of our detailed thermal - conductivity study with a focus on the behavior below 1k .
our temperature- and field - dependent measurements are compatible with a single - gap @xmath3-wave state .
for the synthesis of single crystals of lapt@xmath0ge@xmath1 , a bulk sample of mass 2 g was placed in a glassy - carbon crucible , sealed in a ta tube and enclosed in a quartz ampoule .
the ampoule was heated up to 870@xmath20c within 5h and kept at this temperature for 10h .
then it was cooled down to 845@xmath20c and held for 30 days and furnace - cooled .
single crystals of 1 - 4 mm size were mechanically extracted from the sample @xcite .
we investigated three high - quality single crystals of lapt@xmath0ge@xmath1 selected from the same batch with a residual resistivity ratio ( rrr ) of 17 and a residual resistivity @xmath21 of @xmath22 .
crystal # 1 with a mass of @xmath23 mg and crystal # 2 with 67.2 mg were used for specific - heat measurements by a relaxation method ( heat capacity option in a ppms by quantum design ) . on crystal # 3 we performed transport measurements with a standard two - thermometer - one - heater technique .
for this purpose , we cut it first into a rectangular bar along the principal cubic crystal axes with a cross - sectional area of 0.492mm@xmath24 and a contact distance of 1.08 mm .
this arrangement was used for thermal - conductivity measurements in a ppms ( thermal transport option ) above 2k and in zero magnetic field .
afterwards , a long plate was cut from crystal # 3 again along the principal cubic crystal axes with the dimensions @xmath25 .
this sample # 3a was used to measure thermal conductivity both as a function of temperature and isothermally as function of magnetic field for temperatures below 1k and for magnetic fields of @xmath26 t .
the heat current @xmath27 was applied along the same direction as for the ppms measurement .
a superconducting split - coil magnet generated a magnetic field @xmath28 .
is shown for two different crystals both at zero field and above the upper critical field @xmath30 .
a fit of the results at 2 t with eq .
[ eqn : lapt4ge12_debye_spec_heat ] falls on top of the data , except below 4k , where the phonon contribution follows @xmath31 ( see inset ) .
( b ) the electronic contribution @xmath32 to the zero - field curve follows a bcs behavior with @xmath33 ( solid line ) .
both samples show the same results within the measurement uncertainty .
( c ) and ( d ) numerical evaluation of the two gap @xmath34-model .
( c ) the parameter @xmath35 of the model was determined from the specific - heat jump at @xmath36 for the displayed parameter range of @xmath37 and @xmath38 .
( d ) the variance compared to the data in ( b ) , @xmath39 , reveals a large parameter region around @xmath40 ( blue ) , where the @xmath34-model is able to describe the data equally well ( note the logarithmic scale ) .
the electronic specific heat from the two example parameter sets 1 ( @xmath41 , @xmath42 , @xmath43 ) , and 2 ( @xmath44 , @xmath45 , @xmath46 ) is shown as dashed and dotted curves in ( b ) .
curve 1 falls almost on top of that of the single - band model . ]
the specific heat @xmath47 of lapt@xmath0ge@xmath1 was measured at zero field and in a field of 2 t , which is above the superconducting critical field ( fig .
[ fig : c - vs - t](a ) ) .
our measurements are in good agreement with previous results on polycrystalline samples @xcite .
the data at 2 t , @xmath48 , are used as an estimate of the phonon contribution . to obtain the electronic contribution @xmath49 to the specific heat at zero field ,
the results at 2 t are subtracted from the zero - field data except for a normal - state electronic contribution @xmath50 , which is determined by a fit to @xmath48 between @xmath51 using the sum of a debye and einstein model , @xmath52 where @xmath53 describes the einstein specific heat from a single phonon mode . the fit result is also plotted in fig . [ fig : c - vs - t](a ) . from the fit we obtain @xmath54 , @xmath55 ( debye temperature @xmath56k ) , and an einstein temperature of @xmath57k .
the fit describes the data in the whole temperature range reasonably well , only below 4k there are small deviations and the phonon contribution follows @xmath58 .
this deviation is most likely due to the complicated phonon spectrum generally observed in filled skutterudites caused by a combination of modes from the filler atom la and modes from the cage structure formed by pt and ge @xcite .
the addition of a @xmath59-term is not able to improve the fit .
the deviations of the data from the model in eq .
[ eqn : lapt4ge12_debye_spec_heat ] , however , can be neglected for the following analysis of the specific heat , but it becomes important when we discuss the thermal conductivity .
the result for the electronic contribution to @xmath60 is shown in fig .
[ fig : c - vs - t](b ) as the specific heat coefficient @xmath61 .
we note that the feature in @xmath62 below 1k is an artefact due to the subtraction of the lattice contribution to the specific heat .
the specific heat exhibits a sharp superconducting transition at 8.0k with a width of 0.13k , which is an indication of the good quality of our crystals .
the jump height is larger than expected from the predictions of the weak - coupling bcs model ( @xmath63 ) , which can not reproduce our data .
however , we are able to describe @xmath64 adjusting the gap to @xmath33 .
this value is in resonable agreement with the results from photoelectron spectroscopy : @xmath65 @xcite , and nmr : @xmath66 @xcite .
since superconductivity in other skutterudites is discussed in terms of multiband superconductivity @xcite and there are results in favor of this interpretation for lapt@xmath0ge@xmath1 as well @xcite , we also consider this possibility based on our specific - heat data . in the case of two gaps ,
the specific heat might be described by a weighted sum of the contributions from both gaps within the standard two - gap @xmath34-model @xcite @xmath67 both @xmath68 and @xmath69 are calculated within the bcs theory , but with variable gap sizes @xmath37 and @xmath38 .
the third free parameter of this model is the weighting factor @xmath35 . since our data can already be described with a single gap @xmath70 , one of the two gaps considered for the @xmath34-model must be greater , the other one smaller than @xmath70 .
we therefore evaluate the @xmath34-model numerically in the parameter range @xmath71 , @xmath72 using the data between @xmath73 . using eq .
[ eqn : alpha ] , we determined the corresponding @xmath35 from the height of the specific heat jump at @xmath36 , which we obtained from the single band calculation with @xmath33 .
the results for @xmath35 are shown in fig .
[ fig : c - vs - t](c ) . in order to find all parameter sets @xmath74 , which are able to describe our data
, we evaluate the variance @xmath75 which we plot in fig .
[ fig : c - vs - t](d ) on a logarithmic scale . in this plot
, we can identify the single band case in the corner where @xmath40 .
it also extends vertically and horizontally from this point , where we find either @xmath76 or @xmath77 , respectively . however , there is also a large region away from the single - band case , which has the same low variance .
as an example , the electronic specific heat calculated from parameter set 1 ( see fig . [
fig : c - vs - t](b ) ) lies almost on top of the single - band calculation .
we obtain such a shape and large size of this low - variance region , because a single band is already capable to describe our data . away from the low variance region ,
the results from the @xmath34-model start to deviate from our data , which is demonstrated by the curve from parameter set 2 in fig .
[ fig : c - vs - t](b ) . in summary ,
our results from the analysis with the @xmath34-model reveal two possible scenarios : ( a ) two gaps close to @xmath70 or ( b ) @xmath78 and @xmath79 as long as @xmath80 stays small .
the very good agreement with a single @xmath3-wave gap excludes the possibility to describe @xmath81 of lapt@xmath0ge@xmath1 using a nodal gap structure almost completely .
the question whether there are multiple @xmath3-wave gaps with different gap sizes involved may ultimately be solved by measurements on the substitution series ( la@xmath17pr@xmath4)pt@xmath0ge@xmath1 .
normalized to its normal state value @xmath83 deviates from a linear behavior ( dashed line ) and follows a power law ( solid line ) with an exponent of 0.72 .
@xmath84 was extracted from a fit to @xmath85 at different fields .
a selection of these data is shown in the inset as function of @xmath86 for the equally spaced fields @xmath87 , 0.2 t , ... , 1.4t@xmath88 .
the lines are a guides to the eye . ] the magnetic - field dependence of the @xmath19-coefficient in the superconducting state was determined by an extrapolation of @xmath89 data to zero temperature assuming an exponential behavior at low temperatures in the superconducting state .
figure [ fig : c - vs - h ] clearly shows , that @xmath84 normalized to its normal state value @xmath90 follows a sublinear curve @xmath91 with an exponent of @xmath92 between @xmath93 .
@xmath94 t has been determined from thermal conductivity , see below .
this exponent lies in between expectations for nodal superconductors in simplified models ( @xmath95 ) @xcite and a fully gapped isotropic superconductor ( @xmath96 ) @xcite .
however , these predictions take only certain quasiparticle contributions into account : for an @xmath3-wave superconductor it is the contribution to the density of states from the states within the vortex cores ; for nodal superconductors it is a contribution from the delocalized quasiparticles around the nodes .
deviations are , therefore , expected , if one takes further contributions into account .
in addition to core states and delocalized quasiparticles , respectively , this would be vortex lattice contributions @xcite .
since the temperature dependence of the specific heat does not indicate the presence of nodes in the superconducting gap function , we will focus here on the case of fully gaped superconductors .
a deviation of @xmath97 from the predicted linear dependence was observed in several @xmath3-wave superconductors @xcite .
several possible reasons have been suggested : ( 1 ) multiband superconductivity ; ( 2 ) anisotropic fermi surface ; ( 3 ) non - linear contributions from the vortex lattice , which arise due to a field - dependent vortex - core radius @xcite .
we first discuss case ( 1 ) . in the multiband superconductor mgb@xmath98 , two distinct field scales were identified @xcite , which correspond to the two different upper - critical fields of the two gaps .
interestingly , the field dependence of @xmath99 tracks the field dependence of the thermal conductivity @xmath100 with the same characteristic field scales .
this is not the case in lapt@xmath0ge@xmath1 , as @xmath100 shows a curvature with opposite sign ( see below ) and makes , therefore , a scenario with two gaps of very different size , as in prpt@xmath6ge@xmath1 , rather unlikely .
lapt@xmath0ge@xmath1 has a complex , anisotropic fermi surface with several bands crossing the fermi level @xcite .
this might explain the observed sublinear behavior in @xmath97 .
a similar scenario has been suggested for the borocarbide superconductor luni@xmath98b@xmath98c @xcite .
we can not exclude that a field dependent vortex core radius provides a further mechanism for the observed deviations , however , so far there are no indications for such a scenario in lapt@xmath0ge@xmath1 . .
( a ) in the double - logarithmic representation up to 100k , a clear but smooth drop of @xmath101 at @xmath36 is visible .
a hump appears at lower @xmath29 .
inset : @xmath101 on a linear scale . for an @xmath3-wave superconductor ,
one expects a rapidly decreasing @xmath102 ( e ) below @xmath36 , while @xmath103 ( p ) increases forming a hump .
the sketched @xmath29-dependencies for @xmath102 and @xmath103 are inspired by the theory of j. bardeen , g. rickayzen , and l. tewordt ( brt ) for @xmath3-wave superconductors .
the dashed lines are guides to the eye .
( b ) the low - temperature thermal conductivity can be fitted with a power law @xmath104 ( solid yellow ) leading to a residual term of @xmath105 . a second fit ( dashed red ) with eq .
[ eqn : lapt4ge12_kinetic_phonon_kappa ] using the experimental specific - heat results indicates a dominant phonon contribution scattered by sample boundaries . ]
figure [ fig : kappa - vs - t](a ) presents the thermal conductivity @xmath106 at zero field for temperatures up to 100k .
figure [ fig : kappa - vs - t](b ) shows the section below about 0.5k as a function of @xmath107 , which we will discuss first in more detail .
the data follow a power law , which we fit with @xmath108 below @xmath109 . from the fit
, we obtain a residual term @xmath110 ( @xmath111 ) , which is of the order of our measurement uncertainty .
a sizable residual term is expected for a nodal superconductor @xcite due to pair - breaking impurities .
our observed @xmath112 is small compared to these expectations and experimental results for unconventional superconductors , _
e.g. _ the @xmath113-wave superconductors tl@xmath98ba@xmath98cuo@xmath114 with @xmath115 @xcite , and ceirin@xmath116 with @xmath117 @xcite .
hence , this result points towards a superconductor with a finite gap everywhere on the fermi surface .
the exponent we obtain from the power law fit is @xmath118 . in an @xmath3-wave superconductor ,
the thermal conductivity at sufficiently low @xmath119 is expected to be entirely due to phonons . with the relation @xmath120
one can estimate the phonon contribution @xmath121 from the specific heat @xmath122 and the mean free path @xmath123 .
@xmath124 is the phonon specific heat from a debye model determined above .
the sound velocity @xmath125 can be calculated from @xmath122 using the two equations @xmath126 where @xmath127 is the debye temperature , @xmath128 is the volume of sample # 2 and @xmath129 is the number of atoms in the crystal .
this leads to @xmath56k and @xmath130 m/s .
however , the experimental specific - heat data on lapt@xmath0ge@xmath1 follow only approximately a @xmath131 dependence , but can be better described by a @xmath132 law below 4k ( see inset of fig . [
fig : c - vs - t](a ) ) .
interestingly , the thermal conductivity follows a @xmath133 dependence at low @xmath29 , which is close to the experimentally determined @xmath29-dependence of @xmath122 .
the almost identical exponents hint at a scenario , where phonons scattered on boundaries are the main contribution to @xmath134 at low @xmath29 .
a fit of the thermal conductivity with an adjusted power law @xmath135 , with the velocity @xmath125 calculated above , and @xmath123 as free parameter is shown in fig .
[ fig : kappa - vs - t](b ) .
it leads to a mean free path of @xmath136 mm , which is a reasonable value considering the shortest of the sample dimensions ( @xmath137)mm@xmath138 . this nice agreement suggests first boundary scattering as the main contribution to @xmath123 and a negligible contribution of specular reflections , which would lead to a considerably larger @xmath123 .
secondly , it points towards a negligible electronic contribution to @xmath134 at low @xmath29 and thus to a finite gap at every @xmath139 point of the fermi surface . with the results from the low - temperature thermal conductivity
, we can also understand the behavior in the whole temperature range ( fig .
[ fig : kappa - vs - t](a ) ) . at @xmath36 ,
a clear drop is visible followed by a hump at slightly lower @xmath29 . in general , the thermal conductivity both in the normal state , @xmath140 , and in the superconducting state , @xmath141 , consists of a phonon and an electron contribution @xmath142 assuming an @xmath3-wave superconductor , the drop can be attributed to the decreasing number of electronic heat carriers , hence , @xmath143 decreases rapidly .
the hump indicates an enhanced mean free path of phonons due to the decreasing number of electronic scattering centers , therefore , @xmath144 increases below @xmath36 .
j. bardeen , g. rickayzen , and l. tewordt ( brt ) developed a standard theory for thermal conductivity in the case of an @xmath3-wave superconductor @xcite .
in the spirit of the brt theory , we sketched the temperature dependence for both @xmath103 and @xmath102 in the inset of fig .
[ fig : kappa - vs - t](a ) , which can qualitatively explain our observed temperature dependence of the thermal conductivity . is almost linear and develops a minimum at low fields which becomes more pronounced with higher temperature .
the shaded areas illustrate qualitatively how the electronic and the phonon contribution change with field @xcite .
( b ) we compare the field - dependent thermal conductivity data for @xmath145 with the following materials representing different gap symmetries .
@xmath113-wave # 1 : tl@xmath98ba@xmath98cuo@xmath114 ( @xmath146 ) @xcite , @xmath113-wave # 2 : ceirin@xmath116 ( @xmath146 ) @xcite , multiband : mgb@xmath98 ( @xmath147 ) @xcite , clean @xmath3-wave : nb ( @xmath148 ) @xcite , dirty @xmath3-wave low @xmath29 : inbi ( @xmath149 ) @xcite dirty @xmath3-wave high @xmath29 : ta@xmath150nb@xmath151 ( @xmath152 ) @xcite .
note the different scaling of the field axis to account for the different @xmath153 and @xmath154 .
] we now turn to the field - dependence of @xmath155 for lapt@xmath0ge@xmath1 , which is presented in fig .
[ fig : kappa - vs - h](a ) for selected temperatures between 290mk and 650mk together with a zero - temperature extrapolation .
the measurements were performed on increasing field .
a comparison with measurements during decreasing field showed no hysteresis .
the field - dependence is consistent with temperature - dependent measurements at finite fields , which were performed on warming and which do not show any difference between field- and zero - field - cooling .
a clear kink is visible at @xmath156 .
this is slightly lower than 1.6 t from the extrapolation of the specific - heat data on polycrystals obtained at @xmath157 @xcite and from resistivity data down to 0.3k @xcite .
the latter observation is in agreement with resistivity measurements on our samples ( not shown ) , which show a slightly higher @xmath30 than from thermal transport and specific heat ( fig .
[ fig : c - vs - h ] ) .
this deviation is most likely due to surface effects .
below @xmath30 , the thermal conductivity shows an almost linear field dependence in the zero - temperature limit . for finite temperatures
, a minimum appears at low fields .
it shifts to higher fields and becomes more pronounced as the temperature is increased .
generally , @xmath158 is almost constant below @xmath153 @xcite , a regime which we will not consider here since @xmath159 obtained from magnetization measurements is very small in lapt@xmath0ge@xmath1 . however , there are drastic changes above @xmath153 due to the properties of the vortex state : in addition to the delocalized quasiparticles due to thermal excitations above the gap , there are quasiparticles from the core region , which lead to an increase of @xmath143 @xcite . around the vortex core , a supercurrent flows , which decays over a distance roughly equal to the penetration depth @xmath160 .
the supercurrent with the velocity @xmath161 leads to a doppler shift of the energy of delocalized quasiparticles @xmath162 @xcite .
this changes the excitation spectrum , increases the density of states , and effectively lowers the gap @xmath163 for @xmath139 directions with a component parallel to @xmath161 .
this effect also increases @xmath143 @xcite and becomes important especially for superconductors with nodes in the gap .
scattering of electrons and phonons on the vortex lattice decreases @xmath143 and @xmath144 @xcite .
the sum of all these effects leads to pronounced and typical differences in @xmath158 for materials with different gap structures .
they are all anisotropic with respect to the angle between current and field and also with respect to the angle between field and wave vector @xcite . in the following
we restrict the discussion to our case of a field perpendicular to the current .
the field dependence of @xmath164 can be affected by the quality of the sample as exemplified in fig .
[ fig : kappa - vs - h]b for the case of an @xmath3-wave superconductor @xcite . in fig .
[ fig : kappa - vs - h](b ) , we compare our data with results on different materials .
the materials chosen for this comparison show a field - dependence of the thermal conductivity typical for their superconducting gap symmetry .
the key signatures found in lapt@xmath0ge@xmath1 are the finite slope close to @xmath30 , no residual term @xmath100 at zero field , and a minimum at intermediate fields , which is absent for @xmath145 but increases in amplitude with increasing @xmath29 .
the curve for lapt@xmath0ge@xmath1 compares best with that of inbi , which is a representative example for dirty s - wave superconductors @xcite .
the most important difference to the clean @xmath3-wave case is the infinite slope at @xmath30 found in the latter one .
we also added data for a second dirty superconductor at higher relative temperature to illustrate the appearance of a minimum at intermediate fields very similar to our observation in lapt@xmath0ge@xmath1 .
the good agreement of @xmath158 in lapt@xmath0ge@xmath1 with that of a dirty @xmath3-wave superconductor fits to the specific heat results and the temperature dependence of the thermal conductivity .
we find in addition , that our sample is actually in the dirty limit , since the ginzburg - landau coherence length @xmath165 is of the same order as the superconducting mean free path @xmath166 . here
, we used the upper critical field to determine @xmath167 with @xmath168 . to calculate @xmath169
, we applied @xmath170^{0.5}\,\mathrm{cm}\ , , $ ] @xcite where we use the electronic specific heat coefficient @xmath171 in the unit @xmath172 , the residual resistivity @xmath173 in the unit @xmath174 ( from crystal # 3 , not shown ) , @xmath175 in the unit k , and the aforementioned coherence length @xmath176 in the unit cm . the good agreement with results for a dirty @xmath3-wave superconductor implies , that the field dependence of the thermal conductivity is composed of a phonon part @xmath144 , which decreases with increasing field due to scattering on vortices , and an electronic contribution , which increases due to an enhanced number of localized states in the vortex cores tunneling to ever closer neighboring vortices .
this behavior is indicated in fig . [
fig : kappa - vs - h](a ) . since our specific - heat results indicate that a description within a two - band model is also consistent with the data , we like to analyze our thermal - transport data with respect to this scenario as well . a typical feature of @xmath164 in multiband systems is a plateau at intermediate fields as shown in fig .
[ fig : kappa - vs - h](b ) for mgb@xmath98 , which is attributed to the suppression of the smaller of the two gaps .
we do not observe such change of curvature in lapt@xmath0ge@xmath1 .
however , this change might be very weak as _
e.g. _ in the case of nbse@xmath98 @xcite .
let us now compare our thermal conductivity results with observations and predictions for the case of nodes in the superconducting gap .
there are both theoretical @xcite and experimental @xcite reports on a minimum in @xmath164 for unconventional superconductors .
the slope above the minimum can vary considerably depending on the type of gap and the purity of the sample @xcite ( _ cf . _ fig .
[ fig : kappa - vs - h](b ) ) .
such a minimum is consistent with our observation .
hence , this property does not rule out nodes in the gap of lapt@xmath0ge@xmath1 completely .
however , one would expect a sizable residual term for @xmath145 in a nodal superconductor .
in contrast , both single - band and multiband @xmath3-wave superconductors show an insignificantly small residual term exactly as our results on lapt@xmath0ge@xmath1 .
we performed specific - heat and thermal - conductivity measurements to investigate the superconducting order parameter of lapt@xmath0ge@xmath1 .
the specific heat shows a sharp superconducting transition at 8.0k and its zero - field temperature - dependence in the superconducting state can nicely be described assuming a single bcs @xmath3-wave gap of the size @xmath177 .
the field dependence of the @xmath19-coefficient deviates from the simple linear behavior for @xmath3-wave pairing . such a deviation can arise , if the fermi surface shows strong anisotropies as is the case for lapt@xmath0ge@xmath1 .
we used the specific - heat results to analyze the low - temperature thermal - transport in detail .
the temperature dependence of the thermal conductivity reveals a dominant phonon contribution at low temperatures and a negligible electronic residual term .
together with the dependence of the thermal conductivity on magnetic field , this behavior suggests a single gap bcs @xmath3-wave superconductor .
however , we can not completely exclude a two - band scenario from our experimental results , in particular if both gaps have a similar size @xmath178 or if one of them has only a tiny contribution to the thermal - transport and thermodynamic properties .
the analysis of the specific heat data enables us to limit the possible range of amplitudes @xmath37 and @xmath38 and weighting factors @xmath35 within the two - gap @xmath34-model .
the possible parameter range is comparatively large , since a single - band model is already sufficient to describe our experimental results .
since the sister compound prpt@xmath6ge@xmath1 is discussed to be a multigap and/or unconventional superconductor , further thermal conductivity and specific heat measurements on the substitution series ( pr@xmath4la@xmath17)pt@xmath0ge@xmath1 are highly desired .
they might help to resolve the question , if prpt@xmath6ge@xmath1 and lapt@xmath0ge@xmath1 have compatible superconducting order parameters as was suggested by the smooth evolution of @xmath179 across the substitution series .
we are indebted to r. daou and h. rosner for valuable discussions . | we investigated the superconducting order parameter of the filled skutterudite lapt@xmath0ge@xmath1 , with a transition temperature of @xmath2k . to this end , we performed temperature and magnetic - field dependent specific - heat and thermal - conductivity measurements .
all data are compatible with a single superconducting @xmath3-wave gap .
however , a multiband scenario can not be ruled out .
the results are discussed in the context of previous studies on the substitution series pr@xmath4la@xmath5pt@xmath0ge@xmath1 .
they suggest compatible order parameters for the two end compounds lapt@xmath0ge@xmath1 and prpt@xmath6ge@xmath1 .
this is not consistent with a single @xmath3-wave gap in lapt@xmath0ge@xmath1 considering previous reports of unconventional and/or multiband superconductivity in prpt@xmath6ge@xmath1 . |
in hierarchical scenarios of galaxy formation ( white & rees 1978 ) , dwarfs would be the first galaxies to be formed , from small - amplitude density fluctuations and would latter on merge to form larger galaxies and structures .
the fact that the gas - poor , des are usually found in denser parts of the universe than the gas - rich dirrs may indicate that both have the same origin , the former being in a more evolved evolutionary stage and in the process of being soon incorporated into bigger neighboring galaxies . if this were the case , isolated dirrs like ddo 187 should still have the traces of their formation process and early evolution and , eventually , a similar underlying structure to that of the less distorted des .
it is hence of major relevance to study the structure , kinematics and spatial distribution of stellar populations , including the oldest ones , of these small , gas rich galaxies , checking in particular whether they have old / young subsystems similar to the halo / disk ones of spirals or any other structures that would trace their formation scenario .
the main purpose of this paper is to test the hypothesis of the existence of a two - component , halo / disk structure in the field dirr galaxy ddo 187 .
ddo 187 ( ugc 9128 ) is a resolved , gas - rich dwarf galaxy first cataloged by van den bergh ( 1959 ) .
it shows a morphology typical of dirr galaxies , with the bluest stellar population lying on one - half of its body , and has been subject of several works directed toward the study of its distance , stellar content , neutral gas , and chemical abundances .
aparicio , moles , & garca - pelayo ( 1988 ) showed a color magnitude diagram ( cmd ) and roughly estimated the distance to the galaxy to be @xmath3 mpc , based on the magnitude of the brightest stars .
skillman , kennicutt , & hodge ( 1989 ) found a very low oxygen abundance of @xmath4 .
recently , van zee , haynes , & salzer ( 1997a ) have obtained @xmath5 and 7.73 for two h ii regions in the galaxy .
lo , sargent , & young ( 1993 ) determined its heliocentric h i velocity ( 153 km s@xmath6 ) and using a distance of 4 mpc , obtained its h i and total virial masses to be @xmath7 m@xmath8 and @xmath9 m@xmath8 .
van zee , _ et al . _
( 1997b ) , assuming a distance of 3.4 mpc , obtained @xmath10 m@xmath8 . using surface photometry ,
patterson & thuan ( 1996 ) derived an exponential scale length of 0.25 kpc , for a distance of 4.4 mpc .
recently hoessel , saha , & danielson ( 1998 ) presented the results of their time - extended research of variable stars in ddo 187 .
they find 12 variable stars , and identify two of them as very likely cepheids and three more as possible cepheids . using their periods and luminosities ,
they derive a distance of @xmath11 mpc for ddo 187 .
we present new ground - based @xmath12 , @xmath0 , and @xmath13 photometry of ddo 187 and analyze the properties of this galaxy .
the paper is organized as follows .
observations are described in section 2 .
the cmd of the resolved stars is presented in section 3 .
our photometry reach the tip of the red giant branch ( trgb ) , allowing an independent determination of the distance , which does not agree with the cepheid - based one .
this is discussed in section 4 .
the ionized distribution of h and the properties that can be inferred for the ionizing stars are discussed in section 5 .
section 6 is devoted to the study of the structure , star formation history ( sfh ) , and spatial distribution of stellar populations , discussing , in particular , the hypothesis that ddo 187 had a two - components ( halo / disk - like ) structure . in sec .
7 , several global properties of the galaxy are given .
finally , sec .
8 summarizes the main results of the paper .
images of ddo 187 were obtained in the @xmath12 and @xmath0 johnson cousins filters and in a narrow - band @xmath13 filter with the not ( 2.5 m ) at roque de los muchachos observatory on the island of la palma ( canary islands , spain ) .
the hirac camera was used with a @xmath14 loral ccd binned to @xmath15 .
after binning it provides a scale of 0.22 @xmath16/pix and a total field of @xmath17 arcmin@xmath18 .
total integration times were 3600 s in @xmath12 , 3000 s in @xmath0 , and 1200 s in @xmath13 .
a 400 s exposure was also taken with an @xmath13-continuum filter .
moreover , integrations of 1200 s in @xmath12 and 1000 s in @xmath0 were made of a nearby field to correct the foreground contamination of the broad - band images .
observations of ddo 187 were taken with seeing always in the range @xmath19 .
table [ journal ] gives the journal of observations .
figure [ i m a ] shows one of the @xmath0 band images of ddo 187 .
a beautiful picture of the galaxy can be seen in hoessel _
( 1998 ) .
bias and flat - field corrections were done with iraf .
then , daophot and allstar ( stetson 1994 ) were used to obtain the instrumental photometry of the stars .
eighteen standard stars from the list of landolt ( 1992 ) were measured during the observing run to calculate the atmospheric extinctions for each night and the equations transforming into the johnson
cousins standard photometric system
. a total of about 180 measures in both the @xmath12 and @xmath0 bands of these standards were used .
the transformation equations are : @xmath20 @xmath21 where capital letters stand for johnson cousins magnitudes and lower - case letters refer to instrumental magnitudes corrected for atmospheric extinction .
the @xmath22 values are the dispersions of the fits at the centers of mass of the point distributions ; hence they are the minimum internal zero - point errors .
dispersions of the extinctions for each night vary from @xmath23 to @xmath24 and dispersions of the aperture corrections are of the order of 0.04 . putting all these values together
, the total zero - point error of our photometry can be estimated to be about 0.05 for both bands .
one standard star from oke ( 1990 ) was observed three times to calibrate the @xmath13 images .
the dispersion of the measures was better than 0.01 magnitudes .
daophot and allstar provide with psf model to star profile fitting errors . in general
, they do not reproduce the external errors of the photometry , coming mainly from stellar crowding and blending ( see aparicio , & gallart 1995 and gallart , aparicio , & vlchez 1996a ) , but they provide with an indication of the internal accuracy of the photometry . in this sense ,
our average allstar errors are 0.01 , 0.04 and 0.16 for @xmath25 , 22 and 24 respectively and for @xmath26 , 23.3 and 25.4 , respectively .
the cmd of ddo 187 is shown in figure [ cmd ] .
it is typical of a resolved galaxy with recent star formation ( compare , for example , with the richer cmd of ngc 6822 by gallart et al .
a well populated structure is visible in the region @xmath27 , @xmath28 .
it corresponds to the _ red - tangle _ ( see aparicio & gallart 1994 ; aparicio _ et al .
it is formed by low - mass stars in the rgb and agb phases and it is hence the trace of an intermediate - age to old stellar population .
a well populated blue - plume is also visible at @xmath29 , @xmath30 . formed by intermediate to massive stars in the ms and/or the core he - burning blue - loop phases , it is the trace of significant recent star formation , in good agreement with the presence of h ii regions in the galaxy . at least some of the red stars in the region @xmath31 , @xmath32 are likely agb stars forming a _ red - tail _ structure of the kind discussed by aparicio & gallart ( 1994 ) .
figure [ f_cmd ] shows the cmd of a companion field , close to ddo 187 but far enough to guarantee that it contains no stars of the galaxy .
the distribution of stars in this diagram indicates that most stars brighter than @xmath33 and some of the fainter in the cmd of fig .
[ cmd ] are foreground stars belonging to the milky way . the cmd of fig .
[ f_cmd ] will be used to correct for foreground contamination the star counts in ddo 187 used in section 6 . in summary
, the cmd of ddo 187 indicates that it has had an extended sfh , as evidenced by the presence of young blue stars and a red - tangle of intermediate - age and/or old stars .
a quantitative estimate of the sfh is given in section 6.2 .
hoessel _ et al . _
( 1998 ) made a wide time coverage survey of variable stars in ddo 187 .
they identified two stars in the galaxy as very likely cepheids and three more as possible cepheids . using them ,
they determine a distance of @xmath11 mpc .
this distance implies that the trgb of ddo 187 should be fainter than @xmath34 , and hence difficult to reach by usual ground - based telescopes .
however , this seems not to be the case in the light of the cmd shown in figure [ cmd ] .
an independent estimate of the distance to ddo 187 can be made using the magnitude of the trgb and the relationships of lee , freedman , & madore ( 1993 ) .
we have determined the magnitude of the trgb from the cmd of fig .
[ cmd ] , applying a sobel filter to the luminosity function ( lf ) of stars with colors in the interval @xmath35 , and obtained @xmath36 .
the error has been estimated from the width of the peak produced by the sobel filter around the trgb .
several binning sizes and initial values for the lf were used , the former result being the average of all them .
crowding , blending and other observational effects modify the measured star magnitudes in the sense that , in average , they appear brighter than they are ( see a discussion of this in aparicio , & gallart 1995 and gallart et al .
this would eventually affect the magnitude of the trgb . to check this effect and for the analysis of the stellar populations presented in section 6
, we have performed crowding tests using artificial stars in the way described by stetson ( 1994 ) .
we have added 1190 artificial stars to our images with initial magnitude and color @xmath37 , @xmath38 , roughly corresponding to the position of the trgb .
90% of them were recovered , their average magnitude and color being @xmath39 and @xmath40 , showing that shifts produced by observational effects are negligible at the position of the trgb . the extinction towards ddo 187 can be estimated to be @xmath41 ( burstein & heiles 1982 ) .
this can be neglected and hence the former value for @xmath36 can be assumed to be free of extinction and observational effects . to apply the relationships of lee _ et al . _
( 1993 ) , the metallicity of the stars and the bolometric correction ( bc ) at the trgb are necessary . both can be estimated from the color indices of rgb stars . @xmath42 and @xmath43
have been obtained as the median of the color indices of stars with @xmath44 and @xmath45 respectively and having colors in the interval @xmath46 .
some 35 stars have been used in each case and the quoted errors have been simply obtained as @xmath47 for each sample . from @xmath48 , an average metallicity of [ fe / h ] @xmath49
is found using the calibration by lee _
this fe abundance corresponds to @xmath50 , which is similar to the values obtained from the oxygen abundance of the h ii regions by skillman _
( 1989 ) and by van zee _ et al . _
( 1997a ) ( @xmath51 and @xmath52 , respectively ) .
these values can finally be introduced in the relationships of lee _ et al . _
( 1993 ) to obtain the absolute magnitude of the trgb , @xmath53 , and hence a distance modulus of @xmath54 , which corresponds to a distance to the milky way of @xmath55 mpc . to check whether ddo 187 can be considered a field galaxy ,
its distance to other nearby galaxies has been computed .
the selected galaxies are gr 8 , ddo 187 , andromeda , m81 , m94 and m101 , as well as the barycenter of the local group .
the first two galaxies are apparently the closest dwarfs to ddo 187 .
the other ones are the dominant members of the closest groups .
the distances to the milky way adopted for each one are : gr 8 , 2.2 mpc ( dohm - palmer _
_ 1998 ) ; ddo 190 , 2.9 mpc ( aparicio _ et al .
_ 1999 ) ; andromeda , 0.77 mpc ( freedman & madore 1990 ) ; m 81 , 3.6 mpc ( freedman _ et al .
_ 1994 ) ; m 94 , 4.7 mpc ( mulder & van driel 1993 , adopting @xmath56 kms@xmath6mpc@xmath6 ) and m 101 , 7.0 mpc ( stetson _ et al .
the position of the barycenter of the local group has been calculated neglecting the masses of all the galaxies in the lg except andromeda and the milky way , assuming that the the mass of the milky way is 0.7 times that of andromeda ( peebles 1989 ) and adopting a value of 0.77 mpc for the milky way
andromeda distance ( freedman & madore 1990 ) . in this way , the resulting distances of ddo 187 to each galaxy are : 0.85 mpc to gr 8 ; 1.06 mpc to ddo 190 ; 2.9 mpc to andromeda ; 3.2 mpc to m 81 ; 2.6 mpc to m 94 ; 5.2 mpc to m101 and 2.7 mpc to the barycenter of the local group . hence ddo 187 is likely a field galaxy lying between the local , m81 and m94 groups , the closest galaxy being the dirr gr 8 .
all the distances are summarized in table [ neig ] . the large disagreement between the cepheid and trgb distances merits some comment . while the cepheid p
l relationship is the best distance indicator for nearby galaxies , its reliability has been recently questioned by several authors ( aparicio 1994 ; saha _ et al .
_ 1996 ; tolstoy _ et al .
the problem arises from the fact that , for several dwarf galaxies ( namely , pegasus , leo a and , now , ddo 187 ) , a large disagreement has been detected between cepheid based distances and trgb or red - clump ( rc ) based distances .
table [ dist ] summarizes the distances found with each method for the three galaxies .
column 1 gives the galaxy name ; columns 2 to 4 list the cepheid distance , the number of cepheids used to determine it and the reference ; columns 5 to 7 give the cmd - based distance , the feature ( trgb or rc ) used to estimate it and the reference .
in all cases , cepheid distances were based on just a few stars showing light curves resembling those of bona - fide cepheids . however , latter accurate two - color photometry ( aparicio 1994 ; tolstoy _ et al .
_ 1998 ) demonstrated that the colors of these stars could in no way be reconciled with those of cepheid stars , which is also stressed by saha _
( 1996 ) .
in the case of ddo 187 , hoessel _ et al . _
( 1998 ) identified only two stars as very likely long - period cepheids and three more , with lower quality photometry , as possible cepheids .
they used a control cmd for these stars but perhaps it was not deep enough for the purpose .
figure [ cmd_cef ] shows our cmd with the cepheid candidates of hoessel _ et al . _
( 1998 ) marked with filled dots . from their location in the diagram , these stars are likely agb stars and could well be some kind of long - period variables ( for further discussion on the presence of long - period variables in ddo 187 see hoessel et al .
1998 ) .
considering that ddo 187 is the third case in which a large distance disagreement has been found , it seems clear that cepheid distance estimates based on a few stars , as usually happens for dwarf galaxies , must be accepted with caution .
possibly , only galaxies with a large young star population , like spirals or dwarfs in a phase immediately following a strong star formation burst , are adequate for a confident use of cepheid stars as distance indicators , since only in these galaxies a significant population of cepheids is expected .
the analysis should ever include checking colors in a reliable cmd and a good phase coverage in order to avoid aliased periods ( see also saha _ et al .
_ 1996 on the distance of the postburst dirr ic 10 ) .
in the following we will use @xmath57 mpc for the discussion of the properties of ddo 187 .
figure [ ha_labels ] shows the @xmath13 image of ddo 187 , after subtraction of the continuum .
we have identified 10 h ii regions of different morphology and compactness .
a summary of the results is given in table [ fluxes ] .
column 1 gives the identification number as shown in fig .
[ ha_labels ] , followed by a bracketed c or d indicating whether the h ii region is compact or diffuse , respectively .
column 2 lists an estimate of the size . for regions 1 to 7 , the sizes are the full width at half of maximum , while for the diffuse regions 8 to 10 , the sizes are the square root of the total area used to calculate the flux .
column 3 gives the flux received from each region in erg s@xmath6@xmath58 .
column 4 gives the corresponding @xmath13 luminosities in ergs@xmath6 calculated for a distance of 2.5 mpc .
column 5 lists the number of ionizing photons per second required to produce those luminosities , calculated following kennicutt ( 1988 ) .
column 6 gives the masses required for single ms stars to produce such amounts of ionizing photons ( they have not been calculated for the diffuse regions ) . to calculate these masses , the relation given in fig .
6 has been used .
it has been plotted from calculations made using the models of panaggia ( 1973 ) and the padua stellar evolutionary library ( bertelli _ et al .
it must be kept in mind that , at the distance of ddo 187 , @xmath59 corresponds to 12 pc , so the assumption that the h ii regions are ionized by single stars may well be an oversimplification , even for the compact ones .
the total @xmath13 flux , luminosity and number of ionizing photons are given in the last line of table [ fluxes ] including also the very diffuse emission , not comprised in the regions 1 to 10 listed in the table .
@xmath60 can be used to estimate the current sfr .
it is given by @xmath61 where time increases from the present to the past , the present time being 0 , @xmath62 is the imf , @xmath63 is the sfr in units of number of stars per year and @xmath64 is the ms life - time of a star of mass @xmath65 and @xmath66 is the number of lyman photons emitted per second by a ms star of mass @xmath65 .
while we are interested in stars more massive than some 10 m@xmath8 ( see table 2 ) we need to consider only the interval of time 0 to @xmath67 .
if , for simplicity , we assume that the sfr is constant over this interval , then @xmath68 where @xmath69 is the constant current value of the sfr in yr@xmath6 . using data for stars of metallicity @xmath70 from the stellar evolutionary library of padua ( see bertelli _
1994 ) the ms life - time ( in years ) can be approximated analytically by @xmath71 moreover , for masses in the interval 12m@xmath72m@xmath73 , @xmath66 can be approximated analytically by @xmath74 introducing the imf of kroupa , tout , & gilmore ( 1993 ) , equation ( 4 ) becomes @xmath75 with @xmath76 , @xmath77 , or @xmath78 for @xmath79m@xmath73 , 30m@xmath73 and 40m@xmath73 , respectively .
the result is almost insensitive to the value of @xmath80 , which has been fixed to @xmath81m@xmath8 . using the total value of @xmath60 given in table [ fluxes ]
, the current sfr can be adopted to be @xmath82yr@xmath6 .
it is better to give the sfr in units of mass . from integration of the kroupa _
( 1993 ) imf from 0.1 m@xmath8 to @xmath83 the average mass of the stars results 0.51 m@xmath8 and so the sfr becomes @xmath84m@xmath73yr@xmath6 .
this can be assumed to be the sfr in ddo 187 averaged for the last @xmath85 20 myr .
different imfs would produce different results for the sfr . for comparison , using a salpeter imf ( @xmath86 ) with lower mass limit @xmath87 m@xmath8 , would produce a sfr 10% smaller than the former value .
evident galactocentric gradients in the young stellar population have recently been found in several dirrs : wlm ( minniti & zijlstra 1996 ) ; antlia ( aparicio _ et al . _
1997c ) ; phoenix ( martnez - delgado , gallart , & aparicio 1999 ) ; ngc 3109 ( minniti , zijlstra , & alonso 1999 ) .
all these galaxies have undergone recent star formation episodes in their centers , but lack young stars in outer regions .
however the conclusion can not be drawn from this fact alone that they have two differentiated structures , the outer being an old subsystem tracing out the formation and early evolution of the galaxy .
this result would not be surprising , but sufficiently accurate age distributions and structural and kinematical information are needed before confidently accepting it .
interesting steps towards establishing the primeval nature of the extended structures are supplied by minniti , & zijlstra ( 1997 ) who provide with clues compatible with an old age of the extended population in wlm ( although the presence of an intermediate - age component can not be , in our opinion , excluded ) and include a morphological analysis concluding that flattening is smaller for the halo than for the inner component .
a further important step forward is supplied by the finding that the only globular cluster in wlm is indeed a very old object , sharing the metallicity of the outer component field stars ( hodge _ et al .
. however , wlm has also a very extended gas component ( see minniti , & zijlstra 1997 ) which , at least and in principle , would play against the formation of the extended component before the collapse of the galaxy had finished .
phoenix provides an equally interesting case . lacking gas associated to the main , optical body
, this galaxy has an outer stellar component possibly ( but not certainly ) formed by old stars and rotated @xmath88 with respect to the inner component ( martnez - delgado et al .
this fact , by itself alone , strongly indicates a possible separation of both substructures . together with the presence of young stars in the inner component and a gas cloud at 6 arcmin away from them ( young , & lo 1997 ) it makes phoenix of mayor relevance in the study of extended components .
centering our attention on ddo 187 , we will show in this section how our data , together with information about the gas distribution , could indicate that a stellar population lacking young stars would be associated with an external , diffuse , perhaps differentiated subsystem of this galaxy . on the light of these finding
, we will discuss the possible disk - halo structure on the galaxy . from surface brightness photometry , patterson & thuan ( 1996 , 1998 )
have determined the holmberg radius , ellipticity and position angle of ddo 187 to be @xmath89 , @xmath90 and pa @xmath91 . on the other hand , for the h i component of the galaxy , lo _
( 1993 ) obtain @xmath92 , @xmath93 , pa@xmath94 , where @xmath95 is the major axis measured to 10% of the peak flux . in other words , the gas component has a similar extension to that of the optical light integrated to the holmberg radius , but , to the first approximation , the former shows a spherical morphology , in contrast to the likely flattened structure of the latter . to further investigate the structure of the galaxy
, we have searched the morphology of the distribution of stars to its outermost extension .
we have obtained the number density of resolved stars and fitted it with ellipses .
position angles and ellipticities for different semi - major axes have been obtained with the iraf package . except for the inner @xmath96 ,
the ellipses show pa in the range 20@xmath97@xmath98 , compatible with that of the optical holmberg radius ellipse , but with ellipticities @xmath99 0.70.85 .
see below for more discussion about the different components of the galaxy .
we then obtained the surface density of stars in the galaxy by counting stars in ellipses of increasing semi - major axis @xmath100 .
all the stars shown in the cmd of figure [ cmd ] have been used .
we have adopted @xmath101 and pa @xmath102 for all the ellipses .
these values correspond to those of the outermost ellipses , at @xmath103 .
the results are shown in figure [ densi ] . the background level estimated from our comparison field
is shown by the line in the lower right part of the figure .
after subtracting this background level , we performed an exponential fit to the region @xmath104 .
the innermost part of the galaxy has been excluded to avoid completeness effects .
the fit yields a scale factor @xmath105 , which corresponds to 310 pc at the distance of ddo 187 . after scaling to the same distance ,
this is more than twice the result obtained by patterson & thuan ( 1996 ) for the surface brightness distribution , indicating that the galaxy is larger than was previously thought .
in fact , the surface brightness distribution given by patterson , & thuan ( 1996 ) for ddo 187 extends to 50 , i. e. covering only the inner body of the galaxy .
for this reason , roughly estimated surface brightness scales in @xmath12 and @xmath0 are also plotted in the right - hand vertical axis of figure [ densi ] .
the scales can be obtained making @xmath106 for the case of @xmath107 and @xmath108 for the case of @xmath109 and have been computed comparing the logarithmic number of stars in a region of @xmath110 centered on the galaxy with the integrated magnitudes of the same region .
the surface brightness profile by patterson , & thuan ( 1996 ) is to be preferred for the inner region of the galaxy , the surface brightness scale of figure [ densi ] being useful for the external region .
surface magnitude of the brightest , south - west region of the galaxy , where the youngest stars are placed are @xmath111 and @xmath112 .
errors are the zero - point errors of our photometry . the sfh of a galaxy can be derived in detail from a deep cmd through comparison with synthetic cmds ( see gallart _
the cmd of ddo 187 is not deep enough for such a detailed analysis but it is still possible to sketch the sfh of the galaxy for old , intermediate and young ages .
the galaxy has been divided into three parts for the study of the sfh as shown in figure [ mapa ] .
the inner part corresponds to the optical body of the galaxy down to the holmberg radius , as determined by patterson & thuan ( 1996 , 1998 ) .
the intermediate part matches the gas component as given by lo _
( 1993 ) , excluding the inner component .
the outer part includes the rest of the galaxy out to @xmath113 where , according to the results given above , the galaxy seems to vanish .
the resolved stars have also been plotted in figure [ mapa ] .
while most of the blue and many bright red stars are in the inner component , the fainter red population extends outwards and defines the ellipticity and orientation of the outer component . in what follows
, we will discuss the derivation of the sfh for each region .
note that the different stellar populations will evolve kinematically in different ways and will not necessarily be found in the same region where they were born .
so , properly speaking , what we are going to obtain is the distribution of stellar ages , with indication of stellar metallicities in each region . for simplicity
, we will continue to term the combined distribution as sfh .
we will look for differences in these distributions as a first indication of differences in the underlying structure and/or evolutionary history of the galaxy .
a simplified version of the method proposed by aparicio , gallart , & bertelli ( 1997b ) and used by gallart _
( 1999 ) for leo i has been employed to obtain the sfh of the three regions defined in ddo 187 . in practice , a synthetic cmd with arbitrary , constant sfr of value @xmath114 , the imf of kroupa _ et al . _
( 1993 ) , 25% of binary stars with mass ratios , @xmath115 , uniformly distributed in the interval @xmath116 , and a metallicity , @xmath117 , taking random values from @xmath118 to @xmath119 independently of age have been used .
the @xmath117 interval has been chosen considering the metallicities for h ii regions given by skillman _
( 1989 ) and by van zee _ et al . _
( 1997a ) , the value of [ fe / h ] given in sec . 4 , and the dispersion of colors in the red - tangle .
the resulting synthetic cmd has been then divided into four age intervals : 00.1 gyr , 0.10.3 gyr , 0.34 gyr , and 415 gyr . following the nomenclature introduced in aparicio _
( 1997b ) , each of these synthetic diagrams will be called a _ partial model _ cmd and any linear combination of them will be denoted as _
global models_. eight regions have been defined in the observed and the partial model cmds with the criterion that they sample different age intervals ( see gallart _
_ 1999 and aparicio 1999 ) .
the regions are shown in figure [ cmd_reg ] .
we will denote by @xmath120 the number of stars of the observed cmd lying in region @xmath121 and by @xmath122 the number of stars of partial model ( age interval ) @xmath123 populating region @xmath121 . the number of stars populating a given region in a global model is then given by @xmath124 and the corresponding sfr @xmath125 where @xmath126 are the linear combination coefficients , @xmath127 is a scaling constant , and @xmath128 if @xmath129 is inside the interval corresponding to partial model @xmath123 and @xmath130 otherwise .
the @xmath131 having the best compatibility with the data can be obtained by a least - squares fitting of @xmath132 to @xmath120 . in practice
, several solutions can be found showing similar departures , @xmath22 , from the data . to smooth out the bias that selecting a single good solution would introduce , several
are selected and averaged to obtain the best final solution . in the case of ddo 187 ,
models showing @xmath22 values smaller than 1.5 times the @xmath22 of the best solution have been used .
the resulting @xmath131 for the three regions in which ddo 187 has been divided are shown in figure [ psi ] .
the right - hand vertical scales are normalized to the area covered by each region .
the most interesting result from the sfh analysis is that the central region shows a strong increase of @xmath131 for the last few 100 myr and that the outer region shows no star formation activity , in good agreement with what is expected from the absence of gas .
this will be used below for the discussion of the structure of ddo 187 .
the current sfr obtained in sec .
5 from the h@xmath133 luminosity refers to the last @xmath85 20 myr .
it is a factor of eight smaller than the value derived here for the last 100 myr in the central region of the galaxy , but it is similar to the sfr averaged over the entire life - time of the galaxy .
this picture favors a bursting star formation mode , with the last strong burst having taken place between 20 and 100 myr ago .
if this burst had lasted 10 myr , its intensity would have been of @xmath134 m@xmath8 yr@xmath6 , which is not far from the values given by fanelli , oconnel , & thuan ( 1988 ) for blue compact dwarfs .
it is also interesting to note that the picture of high present - day or very recent star formation activity much higher than the average for the entire life - time of the galaxy is frequently found in galaxies classified as dirr ( ngc 6822 : gallart _ et al .
_ 1996b , c ; pegasus : aparicio , gallart , & bertelli 1997a , gallagher _ et al .
_ 1999 ; antlia : aparicio _ et al . _
1997c ) . on the other hand , galaxies like phoenix ( martnez - delgado _ et al .
_ 1999 ) or lgs 3 ( aparicio , gallart _ et al . _
1997b ) , frequently considered as dsph - dirr transition objects , show similar @xmath131 histories until a few 100 myr ago , failing only in currently producing stars at a high rate .
this would also point to a bursting mode of star formation , the dsph - dirr galaxies being in a quiescent period , and would also imply that an important bias could exist in the classification of dwarfs as bona - fide dirrs towards objects experiencing strong star formation bursts .
the sfr averaged for different periods of time is given in table [ sfrave ] for the three regions defined in ddo 187 .
the first four lines give the sfr averaged for the whole life - time of the galaxy ( assumed to be 15 gyr ) and for the intervals 151 gyr , 10 gyr , and 0.10 gyr ago .
line 5 gives the current sfr derived from the @xmath13 flux .
lines 6 to 10 give the same information as lines 1 to 5 after normalization to the area of each region . some of the results presented here are compatible with ( but not conclusive of ) a two - components , disk / halo - like structure in ddo 187 .
let us assume the two - component possibility as a working hypothesis and check to what extent it could be accepted .
there are three facts supporting it : 1 .
the high ellipticity of the inner star forming region could be due to it being a flattened , disk - like component , different from the almost spherical , low - density , outer region .
the gas component is spatially much more compact than the outer stellar one .
this would be qualitatively compatible with a collapse that left behind the formed stars earlier .
3 . the external component lacks young stars , as it would be expected in an early formed structure which was later on deprived of gas .
these three results combine both age and morphological properties , thereby giving more self - consistency to the two - component hypothesis .
but all them are necessary rather than sufficient conditions . in particular , the sufficient conclusive conditions should be of the three following kinds : 1 .
kinematical , and not only morphological , decoupling should exist between the inner and the outer components .
radial velocities of the stars would be necessary to determine if this is the case , but they are beyond reach of current instruments .
the gas should be associated with the inner , flat component , rather than showing a spherical distribution .
it could be , however , that the sphericity of the gas originates in the heating produced by supernova explosions in the inner region .
3 . the stellar population of the outer component should be old ( older than @xmath85 10 gyr ) , the fact that young stars lack there not being enough . with the time resolution for intermediate and old ages of the analysis presented here
it is not possible to accept or reject the presence of an intermediate - age population with a high confidence level . summarizing
, the two - component hypothesis does not seem unrealistic . but nothing can be definitely stated until more detailed data , ideally including kinematics , are available . by now , this is the case for other dwarfs for which differentiated components have been sought ( antlia , ngc 3109 and with stronger clues , wlm and phoenix ; see references above ) .
for these , as for ddo 187 , more data are needed .
the sfr @xmath131 and the new estimate of the distance obtained here have be used , together with data from other authors , to calculate the integrated properties of ddo 187 listed in table [ global ] .
the bracketed figures refer to the bibliographic sources for each data . first the metallicity , @xmath117 , the distance to the milky way , @xmath135 , and the distance to the barycenter of the local group , @xmath136 , are given .
integrated absolute magnitudes and luminosities , referred to the holmberg radius at @xmath137arcsec@xmath138 ( which match the inner component of the former analysis ) follow .
@xmath139 is the mass in stars and stellar remnants for the whole galaxy , calculated from integration of @xmath131 and assuming that 0.8 of this integral remains locked in stellar objects .
@xmath140 is obtained by multiplying the h i mass by 4/3 to take into account the mass in he .
@xmath141 is the virial mass obtained from the velocity dispersion of the gas .
@xmath142 is the gas fraction , relative to the total mass intervening in the chemical evolution .
@xmath143 is the dark matter fraction , calculated as indicated in the table .
finally , the gas and total mass to luminosity fractions are given .
the hypothesis that dwarf star forming galaxies could have extended , primeval structures is tested for the dirr ddo 187 .
together with the sfh and the structure of ddo 187 , this paper provides a new estimate of the distance and discusses the distribution of h ii regions . from the @xmath0 magnitude of the tip of the red giant branch ( @xmath144 ) , the distance of ddo 187 to the milky way is estimated to be @xmath2 mpc .
the distance to several neighbor galaxies and groups have been computed , showing that it is an isolated , field galaxy : ddo 187 is at 2.6 mpc from m 94 ; at 2.7 mpc from the barycenter of the local group , and at 3.2 mpc from m 81 ; the closest dwarf galaxies to ddo 187 are gr 8 , at 0.85 mpc and ddo 190 , at 1.05 mpc .
the distance of ddo 187 to the milky way ( 2.5 mpc ) found here is quite different from the value of 7 mpc , estimated from cepheid light - curves by hoessel et al .
this is the third case ( pegasus , and leo a are the other two ) in which such a large disagreement is obtained between the two methods and implies that cepheid - based distances of dwarf galaxies must be accepted cautiously .
we suggest that the scarcity of cepheids in dwarf galaxies , derived from the small amount of young stars , would not make worthwhile the effort of determining their distance with cepheids .
ddo 187 has several h ii regions . from analysis of their fluxes
, it is found that the most massive star alive in the galaxy has a mass of some 30 m@xmath8 and that the current sfr is smaller by a factor of 3 than it was 20100 myr ago .
the sfh of ddo 187 has been obtained using synthetic cmds .
ddo 187 shows an overall decreasing sfr , but with high star formation activity over the last few hundred myr in its central region .
the sfr for different regions of the galaxy are summarized in table [ sfrave ] . in particular , the data analyzed here are compatible with a maximum in the sfr between 20 and 100 myr ago that could have had an intensity comparable to that of some blue compact galaxies .
interestingly , this picture of an overall decreasing sfr with a strong enhancement in the recent past is common to other dirr galaxies .
it is suggested that perhaps the dirr morphology is only produced by the recent past sfr activity and that no large intrinsic differences exist between typical dirrs and other dwarf galaxies showing similar intermediate- to old - age sfhs but neither h ii regions nor other traces of very recent star formation activity .
several results suggest that ddo 187 has a two - component halo / disk - like structure : ( i ) differentiated morphologies for the inner and outer stellar components ; ( ii ) a gas component less extended than the outer stellar component , and ( iii ) an outer component lacking young stars .
the working hypothesis that a real halo / disk structure could be present is discussed .
the conclusion is reached that the two - components hypothesis is not unrealistic , but that nothing can be clearly stated until more detailed data , ideally including kinematics , are available .
this paper has been largely improved after careful reading of the manuscript by carme gallart and david martnez - delgado and after the endless , fruitful discussions maintained with them .
we are also grateful to richard patterson and to the referee of the paper , albert zijlstra , for their useful comments and suggestions .
this work has been financially supported by the iac ( grants p3/94 and p90/92 ) , by the dges of the kingdom of spain ( grant pb97 - 1438-c02 - 01 ) and by intas - rfbr grant 95-in - ru-1930 .
data from ned have been used .
aparicio , a. 1994 , , 437 , l27 aparicio , a. 1999 , in the stellar content of local group galaxies .
i.a.u . symp .
192 , edited by p. whitelock aparicio , a. , dalcanton , j. j. , gallart , c. , & martnez - delgado , d. 1997c , , 114 , 1447 aparicio , a. , & gallart , c. 1994 , in the local group : comparative and global properties , eso conference and workshop proceedings no .
51 , edited by a. layden , r. c. smith , and j. storm ( eso garching ) , p. 115
aparicio , a. , & gallart , c. 1995 , , 110 , 2105 aparicio , a. , gallart , c. , & bertelli , g. 1997a , , 114 , 669 aparicio , a. , gallart , c. , & bertelli , g. 1997b , , 114 , 680 aparicio , a. , gallart , c. , chiosi , c. , & bertelli , g. 1996 , , 469 , l97 aparicio , a. , moles , a. , & garca - pelayo , j. m. 1988 , , 74 , 367 aparicio , a. , tikhonov , n. , & karachentsev , i. 1999 , in preparation bertelli , g. , bressan , a. , chiosi , c. , fagotto , f. , & nasi , e. 1994 , , 106 , 275 burstein , d. , & heiles , c. 1982 , , 87 , 65 dohm - palmer , r. c. , skillman , e. d. , gallagher , j. , tolstoy , e. , mateo , m. , dufour , r. j. , saha , a. , hoessel , j. , & chiosi , c. 1998 , , 116 , 1227 fanelli , m. n. , oconnel , r. w. , & thuan , t. x. 1988 , , 334 , 665 freedman , w. l. , & madore , b. f. 1990 , , 365 , 186 freedman , w. l. , hughes , s. m. , madore b. f. , mould , j. r. , lee , m. g. , stetson , p , kennicutt , r. c. , turner , a. , ferrarese , l. , ford , h. , graham , j. a. , hill , r. , hoessel , j. g. , huchra , j. , illingworth , g. d. 1994 , , 427 , 628 gallagher , j. s. , tolstoy , e. , dohm - palmer , r. c. , skillman , e. d. , cole , a. a. , hoessel , j. g. , saha , a. , & mateo , m. 1998 , , 115 , 1869 gallart , c. , freedman , w.l . , aparicio , a. , bertelli , g. , & chiosi , c. 1999 , , submitted gallart , c. , aparicio , a. , bertelli , g. , & chiosi , c. 1996b , , 112 , 1950 gallart , c. , aparicio , a. , bertelli , g. , & chiosi , c. 1996c , , 112 , 2596 gallart , c. , aparicio , a. , & vlchez , j. m. 1996a , , 112 , 1928 hodge , p. w. , dolphin , a. e. , smith , t. r. , & mateo , m. , 1999 , , in press hoessel , j. g. , abbott , m. j. , saha , a. , mossman , a. e. , & danielson , g. e. 1990 , , 100 , 1151 hoessel , j. g. , saha , a. , & danielson , g. e. 1998 , , 116 , 1679 hoessel , j. g. , saha , a. , krist , j. , & danielson , g. e. 1994 , , 108 , 645 kennicutt , r. c. 1988 , , 334 , 144 kroupa , p. , tout , c. a. , & gilmore , g. 1993 , , 262 , 545 landolt , a. u. 1992 , , 104 , 340 lee , m. g. , freedman , w. l. , & madore , b. f. 1993 , , 417 , 553 lo , k. y. , sargent , w. l. w. , & young , k. 1993 , , 106 , 507 martnez - delgado , d. , gallart , c. , & aparicio , a. 1999 , , in press minniti , d. , zijlstra , a. a. 1996 , , 467 , l13 minniti , d. , zijlstra , a. a. 1997 , , 114 , 147 minniti , d. , zijlstra , a. a. , & alonso , m. v. 1999 , , 117 , 881 mulder , p. s. , & van driel , w. 1993 , , 272 , 63 oke , j. b. 1990 , , 99 , 1621 panagia , n. 1973 , , 78 , 929 patterson , r. j. , & thuan , t. x. 1996 , , 107 , 103 patterson , r. j. , & thuan , t. x. 1998 , , 117 , 633 peebles , p. j. e. 1989 , , 344 , l53 saha , a. , hoessel , j. g. , krist , j. , & danielson , g. e. 1996 , , 111 , 197 skillman , e. d. , kennicutt , r. c. , & hodge , p. h. 1989 , , 347 , 875 stetson , p. b. 1994 , , 106 , 250 stetson , p. b. , saha , a. , ferrarese , l. , _ et al .
_ 1998 , , 508 , 491 tolstoy , e. , gallagher , j. s. cole , a. a. , hoessel , j. g. , saha , a. , dhom - plamer , r. c. , skillman , e. d. , mateo , m. , & hurley - keller , d. 1998 , , 116 , 1244 van den bergh , s. 1959 , pub . david dunlap observatory , vol .
ii , n. 5 van zee , l. , haynes , m. p. , & salzer , j. j. 1997a , , 114 , 2479 van zee , l. , maddalena , r. j. , haynes , m. p. , hogg , d. e. , & roberts , m. s. 1997b , , 113 , 1638 white , s. d. m. , & rees , m. j. 1978 , , 183 , 341 young , l. m. , & lo , k. y. 1997 , , 490 , 710 ccccc 97.07.26 & ddo 187 & 21:50 & @xmath0 & 600 97.07.26 & ddo 187 & 22:01 & @xmath0 & 600 97.07.26 & ddo 187 & 22:12 & @xmath12 & 900 97.07.26 & ddo 187 & 22:35 & @xmath12 & 900 97.07.27 & ddo 187 & 21:46 & @xmath12 & 900 97.07.27 & ddo 187 & 22:03 & @xmath12 & 900 97.07.27 & ddo 187 & 22:18 & @xmath0 & 900 97.07.27 & ddo 187 & 22:35 & @xmath0 & 900 97.07.29 & ddo 187 & 21:04 & @xmath13 & 900 97.07.29 & ddo 187 & 21:19 & @xmath13 & 300 97.07.29 & ddo 187 & 21:25 & @xmath13-cont & 400 97.07.29 & field & 22:20 & @xmath12 & 600 97.07.29 & field & 22:32 & @xmath12 & 600 97.07.29 & field & 22:43 & @xmath0 & 500 97.07.29 & field & 22:51 & @xmath0 & 500 ddo 187 & & @xmath145 & 2 & ( 6 ) & & @xmath2 & trgb & ( 1 ) leo a ( ddo 69 ) & & @xmath146 & 4 & ( 5 ) & & @xmath147 & rc & ( 7 ) & & & & & & @xmath148 & trgb & ( 7 ) pegasus ( ddo 216 ) & & @xmath149 & 5 & ( 4 ) & & @xmath150 & trgb & ( 2 ) & & & & & & @xmath151 & trgb & ( 3 ) cccccc 1 ( c ) & 4.7 & 13.79 & 10.32 & 7.53 & 31.5 2 ( c ) & 3.8 & 4.10 & 3.07 & 2.23 & 22.0 3 ( c ) & 4.1 & 3.16 & 2.37 & 1.73 & 20.5 4 ( c ) & 6.3 & 0.58 & 0.43 & 0.31 & 14.0 5 ( c ) & 5.6 & 0.41 & 0.30 & 0.22 & 13.5 6 ( c ) & 5.1 & 0.74 & 0.55 & 0.40 & 14.5 7 ( c ) & 6.7 & 0.82 & 0.61 & 0.44 & 14.8 8 ( d ) & 39.2 & 21.45 & 16.06 & 11.72 & 9 ( d ) & 27.5 & 6.69 & 5.00 & 3.65 & 10 ( d ) & 18.5 & 0.58 & 0.43 & 0.31 & total & & 78.2 & 58.5 & 42.7 & @xmath152(@xmath153m@xmath8yr@xmath6 ) & 3.4 & 1.3 & 0.9 @xmath154 ( @xmath153m@xmath8yr@xmath6 ) & 3.1 & 1.3 & 0.9 @xmath155 ( @xmath153m@xmath8yr@xmath6 ) & 7.6 & 1.2 & @xmath156 ( @xmath153m@xmath8yr@xmath6 ) & 50 & 2.2 & 0 @xmath157 ( @xmath153m@xmath8yr@xmath6 ) & 17.5 & 0 & 0 @xmath158 ( @xmath159m@xmath8yr@xmath6pc@xmath138 ) & 0.59 & 0.16 & 0.018 @xmath160 ( @xmath159m@xmath8yr@xmath6pc@xmath138 ) & 0.53 & 0.16 & 0.018 @xmath161 ( @xmath159m@xmath8yr@xmath6pc@xmath138 ) & 1.29 & 0.15 & @xmath162 ( @xmath159m@xmath8yr@xmath6pc@xmath138 ) & 8.41 & 0.28 & 0 @xmath157 ( @xmath153m@xmath8yr@xmath6 ) & 2.93 & 0 & 0 @xmath163 , @xmath164 & @xmath165 , @xmath166 @xmath167 & @xmath168 @xmath117 & 0.0008 ( 6,7 ) @xmath169 ( mpc ) & 2.5 ( 1 ) @xmath136 ( mpc ) & 2.7 ( 1 ) @xmath170 & 13.07 ( 5 ) @xmath171 & 13.28 ( 4 ) @xmath172 ( @xmath173l@xmath8 ) & 21 @xmath174 ( @xmath173l@xmath8 ) & 15 @xmath175 ( @xmath173l@xmath8 ) & 8.5 @xmath139 ( @xmath173m@xmath8 ) & 8.4 ( 1 ) @xmath140 ( @xmath173m@xmath8 ) & 26 ( 2,3,8 ) @xmath141 ( @xmath173m@xmath8 ) & 102 ( 2,3 ) @xmath176 & 0.76 @xmath177 & 0.66 @xmath178 ( m@xmath8/l@xmath8 ) & 1.7 @xmath179 ( m@xmath8/l@xmath8 ) & 6.8 | if dwarf galaxies are primeval objects in the universe , as hierarchical galaxy formation scenarios predict , they should show traces of their old stellar populations , perhaps distributed in extended , differentiated structures .
the working hypothesis that such a structure could exist is tested for the case of ddo 187 , a field dirr galaxy showing a high gas fraction and low metallicity . for this purpose ,
the structure , star formation history ( sfh ) and other properties of the galaxy are analyzed using the spatial distribution of stars , the color
magnitude diagrams ( cmds ) of about 1500 resolved stars and the fluxes of h ii regions , together with data about the gas distribution . from the @xmath0 magnitude of the tip of the red giant branch ( @xmath1 ) , the distance of ddo 187 to the milky way is estimated to be @xmath2 mpc .
the distance to several neighbor galaxies and groups has been computed , showing that ddo 187 is probably an isolated , field galaxy .
the distance of ddo 187 to the milky way is almost three times smaller than that obtained from cepheid light - curves .
considering that this is the third case in which such a large disagreement is detected , it seems clear that cepheid distance estimates based on a few stars , as usually happens in dwarf galaxies , must be accepted with caution .
the star formation history of ddo 187 has been analyzed .
the central region of ddo 187 shows an overall time decreasing star formation rate with a strong burst in its central region which happened between 20 and 100 myrs ago and a present - day star formation activity three times smaller than the maximum one .
besides this , a spatially extended stellar component has been found that has no young stars and exceeds the size of the gas component . in short
, several results suggest that ddo 187 has a two - component halo / disk - like structure : ( i ) differentiated morphologies for the inner ( flat ) and outer ( spheroidal ) stellar components ; ( ii ) a gas component less extended than the outer stellar component , and ( iii ) an outer component lacking young stars , which are abundant in the inner component .
the working hypothesis that a real halo / disk structure could be present is discussed .
the conclusion is reached that the two - component hypothesis is not unrealistic , but nothing can be definitely stated until more detailed data , ideally including kinematics , are available . |
in this work , we study reaction - diffusion pde systems @xmath0 as well as their discrete analogues ( `` compartmental - systems '' ) . here , @xmath1 @xmath2 is the laplacian operator on a suitable spatial domain @xmath3 , and no flux ( neumann ) boundary conditions are assumed . in biology , a pde system of this form describes individuals ( particles , chemical species , etc . ) of @xmath4 different types , with respective abundances @xmath5 at time @xmath6 and location @xmath7 , that can react instantaneously , guided by the interaction rules encoded into the vector field @xmath8 , and can diffuse due to random motion .
reaction - diffusion pde s play a key role in modeling intracellular dynamics and protein localization in cell processes such as cell division and eukaryotic chemotaxis ( e.g. , @xcite ) as well as in the modeling of differentiation in multi - cellular organisms , through the diffusion of morphogens which control heterogeneity in gene expression in different cells ( e.g. @xcite ) . from a bioengineering perspective
, reaction - diffusion models can be used to model artificial mechanisms for achieving cellular heterogeneity in tissue homeostasis ( e.g. , @xcite ) .
the `` symmetry breaking '' phenomenon of diffusion - induced , or turing , instability refers to the case where a dynamic equilibrium @xmath9 of the non - diffusing ode system @xmath10 is stable , but , at least for some diagonal positive matrices @xmath11 , the corresponding uniform state @xmath12 is unstable for the pde system @xmath13 .
this phenomenon has been studied at least since turing s seminal work on pattern formation in morphogenesis @xcite , where he argued that chemicals might react and diffuse so as result in heterogeneous spatial patterns .
subsequent work by gierer and meingardt @xcite produced a molecularly plausible minimal model , using two substances that combine local autocatalysis and long - ranging inhibition .
since that early work , a variety of processes in physics , chemistry , biology , and many other areas have been studied from the point of view of diffusive instabilities , and the mathematics of the process has been extensively studied @xcite .
most past work has focused on local stability analysis , through the analysis of the instability of nonuniform spatial modes of the linearized pde .
nonlinear , global , results are usually proved under strong constraints on diffusion constants as they compare to the growth of the reaction part . in this note
, we are interested in conditions on the reaction part @xmath8 that guarantee that no diffusion instability will occur , no matter what is the size of the diffusion matrix @xmath11 . we show that if the reaction system is `` contractive '' in the sense that trajectories globally and exponentially converge to each other with respect to a diagonally weighted @xmath14 norm , then the same property is inherited by the pde . in particular ,
if there is an equilibrium @xmath15 , it will follow that this equilibrium is globally exponentially stable for the pde system .
a similar result is also established for a discrete analog , in which a set of ode systems are diffusively interconnected .
we were motivated by the desire to understand the important biological systems described in @xcite for which , as we will show , contractivity holds for diagonally weighted @xmath16 norms , but not with respect to diagonally weighted @xmath14 norms , for any @xmath17 .
closely related work in the literature has dealt with the synchronization problem , in which one is interested in the convergence of trajectories to their space averages in weighted @xmath18 norms , for appropriate diffusion coefficients and laplacian eigenvalues , specifically @xcite , which used passivity ideas from control theory for systems with special structures such as cyclic systems , @xcite which extended this approach to more general passive structures , and @xcite which obtained a generalization involving a contraction - like diagonal stability condition .
our work uses very different techniques , from nonlinear functional analysis for normed spaces , than the quadratic lyapunov function approaches , appropriate for hilbert spaces , followed in these references .
we start by reviewing several useful concepts from nonlinear functional analysis , and proving certain technical properties for them .
@xcite let @xmath19 be a normed space . for @xmath20 ,
the right and left semi inner products are defined by @xmath21 [ existence - of - norm ] as every norm possesses left and right gateaux - differentials , the limits in @xmath22 exist and are finite . for more details
see @xcite .
the right and left semi inner products @xmath23 , induce the norm @xmath24 in the usual way : @xmath25 .
conversely if the norm arises from an inner product @xmath26 , as when @xmath27 is a hilbert space , @xmath28 .
moreover the right and left semi inner products satisfy the cauchy - schwarz inequalities : @xmath29 the following elementary properties of semi inner products are consequences of the properties of norms .
see @xcite for the proof . for @xmath30 and @xmath31 , 1 .
@xmath32 ; 2 .
@xmath33 ; 3 . @xmath34 .
in general , the semi inner product is not symmetric : @xmath35 @xcite let @xmath19 be a normed space and @xmath36 be a function , where @xmath37 .
the strong least upper bound logarithmic lipschitz constants of @xmath38 induced by the norm @xmath24 , on @xmath39 , are defined by @xmath40=\displaystyle\sup_{u\neq v\in y}\frac{(u - v , f(u)-f(v))_{\pm}}{\|u - v\|_x^2},\ ] ] or equivalently @xmath41=\displaystyle\sup_{u\neq v\in y}\lim_{h\to0^{\pm}}\frac{1}{h}\left(\frac{\|u - v+h(f(u)-f(v))\|_x}{\|u - v\|_x}-1\right).\ ] ] if @xmath42 , we write @xmath43 instead of @xmath44 .
[ subadd ] let @xmath19 be a normed space .
for any @xmath38 , @xmath45 and any @xmath37 : 1 .
@xmath46\leq m_{y , x}^+[f]+m_{y , x}^+[g]$ ] ; 2 .
@xmath47=\alpha m_{y , x}^{\pm}[f]$ ] for @xmath48 . 1 . by the definition of @xmath49 , and the triangle inequality for norms , we have @xmath50&=&\displaystyle\sup_{u\neq v\in y}\lim_{h\to0^+}\frac{1}{h}\left(\frac{\|u - v+h((f+g)(u)-(f+g)(v))\|_x}{\|u - v\|_x}-1\right)\\ & = & \displaystyle\sup_{u\neq v\in y}\lim_{h\to0^+}\frac{1}{2h}\left(\frac{\|2(u - v)+2h((f+g)(u)-(f+g)(v))\|_x}{\|u - v\|_x}-2\right)\\ & \leq&\displaystyle\sup_{u\neq
v\in y}\lim_{h\to0^+}\frac{1}{2h}\displaystyle\left(\frac{\|u - v+2h(f(u)-f(v))\|_x}{\|u - v\|_x}-1\right)+\\ & & \displaystyle\sup_{u\neq v\in y}\lim_{h\to0^+}\frac{1}{2h}\displaystyle\left(\frac{\|u - v+2h(g(u)-g(v))\|_x}{\|u - v\|_x}-1\right)\\ & = & m_{y , x}^+[f]+m_{y , x}^+[g ] \end{array}\ ] ] 2 . for @xmath51 ,
the equality is trivial , because both sides are equal to zero . for @xmath52 : @xmath53&=&\displaystyle\sup_{u\neq v\in y}\lim_{h\to0^{\pm}}\frac{1}{h}\left(\frac{\|u - v+h(\alpha f(u)-\alpha f(v))\|_x}{\|u - v\|_x}-1\right)\\ & = & \displaystyle\sup_{u\neq v\in y}\lim_{h\to0^{\pm}}\frac{\alpha}{\alpha h}\left(\frac{\|u - v+(\alpha h)(f(u)-f(v))\|_x}{\|u - v\|_x}-1\right)\\ & = & \alpha m_{y , x}^{\pm}[f ] .
\end{array}\ ] ] @xcite let @xmath19 be a normed space and @xmath54 be a function , where @xmath37 .
the least upper bound lipschitz constant of @xmath38 induced by the norm @xmath24 , on @xmath39 , is defined by @xmath55=\displaystyle\sup_{u\neq v\in y}\frac{\|f(u)-f(v)\|_x}{\|u - v\|_x}.\ ] ] note that @xmath56<\infty$ ] if and only if @xmath38 is lipschitz on @xmath39 .
@xcite let @xmath19 be a normed space and @xmath36 be a lipschitz function .
the least upper bound logarithmic lipschitz constant of @xmath38 induced by the norm @xmath24 , on @xmath37 , is defined by @xmath57=\displaystyle\lim_{h\to0+}\frac{1}{h}\left(l_{y , x}[i+hf]-1\right),\ ] ] or equivalently @xmath58=\displaystyle\lim_{h\to0^+}\sup_{u\neq v\in y}\frac{1}{h}\left(\frac{\|u - v+h(f(u)-f(v))\|_x}{\|u - v\|_x}-1\right)\ ] ] if @xmath42 , we write @xmath59 instead of @xmath60 .
[ m+<m ] @xmath61\leq m_{y , x}[f]$ ] . for any fixed @xmath62 , @xmath63
now using this inequality , we have : @xmath64 since this inequality holds for any @xmath62 , taking @xmath65 we have : @xmath66 from which the conclusion follows using @xmath67 and @xmath68 . the least upper bound ( lub ) logarithmic lipschitz constant generalizes the usual logarithmic norm ; for every matrix @xmath69 we have @xmath70 = \mu_x[a]$ ] . for ease of references , we review next the basic properties of logarithmic norms for finite dimensional operators .
let @xmath19 be a finite dimensional normed vector space over @xmath71 or @xmath72 .
the space @xmath73 of linear transformations @xmath74 is also a normed vector space with the induced operator norm @xmath75 the logarithmic norm @xmath76 induced by @xmath24 is defined as the directional derivative of the matrix norm , that is , @xmath77 where @xmath78 is the identity operator on @xmath27 . since @xmath79 , whenever @xmath80 and @xmath81
, it follows that @xmath82 [ mu ] for any matrix @xmath69 , @xmath83 .
see the appendix for the proof .
[ mu = m+ ] let @xmath84 be a finite dimensional normed space . for any linear operator @xmath85 , @xmath86=m_x[\mathcal{l}].\ ]
] the proof is immediate from the definition of @xmath59 , @xmath87 , and theorem [ mu ] . when @xmath88 or @xmath89 , we identify operators and their matrix representations on the standard basis , and we call the logarithmic norm the _ matrix measure_. in table @xmath90 , the algebraic expression of the logarithmic norms induced by the @xmath14 norm for @xmath91 and @xmath92 are shown for matrices . for proofs , see for instance @xcite . .standard matrix measures for a real @xmath93 matrix , @xmath94 $ ] .
[ cols="^,^",options="header " , ] [ tab - mu ] for ease of reference , we summarize the main notations and definitions in table [ tab - def ] .
[ tab - def ]
suppose @xmath95 , a bounded domain in @xmath96 with smooth boundary @xmath97 and outward normal @xmath98 , and a subset @xmath99 have been fixed .
we denote @xmath100 where @xmath101 is the set of twice continuously differentiable functions @xmath102 .
in addition , we denote @xmath103 , where @xmath104 is the set of all continuous functions @xmath105 . note that for each @xmath106 , for @xmath107 , @xmath108 and for @xmath109 , @xmath110 and both are finite because @xmath111 is a continuous function on @xmath112 and @xmath112 is a compact subset of @xmath96 . for any @xmath113 , and any nonsingular , diagonal matrix @xmath114 , we introduce a @xmath115-weighted norm on @xmath116 as follows : @xmath117 since @xmath118 without loss of generality we will assume @xmath119 for each @xmath106 . with a slight abuse of notation , we use the same symbol for a norm in @xmath120 : @xmath121 [ u = s ] for any @xmath122 , @xmath123 , where @xmath124 note that @xmath125 and @xmath126 .
let @xmath127 , @xmath119 . for @xmath107 (
the proof is analogous when @xmath109 ) , by the definitions of @xmath128 and @xmath129 @xmath130 note that this equality between weighted @xmath131 norms of functions and of vectors depends on our having taken the matrix @xmath115 to be diagonal .
this is the key place where the assumption that @xmath115 is diagonal is being used .
in this section , we study the reaction - diffusion pde : @xmath132 subject to the neumann boundary condition : @xmath133 [ as - pde ] in @xmath134 we assume : * @xmath135 is a ( globally ) lipschitz and twice continuously differentiable vector field with components @xmath136 : @xmath137 for some functions @xmath138 , where @xmath139 is a convex subset of @xmath120 .
* @xmath140 , with @xmath141 , is called the diffusion matrix .
* @xmath95 is a bounded domain in @xmath96 with smooth boundary @xmath97 and outward normal @xmath98 . by a solution of the pde @xmath142 @xmath143 on an interval @xmath144 , where @xmath145 , we mean a function @xmath146 , with @xmath147 , such that : 1 . for each @xmath148 , @xmath149 is continuously differentiable ; 2 . for each @xmath150
, @xmath151 is in @xmath152 ; and 3 . for each @xmath148 , and each @xmath153 , @xmath154 satisfies the above pde .
theorems on existence and uniqueness for pde s such as @xmath134 can be found in standard references , e.g. @xcite .
one must impose appropriate conditions on the vector field , on the boundary of @xmath139 , to insure invariance of @xmath139 .
convexity of @xmath139 insures that the laplacian also preserves @xmath139 .
since we are interested here in estimates relating pairs of solutions , we will not deal with existence and well - posedness .
our results will refer to solutions already assumed to exist .
pick any @xmath145 and suppose that @xmath154 is a solution of @xmath134 defined on @xmath155 .
define @xmath156 by @xmath157 . also define the function @xmath158 as follows : for any @xmath159 , @xmath160 let @xmath161 denote an @xmath93 diagonal matrix of operators on @xmath152 with the operators @xmath162 on the diagonal .
suppose that @xmath154 solves the pde @xmath134 , on an interval @xmath163 , for some @xmath164 $ ] , and let @xmath165 for each @xmath166 and @xmath148 .
we introduce @xmath167 by @xmath168 then , @xmath169 is the derivative of @xmath170 in the space @xmath171 , that is : @xmath172 - \hat v(t ) \right\vert}_{p , q } } \;=\ ; 0,\ ] ] for all @xmath153 .
moreover , @xmath173 fix @xmath174 and @xmath175 . using the definition of @xmath176
, we have : @xmath177 -v_i(\omega , t)\right|=0,\ ] ] for any @xmath148 .
hence for any @xmath178 , there exists @xmath179 such that for any @xmath180 , @xmath181 -v_i(\omega , t)\right|<\frac{\epsilon}{2}.\ ] ] now since @xmath182 is a continuous function of @xmath183 , there exists a ball @xmath184 centered at @xmath183 such that for any @xmath180 , @xmath185 -v_i(\tilde\omega , t)\right|<\epsilon\ ] ] for all @xmath186 . since @xmath187 is an open cover of @xmath112 and @xmath112 is a compact subset of @xmath96 , finitely many of these balls , namely @xmath188 , cover @xmath112 .
now let @xmath189 .
then , for any @xmath190 and any @xmath148 , we have @xmath181 -v_i(\omega , t)\right|<\epsilon.\ ] ] raising to the @xmath131-th power and taking the integral over @xmath95 of the above inequality , we get @xmath191 -v_i(\omega , t)\right|^p\;d\omega<\abs{\omega}\epsilon^p,\ ] ] which by the definition of @xmath192 , it implies that for any @xmath193 , @xmath194 -v(\cdot , t ) \right\vert}_{p , q}}<c\epsilon,\ ] ] where @xmath195 . since @xmath178 was arbitrary , we have proved that @xmath196 - \hat v(t ) \right\vert}_{p , q } } = 0.\ ] ] for a fixed @xmath150 and any @xmath148 : @xmath197 and therefore equation ( [ pde - ode ] ) holds . in this section
we show that @xmath134 is contracting ( meaning that solutions converge exponentially to each other , as @xmath198 ) if @xmath199<0 $ ] , where , as defined before , @xmath200=\displaystyle\lim_{h\to0^+}\sup_{x\neq y\in v}\frac{1}{h}\left(\frac{{{\left\vert x - y+h(f(x)-f(y ) ) \right\vert}_{p , q}}}{{{\left\vert x - y \right\vert}_{p , q}}}-1\right).\ ] ] now we state the main result of this section .
[ main - result0 ] consider the pde @xmath134 and suppose assumption [ as - pde ] holds .
let @xmath201 $ ] for some @xmath113 , and some positive diagonal matrix @xmath115 .
then for every two solutions @xmath202 of the pde @xmath203 and all @xmath150 : @xmath204 in terms of the pde @xmath134 , this last estimate can be equivalently written as : @xmath205 before proving the theorem , we prove a few technical lemmas .
the upper left and right dini derivatives for any continuous function , @xmath206 , are defined by @xmath207 note that @xmath208 and/or @xmath209 might be infinite .
[ dini ] let @xmath210 .
let @xmath211 be a ( globally ) lipschitz function , where @xmath37 .
let @xmath212 be two solutions of @xmath213 .
then for all @xmath214 , @xmath215 when @xmath216 , we understand the right hand side through the limit in @xmath217 . by the definition of right semi - inner product ,
the right hand side of @xmath218 is : @xmath219 hence we just need to show that @xmath220 now using the definition of dini derivative , we have : @xmath221 where @xmath222 .
note that the fourth equality holds because of remark @xmath223 .
[ key-1 ] under the assumptions of lemma @xmath224 , for any @xmath214 we have : @xmath225\|u(t)-v(t)\|_x.\ ] ] by the definition of the strong least upper bound logarithmic lipschitz constant , @xmath226.\ ] ] now apply lemma @xmath224 to the above inequality .
[ key1 ] under the assumptions of lemma @xmath224 , for any @xmath214 we have : @xmath227t}\|u(0)-v(0)\|_x.\ ] ] apply gronwall s inequality , @xcite , to @xmath228 .
[ m+dini - linear ] note that for any bounded linear operator @xmath85 , and any solution @xmath229 of @xmath230 , the above result says that @xmath231\|u(t)\|_x,\ ] ] for all @xmath232 .
[ key4 ] let @xmath233 , as defined above , denote an @xmath93 diagonal matrix of operators on @xmath152 with the operators @xmath162 on the diagonal
. then @xmath234\leq 0 $ ] . to prove the lemma
, we consider the following three cases : * case 1 . *
@xmath235 . by the definition of @xmath234 $ ] , it s enough to show that for any @xmath159 with @xmath236 , and any @xmath178 , there exists @xmath237 , depending on @xmath238 , such that for @xmath239 , @xmath240 ( as @xmath241 , we write @xmath242 instead of @xmath243 ) therefore we ll show that for @xmath244 small enough @xmath245 let s define @xmath246\to\r$ ] as follows : @xmath247 observe that @xmath248 is continuously differentiable : @xmath249 note that in general @xmath250 is differentiable for @xmath251 and its derivative is @xmath252 . now by green s identity , the neumann boundary condition , and by the assumption that @xmath253 , it follows integrating by parts that : @xmath254 since @xmath255 and @xmath256 is continuous and @xmath257 , @xmath258 for @xmath244 small enough and therefore inequality @xmath259 holds . note
that by the definition of @xmath152 , any @xmath159 satisfies the neumann boundary condition .
* case 2 . * @xmath260 .
let @xmath261 since @xmath262 is a continuous function at @xmath260 , and since in case @xmath263 , we showed that @xmath264 for any @xmath251 , we conclude that @xmath265 . * case 3 . * @xmath109 . before proving this case we need the following lemma , which is an easy exercise in real analysis .
( for completeness , we include a proof in an appendix . )
[ p - limit ] let @xmath266 be a lebesgue measurable set with finite measure @xmath267 and let @xmath38 be a bounded , continuous function on @xmath71 . then @xmath268 is an increasing function of @xmath131 and its limit as @xmath269 is @xmath270 .
for a fixed @xmath271 , pick @xmath159 with @xmath272 . by the definition of the norm , @xmath272 implies that for some @xmath273 , @xmath274 .
let @xmath275 by lemma [ p - limit ] , @xmath276 is an increasing function of @xmath131 , hence for any @xmath277 , @xmath278 .
now fix @xmath175 , @xmath277 , and @xmath178 .
define @xmath248 as follows : @xmath279 in both cases @xmath280 and @xmath255 ( the proof is similar to the proof of @xmath255 in case @xmath263 , since both @xmath281 and @xmath281 ) .
therefore , for some small @xmath244 , @xmath282 which implies that : @xmath283 now by lemma [ p - limit ] , since @xmath284 , and @xmath285 as @xmath269 , we can conclude that @xmath286 in other words , for a fixed @xmath178 , there exists @xmath287 such that for any @xmath288 , @xmath289 let @xmath290 .
then for any @xmath193 , @xmath291 which implies @xmath292 [ mm ] for any function @xmath8 , any @xmath113 , and any positive diagonal matrix @xmath115 , @xmath293\leq { m_{p , q}}[f],\ ] ] where @xmath294 is the lub logarithmic lipschitz constant induced by the norm @xmath295 defined on @xmath120 : @xmath296 . by the definition of @xmath297 $ ] , we have @xmath298 fix an arbitrary @xmath178 .
then there exists @xmath299 such that for all @xmath193 , @xmath300 therefore , for any @xmath301 , and @xmath193 @xmath302 for fixed @xmath303 , let @xmath304 .
fix @xmath305 , and let @xmath306 and @xmath307 .
we give a proof for the case @xmath308 ; the case @xmath109 is analogous . using equation @xmath309
, we have : @xmath310 multiplying both sides by the denominator and raising to the power @xmath131 , we have : @xmath311 since @xmath312 , equation ( [ forw ] ) can be written as : @xmath313 now by taking the integral over @xmath112 , using lemma @xmath314 , we get : @xmath315 ( note that for @xmath316 , @xmath317 which we can add to the right hand side of @xmath318 , and also @xmath319 which we can add to the left hand side of @xmath318 , and hence we can indeed take the integral over all @xmath112 . ) hence , @xmath320 now by letting @xmath321 and taking @xmath65 over @xmath303 , we get @xmath322\leq c$ ] . for any @xmath113 , by subadditivity of semi inner product , lemma [ key4 ] , and lemma [ mm ] , @xmath323\leq m^+_{\mathbf{y } , \mathbf{x}}[\tilde{f}]\leq c.\ ] ] now using corollary [ key1 ] , @xmath324 for all @xmath325 [ thm - for - mu ] consider the reaction - diffusion system @xmath134 and suppose assumption [ as - pde ] holds .
in addition suppose for some @xmath113 , @xmath326 , and a positive diagonal matrix @xmath115 , @xmath327 for all @xmath328 , where @xmath329 is the logarithmic norm induced by @xmath128 .
then , for any two solutions @xmath202 of @xmath203 , we have @xmath330 [ tv ] in general the result of theorem [ thm - for - mu ] holds also for time varying systems : @xmath331 when we assume @xmath332 , where @xmath333 is the jacobian @xmath334 , for all @xmath328 and @xmath335 we omit the details of this easy generalization . to prove theorem [ thm - for - mu ]
, we use the following proposition , from @xcite .
[ key3 ] let @xmath84 be a normed space and @xmath39 is a connected subset of @xmath27 .
then for any ( globally ) lipschitz and continuously differentiable function @xmath336 , @xmath337.\ ] ] moreover if @xmath39 is convex , then @xmath338.\ ] ] the proof is immediate from theorem @xmath339 and proposition @xmath340 .
[ contracting ] consider the reaction - diffusion system @xmath134 and suppose assumption [ as - pde ] holds . in addition suppose for some @xmath113 , and a positive diagonal matrix @xmath115 , @xmath341 for all @xmath328 .
then @xmath203 is contracting in @xmath152 , meaning that solutions converge ( exponentially ) to each other , as @xmath198 .
we provide an example of a biochemical model which can be shown to be contractive by applying corollary [ contracting ] when using a weighted @xmath16 norm , but which is not contractive using any weighted @xmath18 norm , so that previous results can not be applied .
even more interestingly , this system is not contractive in any @xmath14 norm , @xmath251 .
the example is of great interest in molecular systems biology @xcite , and contractivity in a weighted @xmath16 norm was shown for ode systems in @xcite , but the pde case was open .
the variant with more enzymes discussed in @xcite can also be extended to the pde case in an analogous fashion .
* example @xmath263*. a typical biochemical reaction is one in which an enzyme @xmath27 ( whose concentration is quantified by the non - negative variable @xmath342 ) binds to a substrate @xmath343 ( whose concentration is quantified by @xmath344 ) , to produce a complex @xmath39 ( whose concentration is quantified by @xmath345 ) , and the enzyme is subject to degradation and dilution ( at rate @xmath346 , where @xmath347 ) and production according to an external signal @xmath348 , which we assume constant ( a similar result would apply if @xmath348 is time dependent , see remark [ tv ] ) . an entirely analogous system can be used to model a transcription factor binding to a promoter , as well as many other biological process of interest .
the complete system of chemical reactions is given by : @xmath349{k_2 } } y.\ ] ] we let the domain @xmath95 represent the part of the cytoplasm where these chemicals are free to diffuse .
taking equal diffusion constants for @xmath343 and @xmath39 ( which is reasonable since typically @xmath343 and @xmath39 have approximately the same size ) , a natural model is given by a reaction diffusion system @xmath350 if we assume that initially @xmath343 and @xmath39 are uniformly distributed , it follows that @xmath351 , so @xmath352 is a constant .
thus we can study the following reduced system : @xmath353 note that @xmath354 $ ] for all @xmath166 ( @xmath139 is convex ) , and @xmath355 , @xmath356 , @xmath357 , @xmath358 , @xmath359 , and @xmath360 are arbitrary positive constants .
let @xmath361 be the jacobian of @xmath362 : @xmath363 in @xcite , it has been shown that @xmath364 for @xmath365 , where @xmath366 .
therefore by corollary [ contracting ] , the system is contracting .
note that a @xmath367 norm @xmath16 is necessary , since with @xmath368 we obtain @xmath369 we ll show that for any @xmath251 and any diagonal @xmath115 , it is not true that @xmath370 for all @xmath371 . without loss of generality
we assume @xmath372 .
then @xmath373 where @xmath374 $ ] and @xmath375 .
we first consider the case @xmath376 .
we ll show that there exists @xmath377 such that for any small @xmath378 , @xmath379 .
this will imply @xmath380 computing explicitly , we have : @xmath381 where we take a point of the form @xmath382 , for a @xmath383 which will be determined later . to show @xmath384 we ll equivalently
show that for any small enough @xmath378 : @xmath385 note that the @xmath386 of the left hand side of the above inequality is @xmath387 where @xmath388 therefore it suffices to show that @xmath389 for some value @xmath377 ( because @xmath389 implies that there exists @xmath299 such that for @xmath193 , @xmath390 holds ) .
since @xmath251 , by assumption , @xmath38 is differentiable and @xmath391 hence , since @xmath383 @xmath392 choosing @xmath393 small enough such that @xmath394 and choosing @xmath395 , or equivalently @xmath396 , large enough , we can make @xmath389 . for @xmath109 , using table [ tab - mu ] , @xmath397 . for large enough @xmath395 , @xmath398 ( and @xmath399 ) and hence @xmath400 .
in this section , we derive a result analogous to that for pde s for a network of identical ode models which are diffusively interconnected .
we study systems of ode s as follows : @xmath401 [ as - ode ] in @xmath402 , we assume : * for a fixed convex subset of @xmath120 , say @xmath139 , @xmath403 is a function of the form : @xmath404 where @xmath405 , with @xmath406 for each @xmath106 , and @xmath135 is a ( globally ) lipschitz function . * for any @xmath407
we define @xmath408 as follows : @xmath409 where @xmath410 is a positive diagonal matrix and @xmath113 .
+ with a slight abuse of notation , we use the same symbol for a norm in @xmath120 : @xmath121 * @xmath411 is a continuously differentiable function . *
@xmath412 with @xmath141 , which we call the diffusion matrix .
* @xmath413 is a symmetric matrix and @xmath414 , where @xmath415 .
we think of @xmath416 as the laplacian of a graph that describes the interconnections among component subsystems .
[ ode ] consider the system @xmath402 and suppose assumption [ as - ode ] holds
. let @xmath201 $ ] , where @xmath294 is the lub logarithmic lipschitz constant induced by the norm @xmath295 on @xmath120 defined by @xmath417 .
then for any two solutions @xmath202 of @xmath402 , we have @xmath418 this theorem is proved by following the same steps as in the pde case and using lipschitz norms and properties of discrete laplacians on finite graphs . for odes
, we can make some of the steps more explicit , and for purposes of exposition , we do so next .
we start with several technical lemmas .
the following elementary property of logarithmic norms is well - known . to see more properties of logarithmic norm
see @xcite .
[ mu - prop ] let @xmath393 be the largest real part of an eigenvalue of @xmath69 . then , @xmath419 . [ negm+p ] for any @xmath113 , @xmath420 , where @xmath421 is the strong least upper bound logarithmic lipschitz constant induced by the @xmath14 norm .
let @xmath422 .
note that since @xmath414 , by the definition of kronecker product , @xmath423 .
in addition because @xmath416 is symmetric and @xmath11 is diagonal , @xmath424 is also symmetric and therefore @xmath425 . also the off diagonal entries of @xmath424 , like @xmath426 , are positive because @xmath416 is a laplacian matrix . by corollary [ mu = m+ ] , it suffices to show that @xmath427 for any @xmath131 .
we first show that @xmath427 for @xmath428 . for @xmath260 , @xmath429 similarly for @xmath109 , @xmath430 now suppose @xmath431 . by lemma [ mu - prop ] , @xmath432 , where @xmath393 is an eigenvalue of @xmath424 . because @xmath423 , @xmath433 is an eigenvalue of @xmath424 ; therefore @xmath434 . to show that @xmath435 , by remark [ m+dini - linear ] , it suffices to show that @xmath436 where @xmath154 is the solution of @xmath437 . by the definition of dini derivative
, it suffices to show that @xmath438 is a non - increasing function of @xmath6 .
let @xmath439 , where @xmath440 with @xmath441 .
here we abuse the notation and assume that @xmath442 .
we ll show that @xmath443 .
first we ll prove the following inequality : [ ab ] for any real @xmath444 and @xmath445 and @xmath446 : @xmath447 for @xmath448 , the inequality is trivial .
suppose @xmath449 , and w.l.o.g @xmath450 and let @xmath451 .
then it suffices to prove that for @xmath452 , @xmath453 let @xmath454 .
we want to show that @xmath455 for @xmath452 . since @xmath456 and @xmath457 for @xmath452 , indeed @xmath455 . as we explained above
, @xmath424 is symmetric and @xmath423 . using this information and the above inequality : @xmath458 since @xmath459 .
note that @xmath460 is differentiable for @xmath251 .
[ mux = mup ] let @xmath461 and @xmath329 denote the logarithmic norms induced by @xmath462 and @xmath295 respectively .
then @xmath463 recall the following properties of kronecker product : * @xmath464 ; * if @xmath69 and @xmath465 are invertible , then @xmath466 .
hence : @xmath467\\ & = & \displaystyle\mu_p(-l\otimes qdq^{-1})\\ & = & \displaystyle\mu_p(-l\otimes d ) .
\end{array}\ ] ] the last equality holds because both @xmath115 and @xmath11 are diagonal , and so are commutative . therefore @xmath468 [ negm+x ] let @xmath469 denote the strong least upper bound logarithmic lipschitz constant induced by the norm @xmath295 on @xmath470
. then , @xmath471=0.\ ] ] by proposition [ negm+p ] , corollary [ mu = m+ ] and lemma [ mux = mup ] , @xmath471\;=\;\displaystyle\mu_{p , q}(-l\otimes d ) \;=\;\displaystyle\mu_p(-l\otimes d ) \;=\;\displaystyle m_{p}^{+}[-l\otimes d ] \;=\;0.\ ] ] [ mm2 ]
let @xmath472 denote the strong lub logarithmic lipschitz constant induced by the norm @xmath295 on @xmath470 and @xmath294 is the lub logarithmic lipschitz constant induced by the norm @xmath295 on @xmath473 .
then , @xmath474\leq { m_{p , q}}[f].\ ] ] the proof is exactly the same as the proof of proposition [ mm ] .
@xmath475 by subadditivity of @xmath472 , proposition [ subadd ] , proposition [ negm+x ] , and lemma [ mm2 ] : @xmath476\;\leq\;\displaystyle m^+_{p , q}[\tilde{f}]+m^+_{p , q}[-l\otimes d ] \;\leq\;m_{p , q}[f ] \;=\;c .
\end{array}\ ] ] now using corollary [ key1 ] , @xmath477 [ force ] assume @xmath8 is a linear operator . then @xmath478 the proof is immediate by subadditivity of logarithmic norm , proposition [ mm2 ] , and corollary [ mu = m+ ] .
note that ( [ force ] ) does nt need to hold if @xmath479 .
consider the following system : @xmath480 where @xmath481 , @xmath482 and @xmath483 . in this example
@xmath484 and @xmath485 .
we ll show that for @xmath486 , @xmath487 while @xmath488 .
by table @xmath90 , @xmath489 @xmath490 consider the reaction - diffusion ode @xmath402 and suppose assumption [ as - ode ] holds .
in addition assume that @xmath8 is continuously differentiable and @xmath327 for all @xmath328 .
then for any two solutions @xmath202 of @xmath402 we have @xmath477 the proof is immediate by theorem @xmath491 and proposition @xmath340 .
in this section we will show that for any @xmath235 , and @xmath492 , and @xmath493 nodes which are interconnected according to a connected , undirected graph , if @xmath494 , where @xmath139 is a convex subset of @xmath120 and @xmath393 is the smallest positive eigenvalue of @xmath426 , then every solution @xmath495 of ( [ discrete ] ) : @xmath496 has the following property : @xmath497 where @xmath498 and @xmath499 is a row vector defined by @xmath500 to show this , we first state the following lemma from @xcite : [ deimling ] for any @xmath501 , @xmath502 for @xmath492 nodes , there is just one possible graph : the complete graph with graph laplacian matrix @xmath503 which results in the following ode system : @xmath504 note that the smallest and only positive eigenvalue of @xmath426 is @xmath505 . for a fixed solution @xmath506 of this ode , let @xmath507 , @xmath498 , and @xmath508 .
then by the definition of @xmath472 , and lemma [ deimling ] we have , @xmath509\\ & = & pw^p(t)m^+_{p , q}[f-2\tilde{d}]\\ & \leq&pw^p(t)\displaystyle\sup_{(x , y)\in v}\mu_{p , q}(j_f(x , y)-2d ) . \end{array}\ ] ] the last inequality results by proposition [ key3 ] , @xmath510\geq m^+_{p , q}[f-2\tilde{d}].\ ] ] therefore @xmath511 , i.e. , in this case , @xmath512 where @xmath513 for @xmath514 , there are two possible graphs : first , the complete graph with graph laplacian matrix @xmath515 which leads to the following ode system : @xmath516 note that the smallest positive eigenvalue of @xmath426 is @xmath517 . for a fixed solution @xmath518 of this ode , define @xmath499 and @xmath519 as follows : @xmath520 and @xmath521 similar to case @xmath492 , we have @xmath522\\ & \leq&pw^p(t)\displaystyle\sup_{(x , y , z)\in v}\mu_{p , q}\left(j_f(x , y , z)-3d\right ) , \end{array}\ ] ] which leads to @xmath523 where @xmath524 taking the @xmath131th roots ,
@xmath525 the second graph has the graph laplacian matrix @xmath526 , with the following ode system : @xmath527 note that the smallest positive eigenvalue of @xmath426 is @xmath528 .
let again for a fixed solution @xmath518 , @xmath529 and @xmath498
. then @xmath530 by lemma [ ab ] , @xmath531 hence @xmath532\\ & \leq&pw^p(t)\displaystyle\sup_{(x , y , z)\in v}\mu_{p , q}(j_f(x , y , z)-d ) .
\end{array}.\ ] ] similar to the previous case : @xmath497 where in this case @xmath533
[ joo]@xcite let @xmath27 , @xmath39 be arbitrary sets and let @xmath534 be an arbitrary function . for any @xmath535 and @xmath326 , denote @xmath536 and @xmath537 the set of all real numbers @xmath538 such that for any @xmath535 , @xmath539 and let @xmath540
. then @xmath541 if and only if for every @xmath542 , @xmath543 . in this case @xmath544
. 1 . @xmath552 .
@xmath553 hence by the definition , @xmath552 .
2 . @xmath554 . for an arbitrary @xmath555
we ll show that @xmath556 and then we can conclude that @xmath554 . since @xmath555 , there exists @xmath557 such that @xmath558 ( otherwise @xmath559 and so @xmath560 ) .
this means that for all @xmath561 , @xmath562 which implies @xmath563 and hence @xmath564 .
3 . @xmath565 . for an arbitrary @xmath542
we ll show that @xmath566 and then we can conclude that @xmath567 .
since @xmath542 , there exists @xmath568 such that @xmath569 ( otherwise @xmath570 ) .
since we assumed that for every @xmath542 , @xmath543 , and since @xmath571 , @xmath568 , we have @xmath572 . by corollary [ n.e.intersection ] , because @xmath569 , then @xmath573 , i.e. there exists @xmath574 such that for all @xmath535 , @xmath575 , which implies @xmath576 and hence @xmath577 .
now we suppose @xmath578 and fix @xmath542 .
we ll show that @xmath543 .
since @xmath542 , there exists @xmath568 such that @xmath569 .
this means that there exists @xmath574 such that for all @xmath535 , @xmath579 , i.e. @xmath580 . since @xmath581 , there exists @xmath582 such that for all @xmath535 , @xmath583 , i.e. @xmath543 . for a fixed arbitrary norm @xmath584 on @xmath470 and a fixed arbitrary matrix @xmath585 , define @xmath586 by @xmath587 where @xmath588 . for any @xmath589 and @xmath326 , let @xmath590 and let @xmath537 be the set of all real numbers @xmath538 such that @xmath591 whenever @xmath589
let @xmath540 .
_ proof of claim @xmath263 .
_ to apply proposition [ joo ] , we ll show that for @xmath542 , @xmath600 , where @xmath590 and @xmath601 is defined as above . by lemma [ f - dec ]
, @xmath602 is decreasing in @xmath244 which implies @xmath603 when @xmath604 . also by the definition of @xmath601
, @xmath542 implies that @xmath591 for any @xmath605 . on the other hand ,
each @xmath606 is a closed subset of @xmath607 , so they are all compact .
hence their intersection is non - empty .
_ proof of claim @xmath608 .
_ by lemma [ f - dec ] , since @xmath611 is non - increasing as @xmath612 , @xmath613 . by corollary [ n - dec ] , since @xmath597 is non - increasing as @xmath612 , @xmath614 by claim @xmath263 , the right hand side of the equalities in claim @xmath608 are equal , and therefore so are their left hand sides : @xmath615 which implies @xmath83
. fix @xmath616 .
then there exists @xmath617 such that @xmath618 .
indeed @xmath619 . using holder s inequality , @xmath620
hence @xmath8 is an increasing function of @xmath131 .
now we ll show that as @xmath269 , @xmath621 .
note that for any @xmath131 , @xmath622 and therefore @xmath623 . to prove the converse inequality , for any @xmath178
, we define @xmath624 which by the definition of @xmath270 has positive measure , i.e. @xmath625 @xmath626 but because @xmath627 , @xmath628 as @xmath269 .
therefore for any arbitrary @xmath178 , @xmath629 which implies @xmath630 m. miller , m. hafner , e.d .
sontag , n. davidsohn , s. subramanian , p. e. m. purnick , d. lauffenburger , and r. weiss .
modular design of artificial tissue homeostasis : robust control through synthetic cellular heterogeneity .
, 8:e1002579 , 2012 . | this paper proves that contractive ordinary differential equation systems remain contractive when diffusion is added .
thus , diffusive instabilities , in the sense of the turing phenomenon , can not arise for such systems .
an important biochemical system is shown to satisfy the required conditions . |
scalar mesons are degrees of freedom of qcd ( quantum chromodynamics ) that possess quantum numbers @xmath0 , where @xmath1 is the total spin , @xmath2 represents parity and @xmath3 is the charge conjugation .
mesons that possess isospin @xmath4 are referred to as isoscalars .
scalars with @xmath4 represent qcd degrees of freedom with some peculiar features : they are defined as carrying quantum numbers of vacuum but are identifiable as resonances ; as mesons , they should be of @xmath5 ( quarkonium ) structure but are too many to simply represent @xmath5 objects ; and they are strongly mixed with each other so that disentangling them is in the majority of cases a highly non - trival task . however , they are also of extreme importance for qcd since their vacuum expectation values represent the cause of the spontaneous breaking of the chiral symmetry in qcd , and this `` higgs mechanism '' is in turn expected to generate goldstone bosons of qcd : pions , kaons and others .
thus understanding isoscalar mesons is a fundamental question of qcd , and one that has by no means been closed after decades of research and debate .
listings of the particle data group ( pdg ) cite the existence of five @xmath6 states in the energy region up to 1.8 gev : @xmath7 or @xmath8 , @xmath9 , @xmath10 , @xmath11 and @xmath12 @xcite . let us review basic experimental data regarding these resonances . * historically , debates regarding scalar mesons have been mostly focused on the @xmath7 resonance ( also known as @xmath8 ) .
the latter is characterised by a decay width that is of the same magnitude as the resonance mass according to pdg , it has a mass of ( 400 - 500 ) mev and a decay width of ( 400 - 700 ) mev .
the existence of this state was suggested in linear sigma models long ago : the state was introduced theoretically as the putative chiral partner of the pion @xcite . since the pion is a well - established quarkonium state , then the chiral - partner state of the pion had to be of the same structure .
additionally , the pion is a pseudo - goldstone bosone its mass is the lowest in the meson spectrum .
therefore , naive expectations were that ( _ i _ ) the chiral partner of the pion would be the lightest scalar state with @xmath4 , i.e. , the @xmath8 meson , and consequently ( _ ii _ ) that the @xmath8 meson was a @xmath5 state . however , currently it is by no means clear that @xmath8 is truly a @xmath5 state , implying that its interpretation as the chiral partner of the pion is in doubt .
we will come back to this issue later in this manuscript . *
first hints about the existence of the @xmath9 resonance were discovered in analyses of @xmath13 scattering data found to exhibit a rapid drop in the cross - section in the energy region between @xmath14 mev and @xmath15 mev ( i.e. , close to the @xmath16 threshold ) in saclay data on antiproton - proton @xcite and berkley data on pion - proton collisions @xcite .
the most precise modern - data analyses result in pole mass and pole decay widths values in the close vicinity of the mentioned energy region
see , e.g. , ref .
@xcite where the pole mass was determined as @xmath17 mev and the pole decay width as @xmath18 mev . *
the @xmath10 resonance is a broad enhancemenet in the @xmath19 , @xmath20 and , to a lesser extent , @xmath16 and @xmath21 channels in the energy region of approximately 1.3 gev .
it is characterised by emergence of two peaks , respectively in the @xmath22 and @xmath20 decay channels @xcite , the disentaglement of which requires a simultaneous analysis of @xmath23 and @xmath24 scattering data . the analysis of ref .
@xcite demonstrates that the @xmath22 decay channel is dominant in the energy region up to 1.35 gev whereas the energy region thereafter is dominated by the opening of the @xmath25 threshold leading to a strong @xmath20 decay pattern .
note that @xmath10 represents a single resonance despite possessing two peaks since a combined dispersive analysis of both @xmath22 and @xmath20 channels yields only one pole .
historically , early hints of the @xmath10 existence were observed in @xmath26 collisions at cern in 1969 with a possible @xmath27 enhancement claimed at @xmath28 gev @xcite .
current edition of the particle data group listings accumulates all available data on @xmath29 estimating an interval for , rather than stating exact values of , its mass and decay width : @xmath30-@xmath31 mev and @xmath32-@xmath33 mev @xcite . *
the discovery of the @xmath34 resonance originated in search for the scalar glueball state .
this resonance is found mostly in pion final states from nucleon - nucleon ( or antinucleon - nucleon ) and pion - nucleon scattering processes .
if such processes produce four pions , then @xmath34 is reconstructed from @xmath27 final states in the @xmath35 channel and from @xmath36 final states in the @xmath35 or @xmath37 channels .
the resonance is therefore at least partly reconstructed in channels containing a double pomeron exchange rendering the state a glueball candidate @xcite .
the pdg cites a world - average mass @xmath38 mev and decay width @xmath39 mev @xcite . *
the @xmath12 resonance is characterised by a predominant @xmath16 decay channel , marking a clear point of distinction between this resonance and the above mentioned ones .
the earliest evidence for the @xmath40 resonance was obtained from the decay @xmath41 at the slac crystal ball detector from @xmath42 annihilation @xcite ; a resonance with a mass of @xmath43 mev , a decay width of @xmath44 mev and the charge conjugation @xmath45 was found ( the latter was because the resonance had been produced in a radiative @xmath46 decay but , interestingly , no final determination of the total spin and parity was possible from these first data ) .
subsequent analyses , such as those in ref .
@xcite , determined the resonance to be of @xmath47 nature .
the pdg cites a world - average mass @xmath48 mev and a decay width @xmath49 mev @xcite . assuming a @xmath50 structure for mesons ( where @xmath51 denotes a constituent quark ) it is possible to construct two @xmath52 states for three quark flavours ( up @xmath53 , down @xmath54 and strange @xmath55 ) once the approximate isospin symmetry in the nonstrange - quark sector is considered ( @xmath56 ) : @xmath57 and @xmath58 .
the latter are per construction pure states .
they can obviously mix by virtue of carrying the same quantum numbers ; the mixed states are expected to be those appearing in the physical spectrum .
clearly , two pure states will produce the same number of mixed states thus only two of the five previously discussed physical states can conceivably be described in this approach .
hence the natural questions are * where in the physical spectrum are isoscalar @xmath50 states located ?
* what structure do the remaining states possess ?
answers to these questions are complicated by the already mentioned fact that states with the same @xmath59 quantum numbers in general , and isoscalar mesons in particular , are expected to mix .
thus an isoscalar ( or indeed any other ) meson need not necessarily be of @xmath5 structure it can also possess admixtures ( or even be predominantly composed ) of tetraquark ( @xmath60 ) , glueball ( bound states of qcd gauge bosons gluons ) or molecular ( e.g. , @xmath19 or @xmath16 ) contributions .
for the cases of the above - stated isoscalar mesons we then note the following : * as indicated above , the @xmath7 resonance is usually considered as the natural option for the chiral partner of the pion , implying that @xmath7 possesses the @xmath50 structure as suggested , e.g. , by ref .
however , a @xmath61 scalar state possesses the intrinsic angular momentum @xmath62 as well as the relative spin of the quarks @xmath63 . for this reason one could also easily expect the scalar @xmath5 state to be in the region above @xmath64 gev
this would imply that ( _ i _ ) isoscalars above 1 gev possess the @xmath5 structure and ( _ ii _ ) @xmath7 represents a tetraquark state or a @xmath19 bound state .
for details on both statements see , e.g. , refs .
@xcite . * as already discussed , the @xmath9 resonance is close to the kaon - kaon threshold rendering a structure analysis for this state rather difficult . the resonance may be interpreted as a quarkonium @xcite , as a @xmath65 state @xcite , as a @xmath16 bound state @xcite , as a glueball @xcite or even as an @xmath66 bound state @xcite . *
the @xmath10 resonance appears to be a suitable candidate for a non - strange @xmath50 state @xcite or a dynamically generated state @xcite . *
the @xmath11 resonance seems to represent a ( predominant ) glueball state @xcite although claims have also been made that , contrarily , @xmath12 is of predominantly glueball nature @xcite . alternatively , @xmath12 can also be interpreted as a predominantly @xmath67 state @xcite .
hence the current state of knowledge allows us to only claim with certitude that all the above mentioned states mix @xcite ; some of them even overlap due to large decay widths @xcite .
the situation is thus rather complicated and it is enhanced even further by the possible existence of a sixth isoscalar state , located very proximal to the set of the already mentioned isoscalars : the @xmath68 resonance .
the existence of the putative new @xmath68 resonance has been claimed by the bes ii collaboration in 2004 @xcite .
the mass and decay width of the resonance were determined by the collaboration as @xmath69 and @xmath70 mev .
the stated large decay width implies a strong overlap of this resonance with @xmath12 ; however , there is a clear point of distinction between the two resonances : @xmath12 is reconstructed predominantly in the kaon decay channels whereas @xmath68 is reconstructed predominantly in the pion decay channels .
+ there are four basic production mechanisms for @xmath12 and @xmath68 via @xmath46 decays @xcite : * ( _ i _ ) @xmath71 , * ( _ ii _ ) @xmath72 , * ( _ iii _ ) @xmath73 , * ( _ iv _ ) @xmath74 .
reactions ( _ i _ ) and ( _ iii _ ) allow for reconstruction of @xmath40 @xcite whereas reactions ( _ ii _ ) and ( _ iv _ ) allow for reconstruction of @xmath75 . importantly , assuming @xmath40 and @xmath75 to be the same resonance leads to a contradiction : such a resonance would have to possess a pion - to - kaon - decay ratio of @xmath76 according to reactions ( _ i _ ) and ( _ ii _ ) and , simultaneously , the pion - to - kaon - decay ratio @xmath77 according to reactions ( _ iii _ ) and ( _ iv _ ) @xcite .
there can be no one resonance with these simultaneous features
thus bes ii data suggest @xmath40 and @xmath75 to represent two distinct resonances .
+ let us finally discuss the relevance of the deliberations in this paper for fair .
the proposed panda experiments at fair will study interactions between antiprotons and fixed - target protons and nuclei in the momentum range of 1.5 - 15 gev .
a stated goal of the panda collaboration is the exploration of light - hadron spectroscopy , including search for glueballs and multiquark states @xcite . in this regard
i would like to suggest the following courses of action in the particular case of possible mixing scenarios in the scalar sector : * any viable search for members of a glueball spectrum must include the ground state as otherwise the spectrum would be incomplete .
the panda collaboration is in a unique position to greatly advance our search for glueballs since ( _ i _ ) the collaboration can build upon extensive experimental work performed in the last decades and ( _ ii _ ) the technical specifications of the panda detector allow for a search for glueballs with unprecedented accuracy in comparison with similar undertakings in the past .
+ however , the glueball state will inevitably overlap with other states that possess the same quantum numbers ( and may be of @xmath5 , @xmath60 or molecular structure ) and the disentanglement of various signals from each of the mentioned resonances will represent a formidable task .
+ * current theoretical results @xcite suggest the scalar glueball to be positioned at ( 1.5 - 1.7 ) gev , i.e. , in the vicinity of the established resonances @xmath10 , @xmath11 and @xmath12 as well as the newly proposed @xmath68 state .
consequently , a viable search for the scalar glueball will have to consider overlap of the glueball not only with @xmath10 , @xmath11 and @xmath12 _ but also _ with @xmath68 . in other words : a viable search for the glueball ground state at panda must consider the possibility of the existence of @xmath68 .
if @xmath68 is not found at panda , then bes results may be put in doubt ; but if @xmath68 is ascertained then its signal will have to be disentangled from the glueball signal in order for the glueball search to be successful , making the search even more challenging and thus even more interesting .
scalar mesons represent an interesting challenge for the future panda experiments at fair : the planned search for the glueball state in the @xmath6 channel will have to consider mixing / overlap between non - strange @xmath5 , strange @xmath5 , non - strange and strange @xmath60 as well as possible molecular - type states . additionally , obtaining a clear signal for
the glueball will inevitably require the confirmation ( or negation ) of the existence of a putative new isoscalar state referred to as @xmath68 by the bes ii collaboration @xcite .
if @xmath68 exists , it will most certainly overlap with the glueball .
thus a viable search for a glueball state simultaneously implies a search for @xmath68 : ignoring @xmath68 will strongly distort any signal for the glueball .
for this reason , the search for a glueball at panda is more than a search for merely one state it is a search where several very close - by states have to be considered simultaneously .
9 m. alston - garnjost , a. barbaro - galtieri , s. m. flatte , j. h. friedman , g. r. lynch , s. d. protopopescu , m. s. rabin , f. t. solmitz , phys .
b * 36 * , 152 ( 1971 ) ; s. m. flatte , m. alston - garnjost , a. barbaro - galtieri , j. h. friedman , g. r. lynch , s. d. protopopescu , m. s. rabin , f. t. solmitz , phys .
lett . * b38 * , 232 ( 1972 ) ; s. d. protopopescu , m. alston - garnjost , a. barbaro - galtieri , s. m. flatte , j. h. friedman , t. a. lasinski , g. r. lynch , m. s. rabin _ et al .
_ , phys .
d * 7 * , 1279 ( 1973 ) .
d. alde _ et al . _
[ serpukhov - brussels - los alamos - annecy(lapp ) collaboration ] , phys .
b * 182 * , 105 ( 1986 ) ; d. alde _ et al .
_ [ ifve - brussels - annecy - los alamos collaboration ] , phys .
lett . * b284 * , 457 ( 1992 ) ; t. a. armstrong _ et al .
_ [ e760 collaboration ] , phys .
b * 307 * , 394 ( 1993 ) ; d. v. bugg , i. scott , b. s. zou , v. v. anisovich , a. v. sarantsev , t. h. burnett and s. sutlief , phys .
b * 353 * , 378 ( 1995 ) ; d. barberis _ et al .
_ [ wa102 collaboration ] , phys .
b * 462 * , 462 ( 1999 ) [ arxiv : hep - ex/9907055 ] ; s. chekanov _ et al .
_ [ zeus collaboration ] , phys .
* 101 * , 112003 ( 2008 ) [ arxiv:0806.0807 [ hep - ex ] ] .
y. nambu and g. jona - lasinio , phys .
rev . * 122 * , 345 ( 1961 ) ; r. delbourgo and m. d. scadron , phys .
* 48 * , 379 ( 1982 ) ; r. delbourgo and m. d. scadron , mod .
lett . a * 10 * , 251 ( 1995 ) [ hep - ph/9910242 ] ; r. delbourgo and m. d. scadron , int .
. phys .
a * 13 * , 657 ( 1998 ) [ hep - ph/9807504 ] .
p. ko and s. rudaz , phys .
d * 50 * , 6877 ( 1994 ) ; j. r. pelaez , mod .
a * 19 * , 2879 ( 2004 ) [ hep - ph/0411107 ] ; d. parganlija , f. giacosa and d. h. rischke , aip conf .
proc . * 1030 * , 160 ( 2008 ) [ arxiv:0804.3949 [ hep - ph ] ] ; d. parganlija , f. giacosa and d. h. rischke , pos c**onfinement8 * * , 070 ( 2008 ) [ arxiv:0812.2183 [ hep - ph ] ] ; d. parganlija , f. giacosa and d. h. rischke , arxiv:0911.3996 [ nucl - th ] ; j. ruiz de elvira , j. r. pelaez , m. r. pennington and d. j. wilson , phys . rev .
d * 84 * , 096006 ( 2011 ) [ arxiv:1009.6204 [ hep - ph ] ] ; d. parganlija , f. giacosa , p. kovacs and g. wolf , aip conf .
proc . * 1343 * , 328 ( 2011 ) [ arxiv:1011.6104 [ hep - ph ] ] ; p. kovacs , g. wolf , f. giacosa and d. parganlija , arxiv:1102.4732 [ hep - ph ] ; d. parganlija , phys .
scripta t * 150 * , 014029 ( 2012 ) [ arxiv:1109.4331 [ hep - ph ] ] ; w. heupel , g. eichmann and c. s. fischer , arxiv:1206.5129 [ hep - ph ] ; d. parganlija , p. kovacs , g. wolf , f. giacosa and d. h. rischke , arxiv:1208.2054 [ hep - ph ] .
d. parganlija , f. giacosa , d. h. rischke , phys .
d * 82 * , 054024 ( 2010 ) [ arxiv:1003.4934 [ hep - ph ] ] ; d. parganlija , f. giacosa and d. h. rischke , acta phys .
supp . * 3 * , 963 ( 2010 ) [ arxiv:1004.4817 [ hep - ph ] ] ; d. parganlija , f. giacosa , d. h. rischke , p. kovacs and g. wolf , int .
j. mod .
a * 26 * , 607 ( 2011 ) [ arxiv:1009.2250 [ hep - ph ] ] ; d. parganlija , acta phys . polon . supp . * 4 * , 727 ( 2011 ) [ arxiv:1105.3647 [ hep - ph ] ] .
d. morgan , phys .
b * 51 * , 71 ( 1974 ) ; a. bramon and e. masso , phys .
b * 93 * , 65 ( 1980 ) [ erratum - ibid .
b * 107 * , 455 ( 1981 ) ] ; m. d. scadron , phys .
d * 26 * , 239 ( 1982 ) ; e. van beveren , t. a. rijken , k. metzger , c. dullemond , g. rupp and j. e. ribeiro , z. phys .
c * 30 * , 615 ( 1986 ) [ arxiv:0710.4067 [ hep - ph ] ] .
t. barnes , phys .
b * 165 * , 434 ( 1985 ) ; j. d. weinstein and n. isgur , phys .
d * 41 * , 2236 ( 1990 ) ; m. p. locher , v. e. markushin and h. q. zheng , eur .
j. c * 4 * , 317 ( 1998 ) [ arxiv : hep - ph/9705230 ] .
s. surovtsev , p. bydzovsky , r. kaminski and m. nagy , phys .
d * 81 * , 016001 ( 2010 ) [ arxiv:0811.0906 [ nucl - th ] ] .
r. molina , d. nicmorus and e. oset , phys .
d * 78 * , 114018 ( 2008 ) [ arxiv:0809.2233 [ hep - ph ] ] ; l. s. geng and e. oset , phys .
d * 79 * , 074009 ( 2009 ) [ arxiv:0812.1199 [ hep - ph ] ] . c. morningstar and m. j. peardon , aip conf .
* 688 * , 220 ( 2004 ) [ arxiv : nucl - th/0309068 ] ; m. loan , x. q. luo and z. h. luo , int .
j. mod .
a * 21 * , 2905 ( 2006 ) [ arxiv : hep - lat/0503038 ] ; y. chen _ et al .
_ , phys .
d * 73 * , 014516 ( 2006 ) [ arxiv : hep - lat/0510074 ] ; s. janowski , d. parganlija , f. giacosa and d. h. rischke , phys .
d * 84 * , 054007 ( 2011 ) [ arxiv:1103.3238 [ hep - ph ] ] .
d. black , a. h. fariborz , f. sannino and j. schechter , phys .
d * 59 * , 074026 ( 1999 ) [ hep - ph/9808415 ] ; d. black , a. h. fariborz and j. schechter , phys .
d * 61 * , 074001 ( 2000 ) [ hep - ph/9907516 ] ; a. h. fariborz , r. jora and j. schechter , phys .
d * 72 * , 034001 ( 2005 ) [ hep - ph/0506170 ] ; f. giacosa , phys .
d * 75 * , 054007 ( 2007 ) [ arxiv : hep - ph/0611388 ] . | i review issues related to the scalar - meson phenomenology considered of interest for the future panda experiments at fair . |
the notion of sterile neutrinos was introduced by b. pontecorvo in 1967 @xcite . in the last years the possibility of transitions of flavor neutrinos into noninteracting sterile states has been widely discussed in the literature .
this interest in sterile neutrinos is connected mainly with the result of the lsnd experiment @xcite that together with the results of atmospheric @xcite and solar @xcite neutrino experiments imply the existence of sterile neutrinos @xcite .
let us start with the discussion of the possibilities of sterile neutrinos to appear in neutrino mixing schemes ( see , for example , refs .
flavor neutrinos @xmath0 , @xmath1 , @xmath2 , are determined by the standard charged - current ( cc ) and neutral - current ( nc ) weak interactions @xmath3 that conserve the electron @xmath4 , muon @xmath5 and tau @xmath6 lepton numbers , @xmath7 if there is neutrino mixing , the conservation laws ( [ lepton numbers ] ) are violated .
neutrino mass term _ that does not conserve lepton numbers has the general form @xmath8 where @xmath9 and @xmath10 are @xmath11-component columns and @xmath12 is a @xmath13 matrix .
there are two different possibilities for @xmath9 : 1 .
only flavor fields @xmath14 ( @xmath15 ) enter into @xmath9 : @xmath16 in this case @xmath12 is a @xmath17 matrix and for the mixing we have @xmath18 where @xmath19 is a unitary matrix and @xmath20 is the field of neutrinos with mass @xmath21 . the nature of neutrinos @xmath20 with definite mass depends on @xmath22 : 1 . if @xmath23 where @xmath24 ( @xmath15 ) are right - handed neutrino fields , the total lepton number @xmath25 is conserved and the neutrinos @xmath20 with definite mass are _
dirac particles_. the corresponding mass term is called `` dirac mass term '' .
if @xmath26 where @xmath27 is the charge - conjugated ( right - handed ) component of the left - handed field @xmath28 ( @xmath15 and @xmath29 is the matrix of charge conjugation ) , there are no conserved lepton numbers and neutrinos with definite masses are _ majorana particles _ ( @xmath30 ) .
the corresponding mass term is called `` majorana mass term '' .
+ from eq .
( [ 3mixing ] ) it follows that in both cases only transitions between active neutrinos @xmath31 are possible .
notice that with the investigation of neutrino oscillations it is not possible to distinguish the case of dirac neutrinos from the case of majorana neutrinos @xcite .
2 . in the column
@xmath9 not only the flavor fields @xmath28 ( @xmath15 ) but also other fields @xmath32 ( @xmath33 ) enter @xmath34 the fields @xmath32 do not enter in the standard cc and nc weak interactions ( [ cc ] ) , ( [ nc ] ) and are called `` sterile '' .
it is possible that there are three sterile fields @xmath32 that are the charge - conjugated components of right - handed neutrino fields , @xmath35 ( @xmath36 ) .
however , the sterile fields could also be fields of some other particles ( susy , ) .
+ in this case , for the neutrino mixing we have @xmath37 where @xmath38 is the number of sterile fields and @xmath20 ( @xmath39 ) is the field of neutrinos with mass @xmath21 .
let us stress that the number of sterile fields depends on the concrete scheme of neutrino mixing .
+ the nature of neutrinos with definite mass depends on @xmath10 . if @xmath40 , the @xmath20 s are majorana particles . if @xmath35 ( @xmath36 ) , the corresponding mass term is called `` dirac - majorana mass term '' .
+ if all the masses @xmath21 of the fields @xmath20 in eq.([mixing ] ) are small , not only transitions between flavor neutrinos @xmath31 but also transitions between flavor and sterile neutrinos @xmath41 are possible .
there are two possible ways to reveal the existence of sterile neutrinos : 1 . through the determination of the number of massive neutrinos .
if this number is larger than the number of flavor neutrinos ( three ) sterile neutrinos must exist .
2 . through the measurement of the total transition probability of neutrino with definite flavor ( @xmath0 or @xmath1 ) into all possible flavor neutrinos , @xmath42 ( @xmath43 ) .
this probability can be determined from the investigation of nc induced neutrino processes . if @xmath44 from the unitarity of the mixing matrix it follows that active flavor neutrinos transform into sterile states .
the data of neutrino oscillation experiments indicate the existence of three different scales of neutrino mass - squared difference @xmath45 : 1 .
@xmath46 from atmospheric neutrino experiments @xcite ; 2 .
@xmath47 ( or @xmath48 ) from solar neutrino experiments @xcite ; 3 .
@xmath49 from the lsnd experiment @xcite .
three different scales of @xmath45 can be accommodated only if at least four massive neutrinos exist in nature @xcite . in other words ,
the existing data indicate that sterile neutrinos exist . in the framework of the minimal scheme with four massive neutrinos , from the existing data it follows @xcite that the dominant transitions of solar neutrinos are @xmath50 .
below we will present the corresponding arguments .
figure [ 4spectra ] shows the six types of spectra of four massive neutrinos that can accommodate the solar , atmospheric and lsnd ranges of neutrino mass - squared differences .
these spectra are divided in two classes : class 1 constituted by the spectra ( i)(iv ) and class 2 comprising the spectra ( a ) and ( b ) . in the spectra of class 1 a group of three close masses
is separated from the fourth mass by the lsnd gap of @xmath51 . in the spectra of class
2 two pairs of close masses are separated by the lsnd gap .
0.99@cccccc ( 1,4 ) ( 0.1,0.2)(0,1)3.8 ( 0.0,0.2)(1,0)0.2 ( 0.4,0.15)(0,0)[l]@xmath52 ( 0.0,0.4)(1,0)0.2 ( 0.4,0.45)(0,0)[l]@xmath53 ( 0.0,0.8)(1,0)0.2 ( 0.4,0.8)(0,0)[l]@xmath54 ( 0.0,3.5)(1,0)0.2 ( 0.4,3.5)(0,0)[l]@xmath55 & ( 1,4 ) ( 0.1,0.2)(0,1)3.8 ( 0.0,0.2)(1,0)0.2 ( 0.4,0.2)(0,0)[l]@xmath52 ( 0.0,0.6)(1,0)0.2 ( 0.4,0.55)(0,0)[l]@xmath53 ( 0.0,0.8)(1,0)0.2 ( 0.4,0.85)(0,0)[l]@xmath54 ( 0.0,3.5)(1,0)0.2 ( 0.4,3.5)(0,0)[l]@xmath55 & ( 1,4 ) ( 0.1,0.2)(0,1)3.8 ( 0.0,0.2)(1,0)0.2 ( 0.4,0.2)(0,0)[l]@xmath52 ( 0.0,2.9)(1,0)0.2 ( 0.4,2.9)(0,0)[l]@xmath53 ( 0.0,3.3)(1,0)0.2 ( 0.4,3.25)(0,0)[l]@xmath54 ( 0.0,3.5)(1,0)0.2 ( 0.4,3.55)(0,0)[l]@xmath55 & ( 1,4 ) ( 0.1,0.2)(0,1)3.8 ( 0.0,0.2)(1,0)0.2 ( 0.4,0.2)(0,0)[l]@xmath52 ( 0.0,2.9)(1,0)0.2 ( 0.4,2.85)(0,0)[l]@xmath53 ( 0.0,3.1)(1,0)0.2 ( 0.4,3.15)(0,0)[l]@xmath54 ( 0.0,3.5)(1,0)0.2 ( 0.4,3.5)(0,0)[l]@xmath55 & ( 1,4 ) ( 0.1,0.2)(0,1)3.8 ( 0.0,0.2)(1,0)0.2 ( 0.4,0.2)(0,0)[l]@xmath52 ( 0.0,0.6)(1,0)0.2 ( 0.4,0.6)(0,0)[l]@xmath53 ( 0.0,3.3)(1,0)0.2 ( 0.4,3.25)(0,0)[l]@xmath54 ( 0.0,3.5)(1,0)0.2 ( 0.4,3.55)(0,0)[l]@xmath55 & ( 1,4 ) ( 0.1,0.2)(0,1)3.8 ( 0.0,0.2)(1,0)0.2 ( 0.4,0.15)(0,0)[l]@xmath52 ( 0.0,0.4)(1,0)0.2 ( 0.4,0.45)(0,0)[l]@xmath53 ( 0.0,3.1)(1,0)0.2 ( 0.4,3.1)(0,0)[l]@xmath54 ( 0.0,3.5)(1,0)0.2 ( 0.4,3.5)(0,0)[l]@xmath55 + ( i ) & ( ii ) & ( iii ) & ( iv ) & ( a ) & ( b ) in the case of the spectra of class 1 the amplitude of @xmath56 transitions in short - baseline ( sbl ) experiments is strongly suppressed and the upper bound of the corresponding oscillation amplitude obtained from the exclusion curves of sbl @xmath57 and @xmath58 disappearance experiments , is smaller than the value of the amplitude of @xmath59 oscillations found in the lsnd experiment @xcite .
thus , the schemes of mixing of four massive neutrinos belonging to class 1 are not compatible with the existing neutrino oscillation data . only the schemes of mixing of four massive neutrinos with mass spectra of class 2
, @xmath60 can describe all the existing neutrino oscillation data . in this case ,
the largest mass squared difference @xmath61 is relevant for the oscillations in sbl reactor and accelerator experiments .
the corresponding transition probabilities are @xcite @xmath62 here @xmath63 is the distance between neutrino source and neutrino detector , @xmath64 is neutrino momentum , and @xmath65 , @xmath66 are the oscillation amplitudes given by @xmath67 where the index @xmath68 runs over the values @xmath69 or @xmath70 .
the quantities @xmath71 for @xmath72 are constrained by the results of sbl reactor @xmath73 and accelerator @xmath74 disappearance experiments in which no indications in favor of neutrino oscillations were found : @xmath75 here @xmath76 where @xmath77 is the upper bound of the amplitude of @xmath78 transitions obtained from the exclusion curves of sbl reactor and accelerator disappearance experiments . using the exclusion plots obtained in the reactor bugey @xcite experiment and in the accelerator cdhs @xcite and ccfr @xcite experiments we have @xcite
@xmath79 further constraints on the quantities @xmath71 for @xmath72 can be found by taking into account the results of solar and atmospheric neutrino experiments .
the survival probabilities of solar @xmath0 s and atmospheric @xmath1 s are constrained by ( see @xcite ) @xmath80 where the index @xmath68 runs over 1,2 in scheme a and over 3,4 in scheme b. let us introduce the quantities @xmath81 with the index @xmath68 running over 1,2 in scheme a and over 3,4 in scheme b. from eqs .
( [ a0_e ] ) , ( [ a0_mu ] ) and ( [ pmin ] ) we conclude @xcite that the four - neutrino schemes a and b are compatible with the results of solar and atmospheric neutrino experiments only for @xmath82 constraints on the elements of the mixing matrix @xmath19 can be also obtained from the limit on the effective number of neutrinos @xmath83 in big - bang nucleosynthesis ( bbn ) ( see , for example , @xcite ) .
the analysis of recent data yields the upper bound @xmath84 at 95% cl @xcite , which implies that @xcite @xmath85 taking now into account the unitarity relation @xmath86 from eqs .
( [ r2 ] ) and ( [ c_s ] ) we come to the conclusion that @xmath87 and @xmath88 are small and @xmath89 and @xmath90 are large . in the approximation @xmath91 in scheme
a we have @xmath92 where @xmath93 and @xmath94 are mixing angles .
the corresponding mixing relations in scheme b can be obtained from eq .
( [ approx - mix ] ) with the change @xmath95 . from eq .
( [ approx - mix ] ) one can see that the dominant transitions of solar neutrinos are @xmath96 and the dominant transitions of atmospheric neutrinos and neutrinos in long - baseline ( lbl ) experiments are @xmath97 .
here we will consider possible model independent methods of searching for @xmath96 transitions in future solar neutrino experiments @xcite . in the sno experiment @xcite solar neutrinos
will be detected through the observation of the cc reaction @xmath99 of the nc reaction @xmath100 and of the elastic - scattering ( es ) reaction @xmath101 because of the large energy thresholds ( @xmath102 for the cc and es processes and 2.2 mev for the nc process ) , mainly @xmath103b neutrinos will be detected in the sno experiment .
let us write the initial flux of @xmath103b neutrinos as a function of energy @xmath104 in the form @xmath105 here @xmath106 is the total flux and @xmath107 is a known function @xmath108 that characterizes the spectrum of @xmath0 s in the decay @xmath109 .
the nc event rate is given by @xmath110 here @xmath111 is the average value of the total transition probability of solar @xmath0 s into all possible flavor neutrinos and @xmath112 is the average value of the cross section of the nc process ( [ sno - nc ] ) .
it is obvious that @xmath113 if @xmath96 transitions take place .
we can not , however , determine the quantity @xmath111 from ( [ nc event rate ] ) in a model independent way , because knowledge of the total flux @xmath106 is needed .
information on the total flux @xmath106 can be obtained from the data of the super - kamiokande experiment .
indeed , we have @xmath114 where @xmath115 is the average cross section of @xmath1-@xmath116 scattering and @xmath117 is the nc contribution to the @xmath118 event rate @xmath119 . in eq .
( [ sigma2 ] ) @xmath120 ( @xmath121 ) is the cross section of @xmath0-@xmath116 ( @xmath58-@xmath116 ) scattering and @xmath122 is the flux of solar @xmath0 s with energy @xmath104 on the earth , that will be measured in the sno experiment through the observation of the cc reaction ( [ sno - cc ] ) .
combining the relations ( [ nc event rate ] ) and ( [ sigma1 ] ) we obtain the ratio @xmath123 that is independent from the the total flux @xmath106 . only measurable ( and known ) quantities enter in the ratio @xmath124 . from eq .
( [ r ] ) it is obvious that @xmath125
only if @xmath126 , _ i.e. _ if solar neutrinos transfer into sterile states . let us notice , however , that the ratio @xmath124 can be different from one only if the probability @xmath127 depends on the neutrino energy .
another possibility to reveal transitions of solar neutrinos into sterile states could be realized by combining the super - kamiokande recoil electron spectrum and the spectrum of @xmath0 s on the earth that will be determined in the sno experiment .
we have @xmath128 where @xmath129 is the recoil electron kinetic energy in the super - kamiokande experiment , @xmath130 and @xmath131 . from relation ( [ sigma3 ] )
it is obvious that if the quantity @xmath132 depends on @xmath129 it means that the probability @xmath133 is different from zero and depends on neutrino energy .
the dependence on @xmath129 of the ratio @xmath134 is presented in fig .
[ reses ] .
the dashed line corresponds to the small mixing angle msw solution of the solar neutrino problem in the case of @xmath96 transitions with @xmath135 , @xmath136 .
the most direct way to search for @xmath96 transitions of atmospheric neutrinos is to investigate the nc process @xmath137 this possibility has been discussed in ref .
if we accept all the existing indications in favor of neutrino oscillations we come to the conclusion that sterile neutrinos must exist . thus , from the phenomenological point of view the problem of sterile neutrinos is connected with the correctness of the existing indications in favor of neutrino oscillations and first of all with the correctness of the indications obtained in the lsnd experiment , the only accelerator sbl experiment in which neutrino oscillations have been observed . through the investigation of nc processes it is possible to search for transitions of active neutrinos into sterile states in solar , atmospheric sbl and lbl experiments .
several possible tests are based only on the unitarity of the mixing matrix and can be done in a model independent way .
y. fukuda _
et al . _ ,
b * 335 * , 237 ( 1994 ) ; r. becker - szendy _ et al .
b ( proc . suppl . )
* 38 * , 331 ( 1995 ) ; w.w.m .
et al . _ , hep - ex/9901024 ; m. ambrosio _
et al . _ ,
phys . lett .
b * 434 * , 451 ( 1998 ) .
et al . _ ,
j. * 496 * , 505 ( 1998 ) ; k.s .
et al . _ ,
lett . * 77 * , 1683 ( 1996 ) ; w. hampel _ et al .
b * 388 * , 384 ( 1996 ) ; d.n .
et al . _ ,
lett . * 77 * , 4708 ( 1996 ) .
peltoniemi , d. tommasini and j.w.f .
valle , phys .
b * 298 * , 383 ( 1993 ) ; e.j .
_ , _ ibid _ * 357 * , 608 ( 1995 ) ; s. mohanty , d.p .
roy and u. sarkar , _ ibid _ * 445 * , 185 ( 1998 ) ; j.t .
peltoniemi and j.w.f .
valle , nucl .
b * 406 * , 409 ( 1993 ) ; q.y .
liu and a.yu .
smirnov , _ ibid _ * 524 * , 505 ( 1998 ) ; d.o .
caldwell and r.n .
mohapatra , phys .
d * 48 * , 3259 ( 1993 ) ; e. ma and p. roy , _ ibid _ * 52 * , r4780 ( 1995 ) ; s. goswami , _ ibid _ * 55 * , 2931 ( 1997 ) ; a.yu . smirnov and m. tanimoto ,
_ ibid _ * 55 * , 1665 ( 1997 ) ; n. gaur _ et al .
_ , _ ibid _ * 58 * , 071301 ( 1998 ) ; e.j .
chun , c.w .
kim and u.w .
lee , _ ibid _ * 58 * , 093003 ( 1998 ) ; v. barger _
_ , _ ibid _ * 58 * , 093016 ( 1998 ) ; j.j .
gomez - cadenas and m.c .
gonzalez - garcia , z. phys .
c * 71 * , 443 ( 1996 ) ; k. benakli and a.yu .
smirnov , phys .
79 * , 4314 ( 1997 ) ; v. barger , t.j .
weiler and k. whisnant , _ ibid _ * 427 * , 97 ( 1998 ) ; s.c .
_ , _ ibid _ * 430 * , 296 ( 1998 ) . s.m .
bilenky , c. giunti and w. grimus , proc . of _
neutrino 96 _ , helsinki , june 1996 , edited by k. enqvist _
et al . _ , p. 174
, world scientific , 1997 ( hep - ph/9609343 ) ; eur .
j. c * 1 * , 247 ( 1998 ) ; s.m .
bilenky , c. giunti , w. grimus and t. schwetz , hep - ph/9903454 . c. giunti , c.w .
kim and u.w .
lee , phys . rev .
d * 46 * , 3034 ( 1992 ) ; r. foot , mod .
lett . a * 9 * , 169 ( 1994 ) ; r. foot and r.r .
volkas , phys .
d * 52 * , 6595 ( 1995 ) ; z. berezhiani and r.n .
mohapatra , _ ibid _ * 52 * , 6607 ( 1995 ) ; j. bunn , r. foot and r.r .
volkas , phys .
b * 413 * , 109 ( 1997 ) ; b. brahmachari and r.n .
mohapatra , phys .
b * 437 * , 100 ( 1998 ) ; j.p .
bowes and r.r .
volkas , j. phys .
g * 24 * , 1249 ( 1998 ) ; m. collie and r. foot , phys .
b * 432 * , 134 ( 1998 ) ; r.r .
volkas and y.y.y .
wong , phys .
d * 58 * , 113001 ( 1998 ) ; d. suematsu , prog .
99 * , 483 ( 1998 ) ; hep - ph/9805305 ; w. krolikowski , acta phys .
b * 29 * , 629 ( 1998 ) ; _ ibid . _ * 29 * , 755 ( 1998 ) ; hep - ph/9803323 ; hep - ph/9808207 ; hep - ph/9808307 ; y. koide and h. fusaoka , phys . rev .
d * 59 * , 053004 ( 1999 ) ; a. geiser , phys .
b * 444 * , 358 ( 1999 ) . s.m .
bilenky , j. hoek and s.t .
petcov , phys .
b * 94 * , 495 ( 1980 ) ; i.yu . kobzarev _
et al . _ , yad . fiz . *
32 * , 1590 ( 1980 ) [ sov .
* 32 * , 823 ( 1980 ) ] ; m. doi _
b * 102 * , 323 ( 1981 ) ; p. langacker _
et al . _ ,
b * 282 * , 589 ( 1987 ) .
y. suzuki , proc . of _ neutrino 96 _ , helsinki , june 1996 , edited by k. enqvist
al . _ , p. 237
, world scientific , singapore , 1997 ; f. vissani and a.yu .
smirnov , phys .
b * 432 * , 376 ( 1998 ) ; k.r.s .
balaji , g. rajasekaran and s. uma sankar , hep - ph/9806256 ; f. vissani , talk presented at the ringberg euroconference _
new trends in neutrino physics _ , tegernsee , germany , may 1998 , preprint desy-98 - 135 , 1998 . | the notion of sterile neutrinos is discussed .
the schemes of mixing of four massive neutrinos , which imply the existence of sterile neutrinos , are briefly considered .
several model independent methods that allow to reveal possible transitions of solar neutrinos into sterile states are presented . |
the possibility of pion condensation in nuclear medium was suggested by a. b. migdal for the first time in the 1970s @xcite .
thereafter , many studies @xcite of `` in - medium '' pion properties have been performed due to the important consequences in subnuclear physics @xcite , in the physics of neutron stars @xcite , supernovas @xcite , and the heavy ion collisions @xcite .
these analysis considered the pion as an elementary object but we know that they can be considered as nambu - goldstone bosons generated by chiral symmetry breaking . hence the internal structure and the mass of the pion can be sensitive to the quantum chromodynamics ( qcd ) vacuum modifications in the finite density environment @xcite ; moreover the finite baryon density , and even the isospin density arising from the neutrality condition , can modify the structure of the qcd ground state , and it can in turn produce significant modifications of the pion properties in the medium . in this work @xcite , we present the possibility of a charged pion ( @xmath5 ) condensation starting from a _ microscopic _ model which is built with quarks as the constituents of pions , and which exhibits chiral symmetry restoration at the finite quark chemical potential @xmath6 or temperature @xmath7 .
we derive an appropriate criterion for @xmath5 condensation , that we use to show the fragility of the @xmath5 condensation with respect to explicit chiral symmetry breaking via a finite current quark mass .
the conditions for the onset of @xmath5 condensation at finite density are investigated using the nambu jona - lasinio ( njl ) model of qcd .
we find that the threshold for @xmath5 condensation for noninteracting elementary pion gas is @xmath8 ( @xmath9 ) for positive ( negative ) @xmath10 , where @xmath10 is the electric chemical potential and @xmath4 the in - medium pion mass .
finally , we clarify the effect of the current quark mass in the electrically neutral ground state , and the effect of a non - vanishing @xmath10 in presence of a finite current quark mass . based on these analyses , we conclude that @xmath5 condensation is forbidden in realistic neutral quark matter within this effective model .
the njl lagrangian of the model is given for a two - flavor quark matter at a finite chemical potential @xcite @xmath11~ , \label{eq : lagr}\end{aligned}\ ] ] here @xmath12 denotes the electron field , and @xmath13 is the quark spinor with dirac , color and flavor indices ( implicitly summed ) .
the bare quark mass is @xmath14 and @xmath15 the coupling constant .
the electrical chemical potential @xmath10 is necessary to keep the system electrically neutral @xcite , while @xmath16 serves as the isospin chemical potential in the hadron sector , @xmath17 .
the quark chemical potential matrix @xmath18 is defined in flavor - color space as @xmath19 , where @xmath20 denotes the identity matrix in color space , @xmath6 is the quark chemical potential related to the conserved baryon number . in the mean field approximation ( mfa ) , we examine the possibility that the ground state develops condensation in the @xmath21 and/or @xmath22 channels , where @xmath23 denotes the pauli matrices . due to the absence of driving forces , we find @xmath24 is always zero
. for convenience we denote @xmath25 and @xmath26 . in the numerical analyses ,
we have fixed @xmath27 mev and @xmath28 so that the model reproduces @xmath29 mev , @xmath30 , and @xmath31 mev in the vacuum with @xmath32 mev .
the pion condensation `` in the vacuum '' ( @xmath33 ) has been investigated with the chiral perturbation approach @xcite and within njl model @xcite . in both studies the threshold for the onset of the condensation
is found to be @xmath34 , i.e. , when the absolute value of the electric chemical potential equals the vacuum pion mass .
we now consider neutral quark matter at @xmath35 and @xmath36 and we want to study the relation between the threshold of @xmath5 condensation at finite density and the in - medium pion masses in the neutral ground state ( for further recent studies at finite baryon and/or isospin chemical potential see also refs .
@xcite ) . in ref .
@xcite , it is shown that at the physical point @xmath32 mev , there is no room for @xmath5 condensation in the neutral phase ( similar results obtained also in ref .
even though the picture changes when the current mass is lowered , for our discussion it is enough to state that we consider a current quark mass of the order of 10 kev . in fig .
[ fig : masse ] we plot @xmath37 and @xmath38 in the neutral phase as a function of @xmath6 . in this figure @xmath39 , @xmath40 denote the in - medium pion masses defined by the poles of the pion propagators in the rest frame , computed in the randomt phase approximation ( rpa ) to the bethe - salpeter ( bs ) equation .
the positive and negative solutions of the bs equation in @xmath41 corresponds to the excitation gaps for @xmath42 and @xmath43 , which are @xmath44 and @xmath45 , respectively . from fig .
[ fig : masse ] we notice that the transition to the pion condensed phase is of second order and it occurs at the point where @xmath46 . for a more detailed mathematical discussion of the numerical results shown in fig .
[ fig : masse ] we refer to @xcite .
, the pion condensate @xmath38 , and meson masses as a function of @xmath6 at @xmath36 in the neutral phase for a toy value of the current quark mass @xmath47 kev.,width=264 ]
in the past years , a large number of the analysis about @xmath5 condensation have been performed in the chiral limit . in this section
we investigate the role of the finite current quark mass in @xmath5 condensation . in order to do this , we set the cutoff @xmath49 and the coupling @xmath15 to the values specified above and we treat @xmath0 as a free parameter . as a consequence ,
the pion mass at @xmath50 , @xmath51 in the following , is a free parameter as well .
plane.,width=264 ] in fig .
[ fig : phasempi ] we report the phase diagram in @xmath52 plane in the neutral case .
the solid line represents the border between the two regions where chiral symmetry is broken and restored .
the bold dot is the critical endpoint of the first order transition .
the shaded region indicates the region where @xmath5 condensation occurs . in the chiral limit @xmath53 our results are in good agreement with those obtained in ref .
indeed , there exist two critical values of the quark chemical potential , @xmath54 and @xmath55 , corresponding to the onset and vanishing of @xmath5 condensation , respectively . when the current quark mass increases , a shrinking of the shaded region occurs till the point @xmath56 for @xmath57 , corresponding to a current quark mass of @xmath58 .
hence the gapless @xmath5 condensation is extremely fragile with respect to the symmetry breaking effect of the current quark mass .
plane at @xmath32 mev.,width=264 ] as a final investigation , in fig .
[ fig : phasediag ] we report the phase diagram of quark matter in the @xmath59 plane when the current quark mass is tuned to @xmath32 mev . at each value of @xmath59
we compute the chiral and pion condensates by minimization of the thermodynamical potential .
the solid line represents the first order transition from the @xmath5 condensed phase to the chiral symmetry broken phase without the @xmath5 condensate .
the bold dot is the critical endpoint for the first order transition , after which the second order transition sets in .
the dashed line indicates the first order transition between the two regions where chiral symmetry is broken and restored , respectively .
the dot - dashed line is the neutrality line @xmath60 which is obtained by requiring the global electrical neutrality condition , @xmath61 .
the neutrality line manifestly shows the impossibility of finding a @xmath5 condensate , in this physical situation .
we have studied the pion condensation in two - flavor neutral quark matter using the nambu jona - lasinio model of qcd at finite density .
we have investigated the role of electric charge neutrality , and explicit symmetry breaking via quark mass , both of which control the onset of the charged pion condensation .
we show that the equality between the electric chemical potential and the `` in - medium '' pion mass , @xmath62 , as a threshold , persists even for a composite pion system in the medium , provided the transition to the pion condensed phase is of the second order ( the same qualitative picture has been found also in a recent analysis in the massive gross - neveu model @xcite ) .
moreover , we have found that the pion condensate in neutral quark matter is extremely fragile to the symmetry breaking effect via a finite current quark mass @xmath0 , and is ruled out for @xmath0 larger than the order of 10 kev .
a final comment is in order . in this
study the contribution to finite baryon density comes only from the constituent quarks and not from nucleons .
the presence of nucleons may make the pion condensation even less favorable .
the reason can be found in the tomozawa - weinberg pion - nucleon interaction , which is isospin odd , giving rise to a repulsive pion self - energy at the physical situation where the neutron density is larger than the proton density ( see ref .
@xcite ) . for this reason
an interesting investigation may consider the pion condensation with the presence of nucleon degrees of freedom within njl model ( see ref .
@xcite for useful prescriptions ) .
sawyer , phys .
* 29 * , 382 ( 1972 ) ; d.j .
scalapino , phys .
* 29 * , 386 ( 1972 ) .
g. baym and d.k .
campbell in : _ mesons and nuclei _
, v. 3 , eds .
m. rho and d. wilkinson ( north holland , 1979 ) .
j. b. hartle , r. f. sawyer and d. j. scalapino , astrophys .
j. * 199 * , 471 ( 1975 ) ; a. b. migdal , g. a. sorokin , o. a. markin and i. n. mishustin , phys .
b * 65 * , 423 ( 1976 ) ; o. maxwell , g. e. brown , d. k. campbell , r. f. dashen and j. t. manassah , astrophys .
j. * 216 * , 77 ( 1977 ) ; a. ohnishi , d. jido , t. sekihara and k. tsubakihara , phys .
c * 80 * , 038202 ( 2009 ) .
h. abuki , m. ciminale , r. gatto , n. d. ippolito , g. nardulli and m. ruggieri , phys .
d * 78 * , 014002 ( 2008 ) ; h. abuki , m. ciminale , r. gatto , g. nardulli and m. ruggieri , phys .
d * 77 * , 074018 ( 2008 ) ; h. abuki , r. anglani , r. gatto , g. nardulli and m. ruggieri , phys .
d * 78 * , 034034 ( 2008 ) .
d. ebert and k. g. klimenko , arxiv:0911.1944 [ hep - ph ] .
m. matsuzaki , arxiv:0909.5615 [ hep - ph ] ; j. o. andersen , l. t. kyllingstad and k. splittorff , arxiv:0909.2771 [ hep - lat ] ; h. abuki , r. gatto and m. ruggieri , phys .
d * 80 * , 074019 ( 2009 ) ; d. ebert and k. g. klimenko , arxiv:0902.1861 [ hep - ph ] ; h. abuki , t. brauner and h. j. warringa , eur .
j. c * 64 * , 123 ( 2009 ) . | we have studied the phenomenology of pion condensation in 2-flavor neutral quark matter at finite density with nambu jona - lasinio ( njl ) model of qcd .
we have discussed the role of the bare quark mass @xmath0 and the electric chemical potential @xmath1 in controlling the condensation .
the central result of this work is that the onset for @xmath2-condensed phase occurs when @xmath3 reaches the value of the in - medium pion mass @xmath4 provided the transition is of the second order , even for a composite pion system in the medium .
finally , we have shown that the condensation is extremely fragile with respect to the explicit chiral symmetry breaking via a finite current quark mass . |
the pierre auger observatory is a facility to study the origin of ultra high energy cosmic rays by a combined measurement of their flux , anisotropy and mass composition .
the southern site of the observatory is located near the town of malarge in argentina and its construction has been completed at the end of 2008 . even during its assembly , data have been collected in a stable manner since january 2004 .
the lateral densities of charged air shower particles at ground level are measured with the surface detector ( sd ) , an array of 1600 water cherenkov detectors distributed on an area of 3000 km@xmath0 . in addition , the longitudinal air shower development can be observed with the fluorescence detector ( fd ) , consisting of 24 wide - angle schmidt telescopes that overlook the atmospheric volume above the surface detector .
the mass composition of cosmic rays above 10@xmath1 is a key observable to discriminate between different models put forward to explain the softening of the cosmic ray energy spectrum at energies between 10@xmath2 and 10@xmath3 ( the so called ankle , see @xcite for a recent review ) .
moreover , a measurement of the ultra high energy cosmic ray composition is essential to understand the nature of the flux suppression observed above 5@xmath4 @xcite . in this article
we will explain , how the cosmic ray composition can be inferred from air shower measurements with the pierre auger observatory .
an estimate of the primary composition from the measurement of the longitudinal air shower profiles with the fluorescence detector will be given in the next section , followed by the description of the mass sensitive observables of the surface detector .
within the heitler model @xcite of cosmic ray induced particle cascades , the maximum number of electromagnetic particles of a shower with energy @xmath5 and mass @xmath6 is reached at a depth of @xmath7 where @xmath8 denotes the electromagnetic radiation length in air .
the details of the hadronic interactions feeding the shower are subsumed into the parameter @xmath9 where @xmath10 denotes the hadronic interaction length , @xmath11 is the charged particle multiplicity and @xmath12 the energy threshold , below which ionization becomes more important than radiative processes .
the change of @xmath13 per logarithm of energy is called elongation rate @xcite : @xmath14 thus , given a prediction of @xmath15 and @xmath16 from air shower simulations , the primary mass composition can be estimated from the average shower maximum @xmath17 and its relative change with energy from the elongation rate . in the following
we will describe the measurement of these two quantities with the auger fluorescence detector . 2 , @xmath13=733@xmath1818 g / cm@xmath0 ) and los morados ( open squares , e=33@xmath182 , @xmath13=751@xmath1816 g / cm@xmath0 ) .
] in the fd composition analysis only events recorded in hybrid mode @xcite are considered , i.e. the shower development must have been measured by the fluorescence detector and at least one coincident sd station is required to provide a ground level time that is needed for a precise determination of the shower direction . the longitudinal energy deposit profile is reconstructed @xcite from the light recorded by the fd using the fluorescence and cherenkov yields from @xcite and @xcite . with the help of data from the observatory s atmospheric monitoring devices @xcite the attenuation of the light on its way from the shower to the detector
is corrected for and the measured light emission heights are converted to atmospheric depth .
[ fig_profile ] shows an example of a reconstructed longitudinal profile of an air shower , that triggered two of the four auger fluorescence detectors .
the data is fitted with a gaisser - hillas function @xcite to extrapolate the profile to depths outside the field of view of the fluorescence telescopes .
the integral over this function gives the calorimetric energy of the shower and the depth where it is maximal is at @xmath13 . from showers that were observed simultaneously by more than one fluorescence detector
, the difference in the reconstructed shower parameters can be used to infer the resolution of the @xmath13 reconstruction .
it is about 20 g / cm@xmath0 .
the energy and shower maximum can only be reliably measured , if @xmath13 is in the field of view of the telescopes ( otherwise only the rising or falling edge of the profile is detected ) . as illustrated in fig .
[ fig_fov ] , this may lead to a biased event selection , if both , events with a small or large @xmath13 , are not included in the analysis because of their poor reconstruction .
the current auger fluorescence detectors cover an elevation range from @xmath20 to @xmath21 and therefore the observable heights for vertical tracks are between @xmath22 , where r denotes the distance of the shower core to the fluorescence detector .
that is , the farther away from a fluorescence detector a track is detected , the smaller becomes the observable upper slant depth boundary @xmath23 . similarly the lower slant depth boundary @xmath24 becomes larger for near showers .
+ the absolute magnitude of the field of view bias depends of course on the , a priori unknown , distribution of the depth of shower maxima of cosmic ray showers .
it is instructive to study the order of magnitude of the field of view bias for simulated showers with a known composition .
the expected @xmath13 distribution for protons with an energy of 10@xmath25 is shown in fig .
[ fig_fovfid0 ] .
the upper viewable slant depth limits are indicated as arrows for different zenith angles . for vertical tracks
it is around 880 g / cm@xmath0at the southern observatory , which is obviously not sufficient to detect the full @xmath13 distribution .
a convolution of the truncated @xmath13 distributions with the arrival directions of isotropic cosmic rays ( @xmath26 ) leads to an effective bias for the measurement of the mean of the distribution . as can be seen in the upper panel of fig .
[ fig_fovfid1 ] , this bias can be as large as 12 g / cm@xmath0 . due to the unknown composition and the theoretical uncertainties of air shower simulations it can not be corrected for a posteriori , since there is no unambiguous correction given a measured @xmath17 . + the strategy followed in this analysis
is therefore to ignore showers that have a too small viewable slant depth range .
the rejection of showers with @xmath27950 g / cm@xmath0almost completely removes the bias on @xmath17 in the simulation , cf .
lower panel of fig .
[ fig_fovfid1 ] .
the _ fiducial volume cuts _ that ensure an unbiased @xmath17 can be derived directly from data as illustrated in fig .
[ fig_fovfid2 ] : the dependence of @xmath17 on the slant depth boundaries is studied and events are only selected in the region where @xmath17 is constant . the @xmath17 results presented here were derived from hybrid data recorded between december 2004 and april 2007 . in order to ensure a good @xmath13 resolution , the following quality cuts were applied to the event sample : the reconstructed @xmath13 should lie within the observed shower profile and the reduced @xmath28 of a fit with a gaisser - hillas function should not exceed 2.5 . moreover , insignificant shower maxima are rejected by requiring that the @xmath28 of a linear fit to the longitudinal profile exceeds the gaisser - hillas fit @xmath28 by at least four .
finally , the estimated uncertainties of the shower maximum and total energy must be smaller than 40 g / cm@xmath0 and 20% , respectively . moreover , the fiducial volume cuts explained in the last section are applied .
+ the systematic uncertainties of the atmospheric properties , event reconstruction and mass acceptance add up to a total estimated systematic uncertainty of @xmath17 that is @xmath29 15 g / cm@xmath0 at low energies and @xmath29 11 g / cm@xmath0 above as a function of energy @xcite . ]
10@xmath2 .
+ in fig .
[ fig_er ] the mean @xmath13 as a function of energy is shown along with predictions from air shower simulations @xcite .
as can be seen , the measurement favors a mixed composition at all energies . a simple linear fit , yields an elongation rate of 54@xmath182 ( stat . ) g / cm@xmath0/decade , but does not describe our data very well ( @xmath30ndf@xmath31 , p@xmath323% ) . from other experiments @xcite .
] allowing for a break in the elongation rate at an energy @xmath33 leads to a satisfactory fit with @xmath30ndf@xmath34 , p@xmath3563% and @xmath36 ( stat . )
g / cm@xmath0/decade below @xmath37 and @xmath38 ( stat . )
g / cm@xmath0/decade above this energy .
this fit is indicated as a gray line in fig .
[ fig_er ] . due to the uncertainties of hadronic interaction at highest energies ,
the interpretation of these elongation rates is , however , ambiguous .
using qgsjetii the data suggests a moderate lightening of the primary cosmic ray composition at low energies and an almost constant composition at high energies , whereas the epos proton elongation rate is clearly larger than the measured one at high energies , which would indicate a transition from light to heavy elements .
theses ambiguities will be partially resolved by the analysis of the @xmath13 fluctuations as an additional mass sensitive parameter . a comparison with measurements of other experiments is presented in fig . [ fig_others ] .
taking into account the individual systematic uncertainties of each experiment there is a reasonable agreement of the @xmath17 values .
it is , however , worthwhile noting that hires uses a different definition of @xmath13 ( see @xcite ) that is about 10 g / cm@xmath0shallower than the one commonly used ( depth of non - parametric shower maximum ) . if this is taken into account , the elongation rates of auger and hires show a discrepancy in the @xmath13-scale of about 30 g / cm@xmath0 .
as we have explained in the last section , the longitudinal air shower development is a good tracer for the primary composition that can be measured very precisely with fluorescence telescopes .
however , these telescopes can only take data in moonless nights , whereas the surface detector has a duty cycle of almost 100% and correspondingly , the available event statistics are about a factor ten higher . because of this , and the obviously independent systematic uncertainties , it is worthwhile to study ground level observables that are sensitive to the primary composition as well . since the amount of electromagnetic particles at ground depends for a given energy on the distance to the shower maximum it depends also on the primary mass .
moreover , the number of ground level muons evolves differently with energy for different primary masses . from the heitler model one expects @xmath39 where @xmath40 depends on the multiplicity and inelasticity of hadronic interaction ( and is thus , as the above @xmath15 , subject to theoretical uncertainties ) .
+ since the water cherenkov detectors of the pierre auger observatory can not explicitly discriminate between the electromagnetic and muonic air shower components on an event by event basis ( but see sec .
[ sec : outlook ] ) , a variety of indirect experimental variables are currently investigated to relate the surface detector data to the primary mass @xcite : the signal rise time , its asymmetry and the signal shape analysis . for each event ,
the water cherenkov detectors record the signal as a function of time ( fadc traces ) .
this provides a possibility to distinguish the muonic and electromagnetic component , since the former travel in almost straight lines through the atmosphere , whereas the latter undergo multiple scattering on their way to ground .
therefore , the two components have different path lengths and arrival times at ground .
this is illustrated in fig .
[ fig_vem ] , where an example of a simulation @xcite of the surface detector response is shown . the _ rise time
_ @xmath41 ]
@xcite is the time it takes until the cumulative signal grows from 0.1 to 0.5 .
it is a measure of the muon to electron ratio in a tank and depends on the primary mass , angle of incidence @xmath42 and distance to the shower core @xmath43 .
the @xmath43 and @xmath42 dependence is marginalized by a comparison to the average rise time @xmath44 and the pull @xmath45 is averaged over the stations of an event to yield @xmath46 @xcite .
this quantity could in principle be compared to air shower simulations . however , as previous studies showed , the number of muons measured in auger can not be reproduced by these calculations @xcite .
therefore , instead of a direct comparison of @xmath46 to air shower simulations , it is related to the shower maximum . for this purpose ,
the subset of events that have been detected by both , the fluorescence and surface detector , are used to establish the relation between @xmath46 and @xmath13 . as can be seen in fig .
[ fig_risetime ] , there is a linear correlation between the two variables , with which the surface detector data can be calibrated to yield @xmath47 , with which it is possible to measure the elongation rate with surface detector data .
= 54@xmath48 , e=55@xmath182 ) ] the sd stations measure one stage of the shower development for near vertical events . in case of inclined showers , however , considerably different shower ages can be observed , depending on whether the station is up- or downstream of the incoming shower direction .
this is illustrated in fig .
[ fig_asymetry ] , where the signal traces of two stations ( both at the same shower plane distance @xmath43 ) are shown .
the upstream trace is rather broad , which is compatible with a large fraction of electromagnetic particles .
the downstream trace is narrower indicating that most of the electromagnetic component has been attenuated and the signal is dominated by muons .
the corresponding _ asymmetry _
@xcite in the rise time is thus a measure of the longitudinal development of the ratio of electromagnetic to muonic particles .
the amplitude of the asymmetry changes with the zenith angle , i.e. distance to the shower maximum .
the angle at which it reaches its maximum can be used as an estimator for the primary composition . since a muon deposits much more energy ( typically 240 mev ) in a water tank than an electron or photon ( about 10 mev )
, spikes are produced over the smoother electromagnetic background in the fadc time traces .
therefore , muons can be identified by searching for sudden variations , in the signal from one fadc bin to the next ( cf .
fig [ fig_vem ] ) .
the expected distributions of variations for purely electromagnetic and muonic traces can be fitted to the measured distribution as shown in fig .
[ fig_muons ] . in that way , the muonic signal can be determined on a statistical basis with a resolution of about 25% and the number of muons as a function of energy can be compared to predictions from air shower simulations to estimate the primary composition .
in this article we presented the measurement of the elongation rate from data collected with the fluorescence telescopes of the pierre auger observatory . when compared to predictions from air shower simulation
the @xmath17 data favors a mixed composition at all energies .
this measurement will be soon updated with larger statistics and an analysis of the fluctuations of the shower maximum . at lower energies
the systematic uncertainties will be reduced by additional telescopes that cover the field of view from 30 to 60 degree @xcite .
+ a variety of surface detector observables are sensitive to the mass composition .
these parameters have a somewhat worse mass resolution than @xmath13 , which is , however , outweighed by the large event statistics collected by the ground array . due to the hybrid design of the pierre auger observatory ,
the correlation between the lateral and longitudinal shower parameters can be studied and additional muon detectors will soon allow for an event - by - event measurement of the muon content of a shower @xcite . 1 j. abraham et al .
[ pierre auger coll .
] , , 101:061101 , 2008 ; r. abbasi et al .
[ hires coll .
] , , 100:101101 , 2008 . v. berezinsky , , arxiv:0710.2750 [ astro - ph ] .
j. linsley and a.a .
watson , phys .
* 46 * ( 1981 ) , 459 ; t.k .
gaisser et al . , proc .
16th icrc * 9 * ( 1979 ) , 258 ; j. linsley , proc .
15th icrc * 12 * ( 1977 ) , 89 .
j. matthews , astropart .
* 22 * ( 2005 ) , 387 ; w. heitler , oxford univ . press , 1954 .
p. sommers , astropart .
phys . * 3 * ( 1995 ) , 349 .
m. unger et al . , nucl .
instrum .
* a588 * ( 2008 ) , 433 ; b. dawson , m.giller and g. wieczorek , proc . 30th icrc ( 2007 ) ; d. gora et al . , astropart . phys .
* 24 * ( 2006 ) , 484 ; m. d. roberts , j. phys .
* g31 * ( 2005 ) , 1291 ; m. ave et al .
[ airfly coll .
] , astropart .
* 28 * , ( 2007 ) 41 ; m. nagano et al .
, astropart .
* 22 * ( 2004 ) , 235 .
f. nerling et al .
, astropart .
* 24 * ( 2006 ) , 421 .
s. y. benzvi for the pierre auger coll .
30th icrc ( 2007 ) ; b. keilhauer for the pierre auger coll .
29th icrc * 7 * ( 2005 ) , 123 . t. k. gaisser and a. m. hillas , proc .
15th icrc * 8 * ( 1977 ) , 353 .
t. bergmann et al . , astropart .
* 26 * ( 2007 ) 420 ; n. n. kalmykov , s. s. ostapchenko and a. i. pavlov , nucl . phys . proc
. suppl . * 52b * ( 1997 ) 17 ; r. engel , t. k. gaisser , t. stanev and p. lipari , proc .
26th icrc ( 1999 ) , 415 ; s. ostapchenko , nucl .
* 151 * ( 2006 ) 143 ; t. pierog and k. werner , phys .
* 101 * ( 2008 ) 171101 .
bird et al .
[ hires coll .
] , phys .
* 71 * ( 1993 ) , 3401 ; t. abu - zayyad et al .
[ hires / mia coll . ] , astrophys . j. * 557 * , ( 2001 ) , 685 .
j. belz for the hires coll .
, these proceedings .
m. unger for the pierre auger coll .
30th icrc , merida ( 2007 ) , arxiv:0706.1495 [ astro - ph ] .
s. argiro _ et al .
_ , nucl .
instrum .
meth . a * 580 * ( 2007 ) 1485 d. heck , g. schatz , t. thouw , j. knapp and j. n. capdevielle , fzka-6019 ( 1998 ) r. walker and a.a .
watson , j. phys .
g * 7 * ( 1981 ) , 1297 .
r. engel for the pierre auger coll . ,
30th icrc ( 2007 ) , arxiv:0706.1921 [ astro - ph ] .
b. e. smith , phd thesis university of leeds ( 2008 ) .
dova for the pierre auger coll .
29th icrc ( 2005 ) , arxiv : astro - ph/0507171 m. risse for the pierre auger coll . ,
these proceedings , arxiv:0901.2525 [ astro - ph ] .
healy for the pierre auger coll .
30th icrc ( 2007 ) , arxiv:0706.1569 [ astro - ph ] .
h. o. klages for the pierre auger coll .
30th icrc , ( 2007 ) .
a. etchegoyen for the pierre auger coll .
30th icrc , ( 2007 ) , arxiv:0710.1646 [ astro - ph ] . | we report on studies of the composition of ultra high energy cosmic rays with the pierre auger observatory .
the detection of longitudinal air shower profiles with the fluorescence detector is described and the measurement of the average shower maximum as a function of energy is presented .
furthermore , mass sensitive parameters that can be obtained from the observatory s surface detector data are discussed . |
aja acknowledges the support of epsrc under grant number gr / s28631/01 . | the effective pair potential between mesoscopic charged particles in a neutralising background medium takes a yukawa form @xmath0 with screening length @xmath1 .
we consider a dilute suspension of such yukawa particles dispersed in a solvent with correlation length @xmath2 and show that the yukawa interaction is ` screened ' if the pair potentials between solvent particles exhibit yukawa decay with the same screening length @xmath1 .
however , if the solvent pair potentials are shorter ranged than the solute yukawa potentials , then the effective potential between pairs of solute particles is ` unscreened ' , i.e. the effective potential between the solute particles is equal to the bare potential at large particle separations .
it is well known that the effective potential between charged particles immersed in a neutralising mobile medium is shorter ranged than the ` bare ' coulomb interaction between the particles the interaction is screened . at large interparticle separations
@xmath3 , the effective pair potential between particles @xmath4 . screening effects due to the neutralising medium are described by an inverse screening length @xmath5 @xcite .
such yukawa potentials are ubiquitous in charged colloidal systems arising in the classic linearised poisson boltzmann or derjaguin landau verwey overbeek ( dlvo ) theories for the effective potential between spherical charged colloids in solution @xcite
. other systems in which effective yukawa potentials arise include those between dust grains in plasma systems @xcite and some proteins in solution @xcite .
the yukawa potential arises not only in diverse physical problems , but is also interesting in its own right @xcite .
note that the motivation for considering effective potentials in complex systems is driven by the fact that a theoretical treatment of the full mixture is often very difficult , particularly when there is a big size asymmetry between different components of the mixture .
it is profitable to integrate out the degrees of freedom of one or more of the components , thereby incorporating formally their influence into an effective hamiltonian for the remaining particles @xcite . in this letter
, we consider a simple model for binary mixtures of charged particles suspended in a neutralising medium .
we examine a dilute suspension of yukawa particles in a solvent with correlation length @xmath2 and we investigate the ( additional ) screening effect that the solvent particles have on the effective interaction between the solute yukawa particles . specifically
, we consider a class of binary mixtures in which the pair potentials are of the form @xmath6 , where indices @xmath7 label the two different species and where @xmath8 is a contribution to @xmath9 that is short ranged in comparison with the yukawa term .
the amplitudes @xmath10 depend on the charges on the particles .
for example , in modelling a binary mixture of charged spherical colloids , we would take @xmath8 to be a hard sphere interaction potential and @xmath11 , where @xmath12 is the ( renormalised ) charge on the colloid of species @xmath13 @xcite .
however , in what follows directly we will not make a particular choice for @xmath8 or for the sign of @xmath10 .
we denote species 1 as the solvent and species 2 as the solute .
we focus on the limit of the number density of the solute particles @xmath14 and investigate the form of the effective potential @xmath15 between the solute particles in this limit . in particular , we obtain the general result that when @xmath16 , @xmath17 , so that there is a yukawa tail @xmath18 present in @xmath19 and @xmath20 , @xmath15 decays faster than @xmath21 as @xmath22 provided the solvent bulk correlation length @xmath23 . on the other hand ,
when @xmath24 and there is no yukawa tail in @xmath19 and @xmath20 , the effective potential @xmath25 for @xmath22 .
this yukawa ` screening ' effect is independent of the sign of the ` charges ' , i.e. independent of the signs of @xmath10 .
the results can be re stated in terms of the solvent - mediated potential @xmath26 , defined via @xmath27 @xcite .
@xmath26 depends on the nature of the solvent and on the solvent - solute interaction . if @xmath28 faster than @xmath21 , as @xmath22 , then we describe the solute
solute interaction as ` unscreened ' .
however , if @xmath26 partially or completely cancels the bare potential @xmath21 at large @xmath3 , we describe the solute
solute interaction as ` screened ' .
in certain mixtures where the bare interactions are purely repulsive we find ` super - screening ' , i.e. @xmath15 can be _ attractive_. we may determine the effective interaction between two solute particles at infinite dilution from the solute
solute radial distribution function @xmath29 using the well known result @xcite @xmath30 , \label{eq : smp1}\ ] ] where @xmath31 is the inverse temperature and @xmath29 is evaluated in the limit @xmath32 .
for large @xmath3 , when @xmath33 is small , it follows that @xmath34 where @xmath35 , and in order to determine the asymptotic behaviour of @xmath15 we must ascertain that of @xmath36 .
the total pair correlation functions @xmath37 are related via the ornstein zernike ( oz ) integral equations @xcite to a set of pair direct correlation functions @xmath38 .
the oz equations can be solved formally in fourier space and the solution written ( for arbitrary concentration ) as @xmath39 where @xmath40 denotes the three - dimensional fourier transform of @xmath37 .
the three functions share the same denominator @xmath41[1-\rho_{2}\hat{c}_{22}(q ) ] -\rho_{1}\rho_{2}\hat{c}_{12}(q)^{2 } , \label{eq : den}\ ] ] where @xmath42 is the density of species @xmath13 .
the numerators @xmath43 may be obtained from refs .
@xcite or @xcite .
taking the inverse fourier transform of eq .
( [ eq : oz2 ] ) , evaluating the integral by a contour integration around a semicircle in the upper half of the complex @xmath44 plane and assuming that the singularities of @xmath40 are simple poles one can express @xmath37 as a sum of contributions from the set of poles at @xmath45 in the upper half of the complex plane @xcite : @xmath46 where @xmath47 is the amplitude associated with the pole at @xmath48 .
the poles are obtained from the set of solutions of @xmath49 .
the amplitude @xmath47 is related to the residue @xmath50 of @xmath51 by @xmath52 .
the poles are either purely imaginary , @xmath53 , or occur as a conjugate pair @xmath54 @xcite .
a purely imaginary pole gives a monotonic contribution to @xmath55 of the form @xmath56 .
a conjugate pair of poles gives a damped oscillatory contribution of the form @xmath57 , where @xmath58 and @xmath59 denote the amplitude and phase respectively @xcite .
in general , there are an infinite number of poles .
however , the asymptotic decay @xmath22 is determined by the pole(s ) with the smallest imaginary part @xmath60 ( or @xmath61 ) .
away from any critical points , the direct pair correlation functions @xmath38 are known to decay as @xmath62 , @xmath22 @xcite .
when @xmath9 has a yukawa contribution it is convenient to separate @xmath38 in the following way : @xmath63 , which defines @xmath64 , the short - ranged piece in @xmath38 , dependent on the form of @xmath8 and on the state point . in fourier space
it follows that @xmath65 where @xmath66 .
we seek the pole(s ) in @xmath40 with smallest imaginary part .
first , we consider the case when there is no yukawa contribution to the solvent potentials , i.e. @xmath67 .
the asymptotic decay of @xmath68 in the pure solvent of species 1 is determined by the pole(s ) in @xmath69 , given by the solution of @xmath70 , with the smallest imaginary part @xmath60 and the bulk solvent correlation length @xmath71 @xcite .
consider now the decay of @xmath37 in the full mixture .
making the separation of @xmath72 given by eq .
( [ eq : dcfsep2 ] ) and substituting into eq .
( [ eq : den ] ) we obtain @xmath73 where @xmath74 and @xmath75 -\rho_1\rho_2[\hat{c}_{12}^{sr}(q)]^2 $ ] .
the equation @xmath76 has a solution @xmath77^{1/2}$ ] , which implies that in the limit @xmath14 there is a pure imaginary pole at @xmath78 , provided the ratio @xmath79 remains finite in this limit . since we have assumed that there is no yukawa contribution in @xmath19 , there is no pole at @xmath80 for the pure solvent and @xmath81 is non - zero and finite .
if in addition , we assume that both @xmath82 and @xmath83 are finite it follows that in the limit @xmath14 , @xmath84 for @xmath85 .
the amplitude of the contribution to @xmath37 from this purely imaginary pole at @xmath80 is given by @xcite : @xmath86 where the prime denotes the derivative with respect to @xmath44 .
it is straightforward to show that in the limit @xmath14 , @xmath87 . using this result and evaluating the numerators @xmath88
, we find that the amplitudes of the contributions to @xmath37 from the purely imaginary pole at @xmath89 are : @xmath90 these results obey the rule @xmath91 , which general considerations demand @xcite .
note that in the limit @xmath14 the amplitudes @xmath92 .
thus the contributions from the pole at @xmath93 to the decay of @xmath68 and @xmath94 are vanishingly small , as one would expect on physical grounds .
note further that in the same limit , @xmath95 tends to a non - zero constant value @xmath96 which is independent of any properties of the solvent , i.e. the pole at @xmath97 gives a contribution to @xmath98 of the form @xmath99 .
thus from eq .
( [ eq : smp2 ] ) one finds that as @xmath22 , @xmath25 . in summary ,
when @xmath67 , the effective interaction between the solute particles is identical to the bare interaction as @xmath22 ; the solute - solute interaction is ` unscreened ' . of course
the argument we have presented supposes that the pole @xmath100 is the leading - order one ; i.e. has the smallest imaginary part .
if the pure solvent has a correlation length @xmath101 then one expects the asymptotic decay of all three correlation functions @xmath102 and @xmath15 is _ longer _
ranged than @xmath21 .
henceforward we restrict consideration to cases where @xmath103 .
we turn now to the more realistic case when @xmath104 , i.e. all three pair potentials have yukawa tails . proceeding in a similar manner as above
( see also the appendix in ref .
@xcite ) the denominator function ( [ eq : den ] ) takes the form @xmath105 , where @xmath106[1-\rho_2 c_{22}^{sr}(q)]-\rho_1\rho_2[c_{12}^{sr}(q)]^2 $ ] , @xmath107\rho_2 \alpha_{22 } + [ 1-\rho_2 c_{22}^{sr}(q)]\rho_1\alpha_{11 } + 2\rho_1\rho_2c_{12}^{sr}(q ) \alpha_{12}$ ] and @xmath108 .
one set of solutions to the equation @xmath76 is given by @xmath109 .
this leads to purely imaginary poles at @xmath110 , provided we assume that the functions @xmath111 are well - behaved ( finite and differentiable ) on the imaginary axis around @xmath112 .
the leading order pole corresponds to @xmath113 , and in the limit of vanishing density @xmath114 ( i.e. @xmath115 ) there is a pole at @xmath116 , so that @xmath117 , @xmath22 . for small concentrations of species 2 we taylor expand @xmath118 in powers of @xmath119 and find the leading order pole is given by @xmath120 using eq .
( [ eq : amp ] ) we calculate the amplitudes @xmath121 of the contributions from this pole to the correlation functions @xmath37 : @xmath122 note that the coefficients of the leading order terms are independent of @xmath64 and that the amplitudes obey the rule @xcite @xmath123 . in the simplest model of a charged system one expects @xmath124 since @xmath125 , the product of the charges on each species .
we refer to this situation as the ` ideal ' mixing rule @xcite . then the coefficients of the leading order terms in eq .
( [ eq : amplitudes_2 ] ) vanish identically and for all three @xmath37 the amplitudes corresponding to the pole at @xmath126 will be zero in the limit @xmath14 .
thus the asymptotic decay of @xmath36 ( and therefore of @xmath15 ) is determined by the next order pole , which generally has @xmath127 , so that @xmath128 decays as @xmath129 , i.e. _ faster _ than @xmath21 , so that the interaction between species 2 solute particles is ` screened ' . in the physical systems
where @xmath130 , there will still be partial screening since one expects the difference @xmath131 to be small and then @xmath15 will decay , as @xmath22 , with the same exponential decay length @xmath1 as the bare potential , but with a reduced amplitude proportional to @xmath131 .
we now display results for a number of model systems that confirm these general predictions .
the effective solute - solute potential @xmath15 for solvent density @xmath132 and @xmath32 calculated using the hnc closure .
solid line : system a , the point yukawa mixture , with @xmath133 and dot - dashed line with non - ideality parameter @xmath134 .
the dashed line is for system b ( @xmath133 ) where the solvent potentials decay faster than for the yukawa case .
results are compared to the bare potential @xmath21 ( dotted line ) . the inset displays the radial distribution functions @xmath135 ( solid line ) and @xmath136 ( dashed line ) for @xmath133 . on this scale ,
there is no visible difference between results in systems a and b.,width=302,height=226 ] the first class of systems is a mixture in which the particles interact via purely repulsive ( point ) yukawa pair potentials for which @xmath137 and @xmath138 @xcite .
we denote this system a. the structure and phase behaviour is described in ref . @xcite . in this system
the effective potential @xmath15 decays faster than the bare potential @xmath21 reflecting the yukawa decay of the solvent ( species 1 ) potentials see fig .
[ fig:1 ] .
we compare with system b in which @xmath21 is the same as in system a but where the solvent potentials are modified slightly to hasten the yukawa decay at large @xmath3 , i.e. @xmath139 and @xmath140 . in both systems we choose pair potential parameters @xmath141 , @xmath142 and the mixing rule @xmath143 where the parameter @xmath144 measures the degree of non - ideality @xcite .
the state point has solvent density @xmath132 and solute density @xmath145 , corresponding to the dilute limit . in system
b we choose @xmath146 , which is sufficiently small that the solvent radial distribution functions @xmath135 and @xmath136 are almost indistinguishable from those in system a @xcite . fig .
[ fig:1 ] displays @xmath15 calculated using the hnc closure to the oz equations @xcite , which is expected to be a very reliable approximation for this model fluid @xcite . for system
a with @xmath133 ( solid line ) there is ` screening ' ; @xmath15 is much shorter ranged than the bare potential @xmath21 ( dotted line ) . in system b with @xmath133 (
dashed line ) @xmath15 is indistinguishable from that in system a at small @xmath3 , but for @xmath147 the results differ significantly there is a maximum near @xmath148 and for @xmath149 , @xmath25 ; the solute - solute interaction is ` unscreened ' . for @xmath134 in system
a ( dot - dashed line ) @xmath15 has a pronounced minimum near @xmath150 and is _ attractive _ for larger @xmath3 as the amplitude @xmath151 see eq .
( [ eq : amplitudes_2 ] ) .
we refer to this scenario as ` super screening ' .
it is remarkable that such effective attraction arises in a system where all the bare interactions are purely repulsive and this constitutes a dramatic signal that for @xmath152 the fluid exhibits liquid - liquid phase separation when the densities of the two components are sufficiently high @xcite .
the effective potential @xmath15 for solvent density @xmath153 and @xmath154 calculated using the py closure for hard - core yukawa mixtures .
dashed line : system c ( @xmath155 , @xmath156 , @xmath157 ) , dot - dashed line : system d ( @xmath158 ) and solid line : system e ( @xmath159 , @xmath160 , @xmath161 ) . in systems
c and d , @xmath15 is shorter ranged than the bare potential @xmath21 ( dotted line ) but in system e , @xmath162 for @xmath163 ; there is no screening .
the inset displays the corresponding total correlation function @xmath68 for systems c , d and e.,width=302,height=226 ]
the second class of systems is that in which the particles have a hard core interaction mimicking ( charged ) spherical colloids . in system
c the pair potentials are of the form @xmath164 , where @xmath165 for @xmath166 and 0 for @xmath167 , the hard - sphere diameter @xcite .
we choose the parameters @xmath168 , @xmath155 , @xmath169 , @xmath161 , corresponding to the case where the sign of the ` charge ' is the same on both species but the magnitude @xmath170 .
we compare with system d in which the ` charges ' on the particles have the same magnitude as in system c but the opposite sign , @xmath171 ; all the other parameters remain the same .
the final comparison is with system e in which the solvent ( species 1 ) particles are neutral hard spheres : @xmath24 while @xmath21 is the same as in systems c and d ( with @xmath172 ) . in fig .
[ fig:2 ] we display the results for @xmath15 calculated using the percus yevick ( py ) closure to the oz equations @xcite for these three systems at the solvent density @xmath153 and solute density @xmath173 .
py should be reasonably accurate for such hard core systems with additional weak yukawa tails .
we see that in system c ( dashed line ) and d ( dot - dashed line ) @xmath15 decays much more rapidly than in system e ( solid line ) where @xmath162 for @xmath163 , i.e. the solute - solute interaction is unscreened . `
screening ' occurs in both systems c and d. perhaps surprisingly , the screening effect is stronger in system c where the solvent particles have the same sign ` charge ' as the solute particles . note that in case d where the solute particles have the opposite sign ` charge ' to the solvent particles , @xmath15 can become quite large and positive at small and intermediate values of @xmath3 , as the magnitudes of the charges are increased . in this case
the screening effect discussed here , which applies to the ultimate @xmath22 decay of @xmath15 , becomes visible only at very large values of @xmath3 .
in this letter we have shown that in yukawa mixtures the interaction between particles of one species can be ` screened ' by the other species .
the screening effect does not depend on the sign of the ` charges ' on the particles see eq .
( [ eq : amplitudes_2 ] ) one observes ` like charge screening ' , in contrast with standard coulombic screening .
our results may have implications for the effective interactions in models of charged binary colloidal suspensions . |
protoplanetary disks are composed of the circumstellar material that feeds stars and planets .
these systems evolve from early gas- and dust - rich phases to phases of dispersal , eventually unveiling planetary systems that form within them ( e.g. * ? ? ?
evidence is growing that there is at least one mode of disk dispersal , working from the inside out .
observations of a wavelength - dependent decrease in infrared excesses ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , or spatially resolved cavities in continuum dust emission ( e.g. * ? ? ?
* ; * ? ? ?
* ) , indicate that dust grains in the inner disk regions are depleted before those at larger radii .
disks that show these properties have been proposed to be in a transition phase ( transitional disks " ) toward their dispersal ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the concept of transitional " disk has been developing ever since its early definitions , under the evidence provided by larger samples , more detailed observations , and improved models ( e.g. * ? ? ?
* ; * ? ? ?
understanding the properties , origin , and evolution of inner disk dispersal and of gaps in protoplanetary disks bears the potential to inform on when , where , and in what physical and chemical environments planets form .
models that produce an inside - out dispersal of disks include the effects of photoevaporation by stellar x - ray and uv radiation ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . in these models ,
disk gaps are opened where thermal heating launches the gas into a photoevaporative wind by providing enough energy to exceed the gravitational potential of the star .
after a gap is opened , the inner disk is expected to be drained onto the central star by viscous accretion , and the inner hole would rapidly grow to larger radii under the effect of photoevaporation .
disks may also disperse under removal of gas by magnetohydrodynamic ( mhd ) winds ( e.g. * ? ? ?
* ; * ? ? ?
another process that has been shown to open gaps in disks is the formation of giant planets ( e.g. * ? ? ?
* ; * ? ? ?
* among several others ) . where and when gaps may be opened depends on the efficiency of giant planet core formation ( usually assumed to increase close to the water snow line , at several to @xmath9au , e.g. * ? ? ?
* ; * ? ? ?
* ) , on their migration , and on the number of planets ( e.g. * ? ? ?
* ; * ? ? ?
in this scenario , the inner region within the gap opened by the planet will also be depleted by viscous accretion onto the central star , eventually forming an inner disk hole . while there is no single model that can entirely match the currently observed properties of transitional " disks
, there is evidence that the photoevaporative and planet scenarios are both happening in disks @xcite .
models are now exploring also the combination of multiple processes , toward a more comprehensive understanding of the evolution of planet - forming disk regions .
recent models by @xcite , for instance , explore the combination of dust evolution and disk photoevaporation , showing how the gas - to - dust ratio decreases in photoevaporating inner disks at radii @xmath10au , producing favorable conditions for rocky planetesimal formation .
the detailed study of the properties of inner disk gaps and holes is greatly advancing thanks to the increase in spatial and spectral resolution of observing techniques .
direct imaging is able to reveal gap sizes ( if their angular size is large enough to be spatially resolved ) and the relative distribution of dust grain populations ( small versus large grains , * ? ? ?
* ; * ? ? ?
* ) , providing tracers of physical processes that act in inner disks .
gas spectroscopy holds unique potential to unveil the chemical evolution of planet - forming regions .
recent developments in sensitivity and spectral resolution of infrared spectrographs allowed the first large surveys of molecular gas in inner disks able to spectrally- ( and sometimes also spatially- ) resolve gas within dust gaps ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the combined analysis of a large sample of disks recently established rovibrational co ( carbon monoxide ) emission at 4.7@xmath0 m as a tracer of inner gaps in molecular gas .
@xcite found evidence for an inside - out depletion sequence where co gas is progressively removed from @xmath11 out to @xmath12au , providing a new probe of gaps over a radial disk region that extends the region probed by millimeter interferometers ( typically limited to @xmath13au depending on the distance from earth ) . by tracing h@xmath14 emission at ultraviolet wavelengths , @xcite found similar evidence for a removal of molecular gas at small radii in transitional disks .
the combination of dust and gas studies of small to large gaps bears the potential of a comprehensive understanding of the physical and chemical evolution of inner disks during planet formation . as a further contribution to this rapidly growing field
, we report on the depletion of h@xmath14o ( water ) and oh ( hydroxyl ) during formation of inner disk gaps .
c c c c c c c c c c c 2.93 & vlt- & 100,000 & @xmath15 & 800010,000 & 900 & @xmath15 & n / a & 58% ( 24 ) & 11% ( 18 ) & this work , 1 + & crires & & & & & & -15 & -14.6 & + + 1037 & _ spitzer_- & 700 & @xmath16 & 7006000 & 300 - 700 & @xmath17 & 0% ( 5 ) & 6385% ( 64 ) & 018% ( 27 ) & this work , 2 + & irs & & & & & -15.7 & -14.8 & -13.7 & + + 55200 & _ herschel_- & 10005000 & @xmath18 & 1001400 & 100 - 300 & @xmath19 & 0% ( 10 ) & 15% ( 80 ) & 15% ( 27 ) & 3 + & pacs & & & & & -14.6 & -14.5 & -14 & + water vapor emission has been observed in protoplanetary disks between 3 and 540@xmath0 m , with significant differences in detection rates depending on the observed wavelength and on the mass of the central star . the most comprehensive survey of water emission in disks by a single instrument has been obtained by the irs spectrograph on the _ spitzer space telescope _
@xcite , which measured hundreds of water emission lines between 10 and 37@xmath0 m in over 60 disks .
_ spitzer _ observations showed that water emission is strongly dependent on the stellar spectral type and tentatively dependent on the evolutionary stage , with high detection rates in disks around solar - mass stars , low in disks around intermediate - mass stars , and similarly low in a small sample of transitional " disks ( 1/6 , see table [ tab : water_surveys ] and * ? ? ? * ; * ? ? ? * ; * ? ? ? * ) .
water emission at _ spitzer _ wavelengths has been interpreted as coming in most disks from an optically thick surface layer within the snow line radius , at temperatures of 300700 k @xcite .
modeling of the velocity - unresolved _ spitzer _ spectra ( @xmath2@xmath3400kms@xmath4 ) provided estimates of the disk emitting region to the inner few au in disks ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . at 12.4@xmath0 m , within the _
spitzer_-irs range , a few water lines have been resolved in velocity in four disks with vlt - visir and gemini - texes ( @xmath2@xmath315 and @xmath33.5kms@xmath4 respectively , * ? ? ?
* ; * ? ? ?
* ) , providing support to the emitting regions estimated by models for the _ spitzer _ spectra @xcite . at shorter wavelengths ( @xmath323@xmath0 m ) , hotter water ( t @xmath20k ) has been spectrally - resolved and analyzed only in four disks so far , by @xcite and @xcite with keck - nirspec and vlt - crires , and by @xcite with irtf - cshell . at longer wavelengths ( @xmath21@xmath0 m ) ,
colder water has been surveyed in disks by _ herschel_-pacs , finding as low detection rates in disks around solar - mass as in disks around intermediate - mass stars ( @xmath22% as based on the 63.3@xmath0 m line , see references in table [ tab : water_surveys ] and * ? ? ?
table [ tab : water_surveys ] summarizes properties , samples , and detection rates for all major surveys of water vapor in protoplanetary disks , including this work .
the general trends identified to date are shown in figure [ fig : spitzer_spt ] , and are compared to expectations produced by recent models of water in disks @xcite ) ; in the case of @xcite , we use the model spectra kindly provided by the authors .
we scale the models to the line fluxes measured in disks around stars of 0.5 m@xmath23 , to highlight global trends with m@xmath24 . ] .
we investigate the measured line fluxes against the masses of the central stars m@xmath24 , as an alternative way to look at the spectral - type dependencies found by @xcite .
water line fluxes generally increase with stellar luminosity @xcite , and with stellar mass .
models explain this with the radial extent of the warm ( @xmath25k ) disk surface where water vapor is abundant due to ice evaporation and efficient gas - phase formation , which increases to larger disk radii around more luminous stars .
model predictions agree well with the increase in _ spitzer _
line fluxes measured in disks around stars of masses between 0.2 and 1.5 m@xmath23 , although 2040% of disks show significantly lower fluxes , relative to what is expected from a pure luminosity effect .
for m@xmath26m@xmath23 , models produce water line fluxes @xmath27 times stronger than observed for 80 - 100% of the disk sample . the low water detection frequency at _ spitzer _ and crires wavelengths in disks around stars of m@xmath26m@xmath23 is still not understood ( e.g. * ? ? ?
in this work , we test the hypothesis that inside - out inner disk depletion may produce the progressive disappearance of water vapor emission at infrared wavelengths .
if water vapor in inner disks is depleted from smaller to larger disk radii , an essential spectral region to include in the investigation of this process is where the hot water emits , in high - energy transitions at 2.9 - 3@xmath0 m ( e.g. * ? ? ?
* ) . in sections [ sec :
obs ] and [ sec : ana ] we describe the new survey of hot water emission as observed with crires at 2.9@xmath0 m , and the comparative analysis with the co dataset at 4.7@xmath0 m . in section [ sec : rad_trends ] , we report the discovery of a correlation between water line fluxes at 2.933@xmath0 m and the size of a gas - depleted region in the inner disk as probed by rovibrational co emission .
we discuss the properties of these molecular gaps , and propose their potential for studies of the evolution of planet - forming disk regions , in section [ sec : disc ] .
the terms disk gap " and hole " have been referred ( sometimes interchangeably ) to radial regions where the dust in the disk shows some degree of depletion as compared to a continuous distribution adopted by models for disk considered primordial " ( e.g. * ? ? ?
* ) . in this work ,
we adopt these terms to identify inner disk regions of various sizes ( namely @xmath280.120 au ) empirically defined to where molecular gas is depleted enough that its infrared emission ( at wavelengths of 2.935@xmath0 m ) is not detected .
the term gap " was utilized in @xcite with this meaning , with the discovery of a strong correlation between velocity - resolved line widths and the @xmath29 vibrational ratios in rovibrational co emission at 4.7@xmath0 m in a large survey of protoplanetary disks . by converting these measurements respectively into a characteristic co emitting radius r@xmath30 and the vibrational excitation temperature , this correlation revealed cooler co gas as larger gaps are formed in disks ( i.e. as r@xmath30 increases ) .
this observational finding suggested an inside - out depletion scenario , which is now being increasingly supported by independent modeling explorations @xcite . to distinguish them from dust gaps or holes detected by other techniques , we call themmolecular gaps / holes " , especially after the discovery in this work that they are shared by three of the most abundant molecules in disks : co , h@xmath14o , and oh ( sections [ sec : ana ] and [ sec : rad_trends ] ) . in absence of constraints on the radial distribution of any undetected residual co gas within the co - depleted inner region ,
we use the term gap / hole interchangeably . an upper limit to the co column density of @xmath31@xmath32 has been estimated by @xcite in one of these gaps / holes , implying that very little residual co gas is left ( if any ) in the inner co - depleted region .
l c c c c c c c c c c c c aatau & 0.85 & 0.71 & -8.31 & -1.43 & 71 & 0.14 & 0.20 & 1.45 & 140 & 0.322 @xmath33 0.011 & 0.337 @xmath33 0.006 & + as205 n & 1.10 & 7.10 & -7.10 & -0.18 & 20 & 0.38 & 0.17 & 5.45 & 125 & 4.390 @xmath33 0.370 & 5.220 @xmath33 0.400 & + as209 & 1.40 & 2.50 & -7.52 & -0.35 & 39 & 0.91 & 0.08 & 3.84 & 125 & 0.829 @xmath33 0.055 & 0.806 @xmath33 0.032 & + cvcha & 2.10 & 8.00 & -6.81 & 0.41 & 30 & 0.25 & 0.18 & 9.09 & 215 & 0.816 @xmath33 0.020 &
1.170 @xmath33 0.030 & 1 + cwtau & 1.20 & 0.76 & -7.80 & -0.49 & 28 & 0.32 & 0.32 & 2.74 & 140 & 1.470 @xmath33 0.180 & 1.760 @xmath33 0.160 & + dftau & 0.53 & 1.97 & -6.93 & -0.63 & 65 & 0.43 & 0.20 & 5.08 & 140 & 1.190 @xmath33 0.130 & 1.380 @xmath33 0.090 & + dgtau & 0.30 & 1.70 & -6.39 & 0.70 & 18 & 0.59 & 0.01 & 7.30 & 140 & 1.030 @xmath33 0.070 & 1.500 @xmath33 0.080 & 2 + doar24e s & 0.70 & 8.80 & -7.44 & -0.92 & 20 & 0.30 & 0.04 & 3.31 & 120 & 2.360 @xmath33 0.380 & 3.130 @xmath33 0.340 & + doar44 & 1.40 & 1.40 & -8.43 & -1.16 & 25 & 0.03 & 0.33 & 1.51 & 125 & 0.610 @xmath33 0.031 & 0.593 @xmath33 0.014 & + dotau & 0.72 & 1.05 & -7.28 & -0.67 & 31 & 0.49 & 0.12 & 3.93 & 140 & 0.892 @xmath33 0.080 & 1.150 @xmath33 0.060 & + drtau & 1.00 & 0.90 & -6.82 & -0.24 & 9 & 0.39 & 0.06 & 7.03 & 140 & 1.420 @xmath33 0.150 & 1.910 @xmath33 0.160 & + ec82 & 0.75 & 3.21 & & 0.03 & 59 & 0.03 & 0.81 & & 415 & 0.348 @xmath33 0.008 & 0.498 @xmath33 0.010 & 3 + exlup08 & 0.80 & 0.60 & -7.50 & 0.30 & 44 & 0.61 & 0.06 & 3.25 & 155 & 1.300 @xmath33 0.300 & 4.200 @xmath33 0.600 & 4 + exlup14 & 0.80 & 0.60 & -10.00 & -1.30 & 48 & 0.51 & 0.07 & 0.25 & 155 & 0.290 @xmath33 0.040 &
0.340 @xmath33 0.030 & 5 + fntau & 0.33 & 0.80 & -8.25 & -1.77 & 20 & @xmath340.12 & 1.69 & 1.12 & 140 & 0.294 @xmath33 0.009 & 0.308 @xmath33 0.005 & 6 + fztau & 0.70 & 0.51 & -7.33 & -0.27 & 38 & 0.20 & 0.98 & 3.70 & 140 & 1.060 @xmath33 0.070 & 1.340 @xmath33 0.070 & + gqlup & 0.80 & 0.80 & -8.15 & -1.10 & 65 & 0.28 & 0.20 & 1.67 & 150 & 1.090 @xmath33 0.090 & 1.030 @xmath33 0.050 & + hd36112 & 2.00 & 30.00 & -7.35 & -0.10 & 22 & 0.03 & 2.26 & 5.15 & 200 & 4.100 @xmath33 0.320 & 4.850 @xmath33 0.390 & + hd95881 & 2.00 & 2.98 & @xmath34 - 5.65 & @xmath341.63 & 55 & 0.06 & 4.65 & @xmath3429.31 & 170 & 4.830 @xmath33 0.900 & 6.990 @xmath33 0.970 & + hd97048 & 2.20 & 33.00 & @xmath34 - 8.16 & @xmath34 - 0.55 & 43 & 0.33 & 12.57 & @xmath342.32 & 180 & 3.010 @xmath33 0.200 & 3.250 @xmath33 0.130 & + hd98922 & 2.20 & 891.00 & @xmath34 - 6.97 & @xmath340.41 & 45 & 0.29 & 8.85 & @xmath347.84 & 350 & 16.500 @xmath33 2.400 & 30.700 @xmath33 0.400 & + hd101412 & 2.30 & 1.36 & @xmath34 - 7.61 & @xmath34 - 0.04 & 80 & & 0.79 & @xmath344.13 & 160 & 1.090 @xmath33 0.060 & 1.720 @xmath33 0.090 & + hd135344b & 1.60 & 8.00 & -8.35 & -1.11 & 14 & 0.05 & 1.70 & 1.72 & 140 & 3.170 @xmath33 0.390 & 3.540 @xmath33 0.400 & + hd139614 & 1.50 & 6.60 & -7.63 & -0.10 & 20 & 0.25 & 2.77 & 3.51 & 140 & 1.370 @xmath33 0.070 & 1.560 @xmath33 0.040 & + hd141569 & 1.90 & 19.10 & -7.65 & -0.05 & 53 & 0.56 & 17.02 & 3.72 & 100 & 0.652 @xmath33 0.026 & 0.441 @xmath33 0.008 & + hd142527 & 3.50 & 69.00 & -7.02 & 0.07 & 20 & 0.04 & 2.33 & 8.69 & 140 & 6.090 @xmath33 0.510 & 7.850 @xmath33 0.500 & + hd144432s & 1.70 & 14.80 & -7.69 & -0.35 & 27 & 0.09 & 0.78 & 3.44 & 160 & 2.200 @xmath33 0.360 & 2.530 @xmath33 0.260 & + hd150193 & 1.90 & 16.20 & -7.45 & 0.10 & 38 & 0.24 & 0.88 & 4.57 & 150 & 5.130 @xmath33 1.260 & 6.470 @xmath33 1.220 & + hd163296 & 2.30 & 36.00 & -7.13 & 0.28 & 46 & 0.14 & 1.40 & 6.75 & 122 & 9.810 @xmath33 2.630 & 11.600 @xmath33 3.600 & + hd179218 & 2.70 & 75.90 & -6.76 & 0.42 & 57 & 0.37 & 20.80 & 10.40 & 240 & 3.550 @xmath33 0.700 & 4.820 @xmath33 0.590 & + hd190073 & 2.85 & 83.20 & -5.82 & 1.38 & 23 & 0.27 & 6.03 & 27.72 & 767 & 3.950 @xmath33 0.790 & 5.620 @xmath33 1.170 & + hd244604 & 3.05 & 97.70 & -7.19 & 0.14 & 50 & 0.11 & 2.76 & 6.98 & 336 & 0.742 @xmath33 0.025 & 0.840 @xmath33 0.019 & + hd250550 & 3.40 & 190.00 & -5.63 & 1.82 & 10 & 0.16 & 0.91 & 35.70 & 280 & 1.690 @xmath33 0.210 & 2.030 @xmath33 0.160 & + htlup & 2.50 & 14.50 &
& -0.71 & 28 & 0.65 & 0.05 & & 150 & 1.500 @xmath33 0.170 & 1.590 @xmath33 0.110 & + imlup & 0.52 & 1.30 & -11.00 & -1.08 & 49 & 1.16 & 0.03 & 0.08 & 190 & 0.500 @xmath33 0.023 & 0.423 @xmath33 0.009 & + irs48 & 2.00 & 14.30 & -8.40 & -0.89 & 42 & 0.38 & 16.22 & 1.76 & 125 & 1.490 @xmath33 0.140 & 2.220 @xmath33 0.230 & + lkha330 & 2.50 & 16.00 & -7.66 & -0.46 & 12 & 0.09 & 4.74 & 4.04 & 250 & 1.130 @xmath33 0.110 & 1.360 @xmath33 0.090 & + rno90 & 1.50 & 5.70 & -7.40 & -0.19 & 37 & 0.21 & 0.23 & 4.44 & 125 & 2.070 @xmath33 0.130 & 2.460 @xmath33 0.160 & + rulup & 0.70 & 0.42 & -7.75 & -0.47 & 35 & 0.37 & 0.14 & 2.41 & 150 & 1.220 @xmath33 0.140 & 1.700 @xmath33 0.110 & + rylup & 1.50 & 2.60 & -7.50 & 0.08 & 68 & @xmath340.04 & 14.13 & 4.01 & 150 & 1.020 @xmath33 0.080 & 1.340 @xmath33
0.050 & 7 + rwaur & 1.34 & 1.70 & -7.50 & -0.20 & 55 & 0.62 & 0.05 & 3.86 & 140 & 1.000 @xmath33 0.040 & 1.380 @xmath33 0.040 & 8 + scra s & 0.60 & 0.76 & -7.53 & -0.65 & 26 & 0.40 & 0.04 & 2.87 & 130 & 2.710 @xmath33 0.310 & 3.040 @xmath33 0.220 & + scra n & 1.50 & 2.30 & -7.36 & -0.13 & 10 & 0.30 & 0.05 & 4.63 & 130 & 2.710 @xmath33 0.310 & 3.040 @xmath33 0.220 & + sr9 & 1.35 & 1.41 & -9.20 & & 60 & @xmath340.07 & 0.47 & 0.68 & 120 & 0.729 @xmath33 0.049 & 0.660 @xmath33 0.018 & 9 + sr21 & 2.20 & 15.00 & -7.90 & -0.70 & 15 & 0.64 & 6.46 & 3.03 & 125 & 1.110 @xmath33 0.090 & 1.080 @xmath33 0.040 & 10
+ ttau n & 2.10 & 7.30 & -7.16 & 0.05 & 20 & 0.57 & 0.07 & 6.35 & 147 & 8.800 @xmath33 0.800 & 12.000 @xmath33 0.800 & + ttau s & 2.50 & & & 0.55 & 25 & 0.20 & 1.10 & & 147 & 8.800 @xmath33 0.800 & 12.000 @xmath33 0.800 & + twcha & 1.00 & 0.38 & -8.13 & -1.66 & 67 & 0.16 & 0.71 & 1.84 & 180 & 0.178 @xmath33 0.004 &
0.174 @xmath33 0.003 & 6 + twhya & 0.70 & 0.23 & -8.70 & -1.30 & 4 & 0.09 & 0.34 & 0.91 & 54 & 0.482 @xmath33 0.023 & 0.303 @xmath33 0.006 & + vssg1 & 0.52 & 1.50 & -7.56 & -0.87 & 53 & 0.36 & 0.27 & 2.65 & 120 & 0.757 @xmath33 0.045 & 1.010 @xmath33 0.060 & + vvser & 3.00 & 85.00 & -7.13 & 0.33 & 65 & 0.16 & 2.89 & 7.38 & 415 & 2.600 @xmath33 0.480 & 2.530 @xmath33 0.180 & + vwcha & 0.60 & 3.37 & -7.50 & -0.77 & 44 & 0.32 & 0.16 & 2.95 & 180 & 1.060 @xmath33 0.050 & 0.847 @xmath33 0.047 & + vzcha & 0.80 & 0.46 & -7.91 & -0.73 & 25 & 0.44 & 0.18 & 2.14 & 180 & 0.354 @xmath33 0.009 & 0.397 @xmath33 0.007 & + waoph6 & 0.90 & 0.67 & -6.64 & -0.73 & 39 & 0.33 & 0.07 & 8.15 & 120 & 1.010 @xmath33 0.070 & 1.000 @xmath33 0.050 & + wxcha & 0.60 & 0.77 & -8.47 & & 87 & 0.45 & 0.11 & 1.09 & 180 & 0.399 @xmath33 0.011 & 0.392 @xmath33 0.007 & 11 +
co emission at 4.7@xmath0 m has been observed with the cryogenic infrared echelle spectrometer ( crires , * ? ? ?
* ) on the very large telescope ( vlt ) of the european southern observatory ( eso ) .
most spectra have been taken as part of the eso large program 179.c-0151 @xcite ; some measurements have been added from programs 079.c-0349 and 081.c-0833 as published in @xcite , from program 091.c-0671 as published in @xcite , and from program 088.c-0898 as published in @xcite .
crires has a resolving power of @xmath35 or 3.2kms@xmath4 , which provides detailed velocity - resolved line profiles of co gas from inner disks and allows us to radially locate the emission through keplerian line broadening ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . in total ,
we include in this work 52 disks that have velocity - resolved co lines from crires observations . within the eso large program 179.c-0151 ,
28 disks were observed at 2.9@xmath0 m to measure their h@xmath14o and oh emission spectra .
three of these disks were previously published in @xcite .
three main - sequence stars were observed to be used as photospheric standards , sampling stellar spectral types from k to m ( spectral types are taken from * ? ? ?
* ; * ? ? ?
* ) : nltt29176 ( k3 ) , hd111631 ( k7 ) , and hip49986 ( m1.5 ) .
these spectra are used to correct the 2.9@xmath0 m h@xmath14o and oh emission spectra , as explained in section [ sec : ana3 ] .
we also include a second epoch of the 2.9 and 4.7@xmath0 m spectra of a strongly variable accreting source , ex lupi , obtained as part of program 093.c-0432 @xcite , and measurements at 2.9@xmath0 m for hd250550 and hd98922 from program 082.c-0491 as published in @xcite .
the crires data from programs 179.c-0151 and 093.c-0432 have been reduced using procedures developed for the eso large program 179.c-0151 as described in @xcite . to complement the crires dataset , we include in this work the keck - nirspec ( r @xmath3 25,000 , * ? ? ?
* ) 4.7@xmath0 m co spectra of dotau , hd244604 , and hd36112 , previously published in @xcite and @xcite .
these targets have h@xmath14o spectra available at 2.9@xmath0 m and/or at 1035@xmath0 m , but have not been observed at 4.7@xmath0 m with crires .
the dataset of h@xmath14o and co emission at 2.9 and 4.7@xmath0 m is analyzed in this work in combination with _
spitzer_-irs spectra covering h@xmath14o emission at 1035@xmath0 m at the resolution of r @xmath3 700
. the _ spitzer _
dataset has been presented and analyzed in previous work @xcite .
the advantage of a combined analysis of water emission over such a large wavelength range is to include lines that probe upper level energies between 1500 and 9000 k and the whole disk region out to the water snow line ( see references in section [ sec : water_surveys ] ) .
the combined sample includes 55 disks around stars with estimated masses between 0.3 and 3.5 m@xmath23 ( table [ tab : sample ] ) , observed in different star - forming regions ( mainly taurus , ophiucus , lupus , chamaeleon for solar - mass stars , and several isolated intermediate - mass stars ) , and spanning evolutionary phases from primordial to transitional disks .
the sample has been assembled from previous work as driven by the availability of co spectra at 4.7@xmath0 m ( available for all 55 disks ) and h@xmath14o spectra at 2.9@xmath0 m ( available for 31 disks in this sample ) and/or at 1035@xmath0 m ( available for 44 disks in this sample ) .
c c c c c c c 2.9273 & @xmath36 & 49 & 8744 + 2.9278 & @xmath37 & 52 & 8388 + 2.9291 & @xmath38 & 48 & 8583 + 2.9292 & @xmath39 & 49 & 8698 + 12.396 & @xmath40 & 7.7 & 5781 + 12.407 & @xmath41 & 4.2 & 4945 + 12.454 & @xmath42 & 1.2 & 4213 + 12.519 & @xmath43 & 1.5 & 4133 + 12.596 & @xmath44 & 0.8 & 3966 + 17.103 & @xmath45 & 3.7 & 3274 + 17.225 & @xmath46 & 1.0 & 2439 + 17.358 & @xmath47 & 0.9 & 2432 + 30.525 & @xmath48 & 14 & 1750 + 30.871 & @xmath49 & 8.7 & 1806 + 32.991 & @xmath50 & 8.3 & 1525 + 33.005 & @xmath51 & 13 & 1504 + 63.323 & @xmath52 & 1.7 & 1070 + 2.9343 & 3/2 p4.5e & 11.7 & 5400 + 2.9346 & 3/2 p4.5f & 11.7 & 5400 + 12.647 & 1/2 r22.5ef & 213 & 15100 + 12.657 & 3/2 r23.5ef & 213 & 15100 + 30.277 & 3/2 r8.5f & 14.3 & 2380 + 30.346 & 3/2 r8.5e & 14.3 & 2380 + 4.7545 & v=1 - 0 p10 & 17.3 & 3300 + 4.7589 & v=2 - 1 p4 & 37.4 & 6160 m ( the spectra from the entire survey are shown in the appendix ) .
the spectrum is shown at the top in black , with its photospheric template in magenta ( veiled and broadened ) . the molecular emission spectrum after correction
is shown at the bottom in grey .
water and oh transitions are marked in blue and orange as in figure [ fig : phot stand].,scaledwidth=50.0% ]
in this section , we present the analysis of the h@xmath14o and oh emission spectra in the crires setting at 2.9052.99@xmath0 m . the analysis of these spectra presents two main challenges : one due to stellar photosphere absorption and one to the blending of emission lines , which can both be addressed by exploiting the high spectral resolution provided by crires .
the level of photospheric absorption at 2.9@xmath0 m depends on the spectral type of the star ( e.g. a cooler m1 star has more photospheric absorption from h@xmath14o and oh than a hotter k3 star , see figure [ fig : phot stand ] ) and on the amount of veiling from hot dust emission ( the ratio between continuum and photospheric flux @xmath53 ) .
we use photospheric templates to correct for photospheric absorption and retrieve the signal from disk gas emission ( section [ sec : ana3 ] ) .
most of the disk emission lines are blended due to their large velocity widths produced by keplerian rotation in the inner disk ( with full widths at half maximum fwhm up to 200 km / s ) .
we use the velocity - resolved co line profiles observed at 4.7@xmath0 m to characterize the h@xmath14o and oh emission at 2.9@xmath0 m ( section [ sec : ana4 ] ) .
the reason to use co emission at 4.7@xmath0 m is that it includes several lines spanning 30008000 k in upper level energy @xmath54 , which are close to and overlap with the upper level energies of h@xmath14o ( @xmath55 800010000 k ) and oh ( @xmath56 54005700 k ) at 2.9@xmath0 m ( table [ tab : lines ] ) .
these lines therefore may share similar excitation conditions and emit from a similar disk region , as supported by previous observations of similar line profiles for 4.7@xmath0 m co and 2.9@xmath0 m h@xmath14o and oh emission in a handful of disks @xcite .
l l c c c[h ! ] aatau & k7 & m1.5 & 2.2 & @xmath57 + cwtau & k3 & ( m1.5 ) & ( 14.0 ) & @xmath57 + dftau & m1 & m1.5 & 2.0 & + dotau & m0 & m1.5 & 7.0 & + exlup14 & m0 & m1.5 & 2.7 & + fntau & m5 & m1.5 & 0.24 & @xmath58 + gqlup & k7 & k7 & 4.8 & + htlup & k2 & ( k7 ) & ( 13.4 ) & @xmath57 + imlup & m0 & ( m1.5 ) & ( 1.7 ) & @xmath57 + rulup & k7 & m1.5 & 13.0 & + scra s & m0 & m1.5 & 13.3 & + twcha & k7 & m1.5 & 4.4 & + twhya & k7 & m1.5 & 0.8 & @xmath58 + vwcha & k5 & k7 & 5.4 & + vzcha & k6 & m1.5 & 6.8 & + waoph6 & k6 & k7 & 2.2 & + wxcha & m0 & m1.5 & 11.4 & we take each 2.9@xmath0 m spectrum that has substantial stellar photospheric absorption and we correct it by subtraction of the photospheric template that best matches its absorption spectrum . given the coarse grid of spectral type templates obtained in this survey ( figure [ fig : phot stand ] ) , the best matching template is usually easily identified for each science spectrum .
we apply the following procedure to the spectral range between 2.925 and 2.935@xmath0 m , where the s / n is higher , the spectrum is less affected by detector or telluric gaps , and the h@xmath14o and oh lines are more isolated and stronger ( figure [ fig : phot stand ] ) .
the template spectrum is first shifted to compensate for velocity differences with the science target , by cross - correlating the h@xmath14o and oh photospheric features in the science and template spectra .
to match the science spectrum , the template is then artificially veiled to account for circumstellar disk emission , and its absorption lines broadened .
this is achieved by adding an excess flat continuum and by convolving the template spectrum with a gaussian of variable width ( for the technique see also * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the best fit is found by minimizing the residuals after subtraction of the veiled and broadened template from the science spectrum , over a grid of veiling and broadening values .
residuals are measured as the standard deviation of pixel values from the median - smoothed de - veiled spectrum , in a spectral range free of disk molecular emission ( 2.9302.934@xmath0 m ) .
figure [ fig : phot_corr ] shows an example of the photospheric correction procedure as applied to the spectrum of vzcha ; plots for all the other targets are reported in the appendix .
table [ tab : phot_corr ] lists , for each target , the best photospheric standard match and the estimated veiling at 2.9@xmath0 m .
veiling measurements may be refined in future work by using a denser sampling of template spectral types .
the analysis of co lines has been presented in @xcite , and we provide here a short summary .
we built high s / n co line profiles for each disk by stacking rovibrational transitions observed in the 4.7@xmath0 m spectra . the stacked line profiles preserve the line flux as well as the line shape , through a homogeneous analysis that we adopted for the entire disk sample .
we stacked those lines that are most commonly available across the sample ( due to the observing strategies and the detector characteristics ) , that have similar flux , and that are not blended with higher - level transitions .
these requirements led us to typically use the @xmath59 lines between p7 and p12 , and @xmath60 lines between p3 and p6 .
each line was normalized to its local continuum , re - sampled on a common velocity grid , and stacked by weighted average .
the @xmath59 and @xmath60 stacked line profiles were then used to identify and separate multiple emission components , where present .
the @xmath60 lines , with higher @xmath54 ( table [ tab : lines ] ) , are mostly excited in the hotter innermost disk region , while the @xmath59 , with lower @xmath54 , often show additional flux contributions from colder gas at larger disk radii .
this has been empirically demonstrated through the systematic decomposition of co lines , where a broad ( inner ) component bc " is identified by the @xmath60 lines that match the wings of @xmath59 lines , and a narrow ( outer ) component nc " is obtained from the residuals by subtraction of the @xmath60 profile ( which typically includes only bc ) from the @xmath59 profile ( which includes both bc and nc ) .
an example of this procedure is included in figure [ fig : mol comp ] . in several disks ,
the two velocity components are clearly seen by visual inspection of the @xmath59 lines , which show a narrower peak on top of broader wings @xcite .
those disks that , following this analysis , have been found to have only one co component ( i.e. , those where the @xmath59 and @xmath60 stacked line profiles match ) , have been identified as single - component " disks ( sc " , e.g. sr21 in figure [ fig : mol comp ] ) . . .
m has two velocity components ( the broad component bc marked in red , the narrow component nc marked in cyan ) .
the spectrum of sr21 shows an example of a single - component disk " disk ( sc ) , where only a narrow co component is found .
the gallery of co line profiles is shown in @xcite .
_ center & bottom : _ examples of fits to h@xmath14o and oh line profiles at 2.93@xmath0 m by using the co emission components ( fits to other disks are shown in the appendix ) .
the spectrum is shown in black with grey error - bars , centered in velocity on ( for oh ) or in between ( for h@xmath14o ) the lines for visualization .
the best - fit line profiles obtained by combining co lines are shown in cyan ( using the nc ) , red ( using the bc ) , and in blue for the single - component disk .
the photospheric template , where used to remove the stellar photosphere , is shown at the top in magenta for reference.,title="fig:",scaledwidth=50.0% ] m has two velocity components ( the broad component bc marked in red , the narrow component nc marked in cyan ) .
the spectrum of sr21 shows an example of a single - component disk " disk ( sc ) , where only a narrow co component is found .
the gallery of co line profiles is shown in @xcite . _
center & bottom : _ examples of fits to h@xmath14o and oh line profiles at 2.93@xmath0 m by using the co emission components ( fits to other disks are shown in the appendix ) .
the spectrum is shown in black with grey error - bars , centered in velocity on ( for oh ) or in between ( for h@xmath14o ) the lines for visualization .
the best - fit line profiles obtained by combining co lines are shown in cyan ( using the nc ) , red ( using the bc ) , and in blue for the single - component disk .
the photospheric template , where used to remove the stellar photosphere , is shown at the top in magenta for reference.,title="fig:",scaledwidth=50.0% ] m has two velocity components ( the broad component bc marked in red , the narrow component nc marked in cyan ) .
the spectrum of sr21 shows an example of a
single - component disk " disk ( sc ) , where only a narrow co component is found .
the gallery of co line profiles is shown in @xcite . _
center & bottom : _ examples of fits to h@xmath14o and oh line profiles at 2.93@xmath0 m by using the co emission components ( fits to other disks are shown in the appendix ) .
the spectrum is shown in black with grey error - bars , centered in velocity on ( for oh ) or in between ( for h@xmath14o ) the lines for visualization .
the best - fit line profiles obtained by combining co lines are shown in cyan ( using the nc ) , red ( using the bc ) , and in blue for the single - component disk .
the photospheric template , where used to remove the stellar photosphere , is shown at the top in magenta for reference.,title="fig:",scaledwidth=50.0% ] l c c c c c broad ( bc ) & 50200 & 0.040.3 & 25/25 & 15/20 & 17/20 + narrow ( nc ) & 1070 & 0.23 & 23/25 & 3/20 & 3/20 + single ( sc ) & 670 & 0.320 & 28/28 & 1/10 & 4/10 o line at 12.396@xmath0 m as observed in drtau with vlt - visir at a resolution of @xmath61kms@xmath4 @xcite .
the rovibrational co line profiles of the broad and narrow components are shown in red and cyan , scaled to the visir line peak for comparison.,scaledwidth=40.0% ] o line fluxes in disks that have both crires and _ spitzer _ spectra available .
line fluxes from each disk are linked with a line , and are normalized to their 33@xmath0 m line flux ( the lowest - energy line that we include in the analysis , see table [ tab : lines ] ) .
upper limits are shown with red arrows .
the wavelength dependence of the decrease in the line fluxes is found both in the line - to - continuum strength , illustrated for three disks in figure [ fig : multiwl_lines ] , and in the integrated line fluxes , as shown here.,scaledwidth=50.0% ] we use the stacked co line profiles to reproduce the observed h@xmath14o and oh emission features , to test if their shapes and width are equal to those of co implying that they emit from a similar disk region .
we adopt the co line profile at each rest wavelength of h@xmath14o and oh transitions , combine them into a single feature if the lines are broad and overlap ( this typically happens for the broad component ) , and then vary the peak flux of the combined line profile to fit the 2.9@xmath0 m spectrum ( see figure [ fig : mol comp ] and the appendix ) .
h@xmath14o and oh lines at 2.9@xmath0 m are typically detected in double - component disks only ( i.e. when the bc is not detected in co , h@xmath14o and oh at 2.9@xmath0 m are typically not detected either ) .
where detected , h@xmath14o and oh lines at 2.9@xmath0 m are dominated by a broad component that matches the co bc profile ( figure [ fig : mol comp ] and table [ tab : molec summary ] ) ; we therefore conclude that these lines are emitted at similar disk radii .
this was first recognized in the 35@xmath0 m spectrum of ex lupi by @xcite , and applies to the majority of disks in the sample studied here ( see table [ tab : molec summary ] ) .
disks that lack the co broad component ( the single - component " disks ) do not show h@xmath14o emission at 2.9@xmath0 m either .
this is a first hint that disks with evidence for an inner co - depleted region are also depleted in water , which will be further detailed and discussed in the next sections . in these disks ,
oh is detected in a few cases .
this finding agrees with previous work that detected oh but not h@xmath14o in some disks around intermediate - mass stars @xcite ; disks in this stellar mass bin are in fact found to typically have only a single , narrow co component @xcite .
while h@xmath14o and oh emission at 2.9@xmath0 m comes from similar disk radii as the co broad component ( @xmath62au ) , a survey of velocity - resolved h@xmath14o and oh emission at @xmath632.9@xmath0 m is still missing , and is essential to determine what disk radii are probed at longer wavelengths .
h@xmath14o emission at 12.4@xmath0 m may come from a region at larger disk radii that matches the emitting region of the narrow co component , as shown in figure [ fig : drtau_visircomp ] for one disk .
in fact , the 12.4@xmath0 m h@xmath14o lines have lower energy than the 2.93@xmath0 m lines ( table [ tab : lines ] ) , and may therefore match the colder / outer component of co emission rather than the hotter / inner one .
by combining h@xmath14o and oh spectra between 2.9 and 35@xmath0 m , we find a wavelength dependence in the decrease of line fluxes .
when the emission is detected at 2.9@xmath0 m , it is always detected at longer wavelengths up to 33@xmath0 m .
on the contrary , when lines at 3033@xmath0 m are detected , those at 12 or 2.9@xmath0 m may not necessarily be ( see figures [ fig : multiwl_lines ] and [ fig : multiwl_lines2 ] , and table [ tab : fluxes ] ) .
the decrease of water emission is found as a sequential intrinsic decrease of the observed line fluxes and of the line - to - continuum strength , where water line fluxes at short wavelengths decrease when water line fluxes at longer wavelengths are still strong . since the emission probes hotter to colder gas from short to long wavelengths ( following a global trend in @xmath64 energies that decrease with increasing wavelength
see table [ tab : lines ] , figure [ fig : multiwl_lines ] , and * ? ? ?
* ) , this wavelength dependence of the line flux decrease may reflect the physical removal of molecular gas in the disk from the inside - out ( as found in the disk of twhya , * ? ? ? * ) . in section [ sec : ana ]
, we reported that h@xmath14o and oh emission at 2.9@xmath0 m comes from a disk region that is consistent with that of the broad co emission at 4.7@xmath0 m , so that we can adopt r@xmath30 as descriptive of h@xmath14o and oh emission too . the fact that the high - energy h@xmath14o lines at 2.9@xmath0 m disappear when the broad component of co emission disappears from the observed spectra , suggests that the depletion of the hotter / inner co gas and the depletion of the hotter / inner h@xmath14o gas proceed together .
we now combine measurements from @xcite and this work to test the dependence of h@xmath14o and oh line fluxes on r@xmath30 as a tracer of molecular gaps in inner disks .
r@xmath30 measures the location of co emission in the disk by taking into account the stellar mass and disk inclination , which are both linked to the measured line fwhm through kepler s law , as @xmath65 where m@xmath24 is the stellar mass , @xmath66 the disk inclination , g the gravitational constant , and we take the line velocity at the half width at half maximum ( fwhm/2 ) as in @xcite . as a reference comparison for r@xmath30
, we take the midplane water snow line radius r@xmath67 , defined as : @xmath68 where m@xmath69 the accretion rate , and @xmath70 the water freeze - out temperature .
equation [ eqn : r_snow ] is the midplane snow line radius for a viscously heated disk as derived by @xcite viscosity parameter of 0.01 , and a rosseland mean dust opacity of 770 cm@xmath58 g@xmath4 that normalize the extra terms in eqn .
[ eqn : r_snow ] as formulated in @xcite .
] , who used @xmath72 k as in @xcite .
this formula assumes that the disk midplane heating by stellar irradiation is negligible , and this has been shown to be valid for disks around stars younger than @xmath73myr with mass up to @xmath74m@xmath23 and accretion rates between @xmath75 and @xmath76m@xmath23yr@xmath4 @xcite .
some disks in our sample exceed these limits in stellar mass and age , while some have inner dust gaps that may change the disk temperature structure . for simplicity
we adopt the same equation for the entire sample . by normalizing to r@xmath67 , we put r@xmath30 from individual disks in reference to a disk radius of equal temperature across the entire disk sample , as set by water ice evaporation in the disk midplane
. r@xmath30/r@xmath67 therefore enables a comparison of results from individual disks within a common reference framework .
figure [ fig : r_fl_corr ] shows that line fluxes of h@xmath14o , oh , and co decrease with increasing r@xmath30/r@xmath67 , i.e. as the inner gap radius increases and reaches the snow line radius times larger than those formulated by @xcite at stellar masses between 1 and 3m@xmath23 ; while the r@xmath30/r@xmath67 values slightly shift depending on the formula used for the snow line radius , the global trends presented here are preserved . ] .
h@xmath14o line fluxes at 2.9@xmath0 m show a decreasing linear correlation at radii @xmath77 logr@xmath30/r@xmath78 , beyond which they fall under the detection limit ( @xmath79 ergs@xmath4@xmath32 distance/140pc ) .
a linear fit to the detected lines provides ( using the bayesian method by * ? ? ?
* ) : @xmath80 where flux@xmath81 is the measured line flux normalized to a distance of 140 pc . at 12.5 and 33@xmath0 m ,
h@xmath14o lines are detected out to logr@xmath30/r@xmath82 , and decrease only at logr@xmath30/r@xmath83 .
oh emission follows similar trends , but oh lines are tentatively detected up to larger disk radii ( especially the cold lines at 30@xmath0 m ) .
variability in accretion luminosity may contribute to the order - of - magnitude scatter in line fluxes along these trends ; however , a factor up to @xmath84 change in line fluxes has been observed so far only in the extreme accretion outburst of ex lupi , and not during the typically lower variability of other young stars included in this sample @xcite .
similarly to h@xmath14o and oh , the co @xmath59 line flux decreases with r@xmath30/r@xmath67 , but it is detected in disks well beyond the snow line radius .
we interpret this as evidence that what regulates the observed h@xmath14o and oh emission is the molecular gas left within the snow line radius , or in other words that h@xmath14o and oh emission disappears as inner disks are depleted from the inside out to the snow line ( i.e. as r@xmath30 grows larger and approaches r@xmath67 ) , beyond which the reservoir of water vapor is strongly limited due to freeze - out at low temperatures .
co gas is instead still detected beyond the snow line because it freezes out at lower temperatures ( @xmath12k ) , found at larger disk radii . as a basic test of this inside - out depletion scenario , we use a simple disk model composed of concentric rings of increasing radius , where each ring is a slab of gas described by one temperature and one column density as in @xcite .
we adopt power - law profiles to set how the temperature and column density of each ring decrease with increasing disk radius , as @xmath85(r ) = @xmath86 ( r / r@xmath87)@xmath88 and @xmath89(r ) = @xmath90 ( r / r@xmath87)@xmath91 .
we calibrate the model using the observed co line fluxes , by adopting model parameter values from the radial temperature profile measured by @xcite in rovibrational co emission , as @xmath92k , @xmath93 , @xmath94@xmath32 , and @xmath95 . in figure
[ fig : r_fl_corr ] ( panel g ) we show how the integrated co line flux from the model decreases with increasing gap size , by removing one ring at a time from r@xmath96au to r@xmath97au to mimic an inside - out removal of hotter to colder gas ( the radial grid of disk rings is shown by the dots ) .
a radially decreasing power - law temperature profile reproduces the decrease in the co line fluxes as a curve , where as the inner hole radius increases , colder gas emission beyond the hole produces weaker co lines at 4.7@xmath0 m .
co line fluxes that exceed the model at logr@xmath30/r@xmath98 come from disks where the emission is pumped by uv radiation @xcite .
the same ring slab model is then used for h@xmath14o emission .
such a model produces line fluxes stronger than observed at 12.5 and 33@xmath0 m .
we therefore include a step function for the column density , to simulate the presence of a radius in the surface of the disk beyond which water vapor is reduced by freeze - out or by chemistry in cold regions ( e.g. * ? ? ?
we keep @xmath95 as used for co at disk radii smaller than @xmath99au , and adopt @xmath100 to produce a steeper decrease in water column density at larger disk radii ( where the gas temperature decreases to @xmath101k ) .
this model , shown in panels a , b , and c of figure [ fig : r_fl_corr ] , approximately reproduces both the absolute fluxes and the wavelength - dependent radial decrease in water emission at 2.9333@xmath0 m .
a step function in the column density produces a decrease / cutoff in the 12.5 and 33@xmath0 m line fluxes at logr@xmath30/r@xmath102 , which is consistent with the observed water line fluxes in the _ spitzer _ spectra .
the purpose of the ring slab model is to test in the simplest terms if an inside - out removal of disk gas can explain the observed line flux decrease within the snow line .
clearly , this simple model ignores several important factors that should be considered , such as a realistic physical disk structure , disk chemistry , and their adaptation to the formation and development of disk gaps .
however , the model seems to have the key ingredients needed to qualitatively reproduce the observed data , by including a radially decreasing temperature and the inside - out removal of molecular gas .
detailed modeling of the water spectrum and of its dependence on disk regions of water vapor depletion at and beyond the snow line has been presented by @xcite .
recent explorations by @xcite for a disk around an intermediate - mass star found a wavelength - dependent decrease in the _ spitzer _ water line fluxes for a grid of increasing disk hole sizes , in qualitative agreement with what we measure in the data out to the snow line radius ( figure [ fig : r_fl_corr ] , panels b and c ) .
these models produce an excess flux when the gap reaches the snow line radius , contrarily to the observations ( see section [ sec : disc_ratios ] ) .
the wavelength - dependent decrease in water vapor emission between 2.9 and 33@xmath0 m is summarized in figure [ fig : tr_water ] , where we show the temperature - radius ( t - r ) diagram originally defined for co rovibrational emission by @xcite and we color - code the individual disks by their h@xmath14o detections .
for r@xmath30/r@xmath67 between 0.1 and 1 , all disks have lost their detectable h@xmath14o emission at 2.9@xmath0 m ( their line fluxes are @xmath103 ergs@xmath4@xmath32 distance/140pc ) , while some still retain detectable emission at 12 or 33@xmath0 m . beyond the snow line ( r@xmath30/r@xmath104 ) , h@xmath14o emission between 2.9 and 35@xmath0 m is not detected in any disk . as a diagnostic of r@xmath30/r@xmath67 , we illustrate in figure [ fig : tr_water ] the wavelength - dependent h@xmath14o emission decrease shown earlier in figure [ fig : multiwl_lines ] , here in reference to the increasing disk gap size .
figure [ fig : r_fl_corr ] ( panels h and i ) shows molecular line flux ratios as measured in the crires dataset ( 2.94.7@xmath0 m ) .
we find a constant oh / h@xmath14o ratio of @xmath105 in the range @xmath77 logr@xmath30/r@xmath78 .
as for the co / h@xmath14o ratio , we find an increasing trend with increasing radius in the range @xmath77 logr@xmath30/r@xmath106 .
this trend is due to a decreasing h@xmath14o line flux , while the co line flux is less steeply decreasing over these disk radii ( figure [ fig : r_fl_corr ] , panels a and g ) . by fitting the measured line ratios
, we find the following linear relation ( using * ? ? ?
* ) : @xmath107 these trends are further analyzed and discussed in section [ sec : disc_ratios ] .
water emission from protoplanetary disks has been detected by the _
spitzer_-irs in over 60 disks ( see section [ sec : water_surveys ] ) .
this sample revealed two major trends : 1 ) water line fluxes increase with increasing stellar luminosity ( and mass , see figures [ fig : spitzer_spt ] and [ fig : lum_trends ] , and * ? ? ? * ) , and , in contrast with the first trend , 2 ) the water vapor frequency is much higher in disks around stars of @xmath108 m@xmath109m@xmath23 ( with a frequency of 6080% ) than in disks around stars of m@xmath26m@xmath23 ( with a frequency of @xmath110% ; see table [ tab : water_surveys ] , figure [ fig : spitzer_det_rates ] , and * ? ? ? * ) .
recent chemical modeling of the molecular content of inner disks by @xcite , while able to reproduce the luminosity trend , was still unable to reproduce the absence of water emission in disks around intermediate - mass stars .
we have demonstrated above that the strength of the infrared water spectrum is linked to an inside - out depletion scenario in disks , by finding : 1 ) a similar emitting region and detection frequency for the high - energy h@xmath14o lines at 2.9@xmath0 m and the innermost co gas ( the broad component ) , 2 ) a wavelength - dependent decrease of water line fluxes , where hotter emission ( as probed at 2.9@xmath0 m ) decreases before colder emission ( as probed at 33@xmath0 m ) , and 3 ) a correlation between the decrease of water line fluxes at 2.933@xmath0 m and the increase in size of an inner molecular hole measured from co emission .
one major implication of this analysis is that the low frequency of water vapor detections in disks around m@xmath26m@xmath23 is found to be linked to inner holes that have depleted the molecular content of their disks out to close to or beyond the snow line , contrary to most disks around solar - mass stars .
other independent techniques find signatures of dust inner holes in a growing number of disks around intermediate - mass stars ( e.g. * ? ? ?
* ; * ? ? ?
* ) . with the analysis of disks that have co emission observed at 4.7@xmath0 m and h@xmath14o at 2.9@xmath0 m and/or 1033@xmath0 m ( table [ tab : fluxes ] )
, we also find that the depletion scenario is not limited to the known transitional disks or to disks around intermediate - mass .
as shown in figure [ fig : tr_water ] , several disks around solar - mass stars show at least some degree of water - depletion ( those that we find to have @xmath111r@xmath30/r@xmath112 ) . by interpreting results from _ spitzer _ and vlt - crires surveys in the context of this analysis ( figure [ fig : spitzer_det_rates ] ) , we could conclude that @xmath282040% of disks around solar - mass stars may have molecular gaps in inner disks , while this fraction may be @xmath2880% for disks around intermediate - stars .
the analysis of velocity - resolved molecular line spectra suggests a path of disk evolution where gas - phase co and water are depleted from the planet - forming region in an inside - out fashion , a path that may be shared by disks around stellar masses of @xmath113m@xmath24/m@xmath23@xmath114 ( as probed in this sample ) . by detecting h@xmath14o emission only at long wavelengths in sr9 , imlup , hd35929 , ec82 , twhya , doar44 and hd163296 ( the last three of this list
have been previously studied by * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , we propose that these disks may be in an intermediate phase of depletion ( @xmath111r@xmath30/r@xmath112 ) and still retain some water vapor in their planet - forming region .
disks that have 2.933@xmath0 m water line fluxes lower than @xmath115ergs@xmath4@xmath32 ( distance/140pc ) , have r@xmath30/r@xmath116 and have already depleted their molecular gas in the planet - forming region within the snow line .
this line flux limit may be used to identify molecular holes / gaps in inner disks when only the line flux , and not the line profile or a direct image , can be obtained . in figure [ fig : tr_water ] we include for comparison also the case of disks around low - mass stars ( m@xmath117 m@xmath23 ) , observed with spitzer at 1019@xmath0 m @xcite , divided by 10 to re - scale the strong emission observed from organic molecules ( hcn and c@xmath1h@xmath1 ) .
rovibrational co emission at 4.7@xmath0 m has not been obtained for any of these disks so far , so we can not include them in the t - r diagram . however , their spectra at _ spitzer _ wavelengths could be distinguished from that of disks with large gaps by considering the emission from organics .
@xcite suggested that disks around low - mass stars may be too cold to produce strong water emission .
the luminosity dependence shown in figure [ fig : spitzer_spt ] may provide an explanation for the non - detection of water in these disks with _ spitzer _
( only a tentative detection was reported by * ? ? ?
* ) : models produce fluxes at or below the measured sensitivity limits , due to a radial extent of the warm water - emitting layer that is @xmath9 times smaller than in a disk around a solar - mass star ( @xmath11 au , see * ? ? ?
however , according to models , these disks would still show prominent organic emission lines from a disk atmosphere at 0.110au only mildly uv - irradiated as compared to higher - mass stars .
therefore , a continuous disk around low - mass stars would produce emission at least from organics at 1315@xmath0 m , while no molecular emission at 1019@xmath0 m may be the signature of inner disk depletion with r@xmath30/r@xmath118 also in disks around low - mass stars .
previous studies proposed the idea that uv radiation may be responsible for the depletion of water vapor in inner disk surfaces , through an efficient photo - dissociation of h@xmath14o into oh @xcite .
current thermochemical disk models , however , show that the balance at thermochemical equilibrium between the production and destruction of the water molecules produces large columns of water vapor even in highly - uv - irradiated disk surfaces ( e.g. * ? ? ?
in fact , models produce more water emission than in less uv - luminous stars ( or less massive stars , see figure [ fig : spitzer_spt ] ) , because h@xmath14o and oh are able to self - shield to uv radiation at the column densities typically measured in inner disks ( @xmath119@xmath32 , * ? ? ? * ) .
therefore , it is still unclear whether uv photo - dissociation alone may be responsible for a global depletion of water vapor in inner disks .
further , co and h@xmath14o emission at 2.94.7@xmath0 m recede at larger disk radii together ( section [ sec : ana ] ) , and the co molecule is able to self - shield to uv photo - dissociation at even lower columns ( @xmath120@xmath32 , * ? ? ?
* ) . as a consequence , a depletion of the molecular column densities to below the self - shielding limits
may be required , before uv photo - dissociation may deplete any residual co and h@xmath14o gas .
the destruction of h@xmath14o by uv radiation has been observationally tested by measuring if oh emission is consistent with prompt emission from excited oh , produced by photo - dissociation of h@xmath14o @xcite . however ,
h@xmath14o line fluxes increase ( instead of decreasing ) with increasing uv radiation , by considering the accretion luminosity as a proxy ( at least up to moderate accretion luminosities ; see * ? ? ?
* ; * ? ? ?
* and figure [ fig : accretion_figure ] )
. the measured @xmath29 vibrational ratios of co lines , which are sensitive to uv pumping and to the uv irradiation that reaches the disk ( e.g. * ? ? ?
* ) , provide another way to test if and how h@xmath14o line fluxes are linked to uv irradiation .
the ratio @xmath29 decreases as water emission decreases from shorter to longer wavelengths , and @xmath29 is not anomalously high in several disks around intermediate - mass stars that show no water emission ( figure [ fig : tr_water ] ) .
however , the increase in @xmath29 with inner gap size for r@xmath121au demonstrates that enough uv radiation is present to pump the co at large radial distances . in this regime of disk gaps around intermediate - mass stars ,
uv photo - dissociation of water could contribute more significantly to the depletion of residual water vapor .
we find that this regime is activated after the inner disk region at @xmath122au has already been depleted from co and water .
if not through uv photo - dissociation , the depletion of co and h@xmath14o may be initiated in inner disks by direct removal of disk gas .
disk photo - evaporation by x - ray and uv photons potentially provides an inside - out dispersal mechanism , although this is proposed to be efficient only beyond the radius where the gas is gravitationally bound to the star ( @xmath12312au in disks around solar - mass stars ) .
another removal mechanism could be through portions of mhd winds ( e.g. * ? ? ?
* ; * ? ? ?
* ) that may be probed at optical wavelengths through a low - velocity component in the forbidden oxygen lines @xcite .
a slow disk wind has been proposed to contribute also to the co narrow component in double - component disks @xcite , and may play a role in the depletion of molecular gas in inner disks . a promising direction for future work is to test if wind models , potentially combined to grain growth and planetesimal formation , can reproduce the formation and development of the molecular holes revealed by the analysis of co and h@xmath14o velocity - resolved infrared spectroscopy .
in figure [ fig : r_fl_corr ] , we showed linear trends in the co / h@xmath14o and oh / h@xmath14o line flux ratios , as measured in the 2.94.7@xmath0 m lines observed with crires .
both ratios were characterized only out to log r@xmath30/r@xmath124 , due to the absence of 2.93@xmath0 m h@xmath14o detections at larger radii .
to explore these trends over a larger disk region , we include in figure [ fig : mol_ratios ] the line flux ratios as measured using h@xmath14o line detections at longer wavelengths from the _ spitzer _ dataset , which is sensitive to emission at larger disk radii ( sections [ sec : ana ] and [ sec : rad_trends ] ) .
we find that the co / h@xmath14o line flux ratio of all three water lines show similar radial trends .
the increase in co / h@xmath14o is confirmed out to log r@xmath30/r@xmath125 ( figure [ fig : mol_ratios ] , left plot ) by all three water lines . at larger radii , the _ spitzer _
lines suggest a decreasing co / h@xmath14o out to the snow line radius . beyond the snow line ,
lower limits show that the ratio may increase again . following a similar procedure
, we show the measured oh / h@xmath14o line flux ratios in figure [ fig : mol_ratios ] , right plot . the constant @xmath105 ratio with disk radius
is confirmed also in this case by all three oh / h@xmath14o ratios as measured at different wavelengths ( 2.93@xmath0 m using the crires data , 12.5 - 12.6@xmath0 m and 30 - 33@xmath0 m using the _ spitzer _ data ) .
some disks show a higher oh / h@xmath14o ratio of @xmath126 at 12.533@xmath0 m , perhaps linked to phases of stronger production of oh through increased uv irradiation ( among these disks are in fact exlupi and dgtau , where enhanced production of oh through uv photodissociation has been found by * ? ? ?
* ; * ? ? ?
. a decreasing trend of oh / h@xmath14o may be present at log r@xmath30/r@xmath127 . beyond the snow line ,
lower limits suggest that the ratio may increase .
models propose that the oh / h@xmath14o ratio in the observable column density of gas should increase significantly at the snow line , due to a drop in the column density of h@xmath14o ( e.g. * ? ? ?
future velocity - resolved observations must test if the 12.533@xmath0 m water lines have velocity profiles that match the 12.630@xmath0 m oh lines , indicating a similar emitting region ( so far , no oh lines have been velocity - resolved at wavelengths @xmath27@xmath0 m ) .
an open question is if these observed trends are due to the disk physical / chemical radial structure in a static primordial " disk , rather than by the formation and evolution of disk gaps that may change it .
the decrease of co / h@xmath14o and oh / h@xmath14o at log r@xmath30/r@xmath127 out to the snow line radius may be due to changes in the disk chemistry as molecular holes expand out to the snow line radius .
a promising direction for future modeling is to calibrate the disk chemistry with the measured molecular line flux ratios .
the explorations suggested above within the framework of models that include disk dispersal processes may provide important constraints on the evolution of planet - forming regions .
interesting in this context is , for instance , the notable discrepancy between the data and the model shown in figure [ fig : r_fl_corr ] ( panel b and c ) , where @xcite predict an increase in water emission when disk gaps reach the snow line radius , as due to a new reservoir of water emission released by photo - desorption of icy grains ( i. kamp , private communication ) .
the fact that this is not observed in the data may suggest that the desorption efficiency is over - estimated , either by the adopted desorption rates or because dust grains are larger ( i.e. the total surface available for ice desorption smaller ) or are removed from disk surfaces . in the evolved disk of twhya , to reproduce weak water lines observed at @xmath128@xmath0 m with _ herschel _ , @xcite proposed that icy dust grains may have been removed from the disk surface by settling towards the midplane , hiding a large reservoir of icy solids in a region shielded from uv irradiation and photo - desorption .
\1 ) this analysis combines a sample of 55 protoplanetary disks that have co rovibrational emission spectra observed at 4.7@xmath0 m as well as h@xmath14o and oh emission spectra observed at 2.9 and/or 1033@xmath0 m .
h@xmath14o and oh spectra at 2.9@xmath0 m ( from vlt - crires ) are published here for the first time for 25 disks ; co 4.7@xmath0 m spectra ( from vlt - crires ) and h@xmath14o and oh spectra at 1033@xmath0 m ( from _ spitzer_-irs ) have been published previously .
this combined sample probes the molecular content of inner disks at radii of @xmath280.0520au .
\2 ) the velocity - resolved line profiles of h@xmath14o and oh emission at 2.9@xmath0 m are matched by the broader / inner component of 4.7@xmath0 m co emission as identified by @xcite .
these emission lines probe a similar inner disk region at @xmath280.040.3au , as shown by their similar profiles , and when co is depleted in this region , h@xmath14o is depleted too .
h@xmath14o lines at longer wavelengths ( 1233@xmath0 m ) , probing progressively colder water , may be co - spatial with the colder component of co emission at disk radii of @xmath129au .
\3 ) we propose the infrared water spectrum as a tracer of inside - out clearing of inner disks around stars of masses between 0.3 and 3.5 m@xmath23 .
we find a wavelength - dependent decrease of water line fluxes where water emission is absent at 2.9@xmath0 m but still present at longer wavelengths ( 1233@xmath0 m ) in some disks .
we discover a correlation between water line fluxes at 2.933@xmath0 m and the size of inner disk gaps / holes as probed by co emission , where water emission decreases as @xmath130 approaches the snow line radius .
all taken together , we interpret this as due to the progressive depletion of inner disks in an inside - out fashion , where inner / hotter gas is depleted before outer / colder gas .
\4 ) this analysis helps clarifying the origin of the absence of infrared h@xmath14o emission in disks around intermediate - mass stars : we find that these disks have inner gaps or holes in molecular gas ( i.e. they have large r@xmath6 ) .
this adds to growing evidence from other studies of gaps and holes in dust emission ( e.g. * ? ? ?
* ; * ? ? ?
* ) , as well as to recent model explorations @xcite .
\5 ) we measure the radial profiles of co / h@xmath14o and oh / h@xmath14o line flux ratios in inner disks out to the water snow line .
we find an increasing co / h@xmath14o with disk radius within the snow line , possibly followed by an inversion in the trend with a co / h@xmath14o ratio decreasing with disk radius as the molecular hole expands out the snow line radius .
for oh / h@xmath14o , we find a constant ratio within the snow line , again possibly followed by an inversion in the trend . to conclude
, our work finds that the decrease in co and h@xmath14o line fluxes is linked to the depletion of molecular gas through inside - out formation of inner disk gaps or holes .
line flux measurements and trends from this work provide an opportunity to model the properties of evolving inner disks , to investigate how physical changes ( e.g. in the gas - to - dust ratio , dust grain growth , and the removal of inner disk gas ) may affect the composition of planet - forming disk regions ( e.g. through an enhanced uv penetration at large disk radii , and the depletion of molecular gas from inner disks ) .
measured radial trends in the chemistry and depletion of water during the evolution of protoplanetary disks may offer new ground to investigate the origins of radial gradients in composition measured in solar system bodies ( e.g. * ? ? ?
* ) and potentially also in exoplanet populations
. the _ james webb space telescope _ ( _ jwst _ ) will provide r @xmath131 ( or @xmath132km / s ) spectroscopy of molecular spectra from the 2.9@xmath0 m lines included in this analysis ( with _ jwst_-nirspec ) to the longer wavelengths covered by _ spitzer _
( up to 28@xmath0 m , with _
jwst_-miri ) .
the observed molecular emission spectra will include co , h@xmath14o , oh , and co@xmath14 .
the broad spectral coverage will provide the data to study more molecular trends than those included in this work , and to better characterize the chemistry and its evolution when gaps form in disks .
_ jwst _ has also the potential to study infrared molecular emission in disks around low - mass stars , and disk samples not limited to the closest star - forming regions at 120 - 180pc .
a _ jwst _ spectrum at 328@xmath0 m may distinguish the evolutionary phase of inner disks from the measured line fluxes and the molecular species detected ( figure [ fig : tr_water ] ) , spanning a range of disk radii and gap sizes ( @xmath30.0520au , * ? ? ?
* ) that is highly complementary to the larger scales probed by millimeter interferometry with the atacama large millimeter array alma ( @xmath133au at 140pc , e.g. * ? ? ?
* ) and by direct imaging techniques .
the joint use of these observatories can transform our understanding of the development of disk gaps from the innermost disk regions to the larger - scale disk structures , and of the physical and chemical evolution of planet - forming environments .
+ we thank s.antonellini and i.kamp for providing water emission spectra from their prodimo model explorations , and for insightful discussions on water emission in disks and on its dependence on disk model parameters .
we also thank a.carmona for providing the crires spectrum of hd139614 .
a.b . wishes to thank all colleagues who made this work possible through their pioneering work in infrared spectroscopy of disks .
thanks simon bruderer for the seminal work he did in 2011 - 2012 , which provided knowledge and tools to enable this work .
acknowledges financial support by a nasa origins of the solar system grant no .
oss 11-oss11 - 0120 , a nasa planetary geology and geophysics program under grant nag 5 - 10201 .
e.v.d . acknowledges european union a - erc grant 291141 chemplan .
this work is partly based on observations made with eso telescopes at the paranal observatory under programs 179.c-0151 , 093.c-0432 , 079.c-0349 , 081.c-0833 , 091.c-0671 , 082.c-0491 , and 088.c-0898 .
this work is partly based on observations made with the spitzer space telescope , which is operated by the jet propulsion laboratory , california institute of technology .
this publication makes use of data products from the wide - field infrared survey explorer , which is a joint project of the university of california , los angeles , and the jet propulsion laboratory / california institute of technology , funded by the national aeronautics and space administration .
this research has made use of the vizier catalogue access tool , cds , strasbourg , france .
the original description of the vizier service was published in a&as 143 , 23 .
this work is partly based on observations obtained at the w. m. keck observatory , which is operated as a scientific partnership among the california institute of technology , the university of california and the national aeronautics and space administration .
the observatory was made possible by the generous financial support of the w. m. keck foundation .
we measure emission line fluxes by fitting a first - order polynomial continuum over neighboring pixels and by summing continuum - subtracted pixel flux values over the spectral range of each line . in the _
spitzer_-irs spectra , to select pixels of continuum free from line emission , we use models of the strongest emitters at 1035@xmath0 m as identified in previous work ( water , oh , hcn , c@xmath1h@xmath1 , co@xmath1 , hi ; see e.g. * ? ? ?
* ; * ? ? ?
the following are the ranges adopted : the 12.52@xmath0 m line is measured on 12.5012.537@xmath0 m , and its continuum fitted on 12.4812.5@xmath0 m and 12.712.72@xmath0 m ( at the two sides of the line ) ; the 33@xmath0 m line on 32.9233.06@xmath0 m and its continuum on 32.5332.63@xmath0 m and 33.433.47@xmath0 m ; the 12.65@xmath0 m oh line on 12.6412.68@xmath0 m and its continuum on 12.4812.5@xmath0 m and 12.712.72@xmath0 m ; the 30@xmath0 m oh lines on 30.24530.385@xmath0 m and 30.6230.73@xmath0 m and their continua on 30.130.17@xmath0 m and 30.3830.43@xmath0 m , and 30.5730.62@xmath0 m and 30.7330.83@xmath0 m ( we sum the flux of these oh doublets to increase the s / n ) .
oh detections at 30@xmath0 m are only tentative in some spectra of disks around intermediate - mass stars , and will need to be confirmed by higher - spectral - resolution observations .
line flux errors are estimated as the standard deviation of the distribution of measured line fluxes after re - sampling the spectrum using the local noise , as in @xcite . following this procedure , we find values consistent within 1 @xmath134 ( or a few % ) to those measured by @xcite , adopting as a test case the 17.22@xmath0 m water line .
line fluxes and errors are reported in table [ tab : fluxes ] ; in all figures throughout this paper , line fluxes are normalized to a common distance of 140pc for a homogeneous comparison , using distances from table [ tab : sample ] . in the crires spectra ,
line fluxes are measured in the continuum - normalized spectra and then flux calibrated using fluxes measured by the wide - field infrared survey explorer ( wise , * ? ? ?
* ) at 3.35 and 4.6@xmath0 m as released by @xcite . in some cases where the wise 4.6@xmath0 m measurement disagrees with previous photometry
, we adopt the _
spitzer_-irac measurement or an interpolated value between wise measurements at 3.35@xmath0 m and 11.6@xmath0 m .
the adopted continuum fluxes are reported in table [ tab : sample ] .
unlike in the spitzer spectra , the lines observed with crires are velocity - resolved and present a broad range of widths that makes it inappropriate to adopt fixed pixel ranges for their measurement .
to adopt a homogeneous method , we therefore measure the line flux included within the full width at 10% of the line peak .
table [ tab : fluxes ] reports the fluxes and errors measured for the lines listed in table [ tab : lines ] .
the four water lines at @xmath32.93@xmath0 m are measured together into a single flux value , because in most spectra they are blended together due to their large line widths .
aatau & @xmath34 0.25 & 0.61 @xmath33 0.04 & 2.48 @xmath33 0.10 & 0.11 @xmath33 0.00 & 0.29 @xmath33 0.03 & 2.01 @xmath33 0.08 & & 2.74 @xmath33 0.02 + as205 n & 7.33 @xmath33 2.71 & 11.26 @xmath33 0.61 & 29.32 @xmath33 1.88 & 2.06 @xmath33 0.76 & 6.20 @xmath33 0.61 & 30.84 @xmath33 1.34 & 17.98 @xmath33 7.19 & 13.38 @xmath33 5.35 + as209 & & & & & & & 1.04 @xmath33 0.03 & 0.71 @xmath33 0.02 + cvcha & & & & & & & 3.55 @xmath33 0.11 & 1.97 @xmath33 0.07 + cwtau & @xmath34 0.46 & 3.59 @xmath33 0.09 & 8.42 @xmath33 0.06 & 1.17 @xmath33 0.21 & 1.54 @xmath33 0.07 & 10.66 @xmath33 0.13 & 5.80 @xmath33 0.93 & 5.51 @xmath33 0.88 + dftau & 4.55 @xmath33 0.59 & & & 1.65 @xmath33 0.21 & & & 7.99 @xmath33 0.72 & 3.25 @xmath33 0.29 + dgtau & & 2.49 @xmath33 0.20 & 4.70 @xmath33 0.70 & & 5.50 @xmath33 0.18 & 16.60 @xmath33 0.64 & 5.42 @xmath33 0.43 &
+ doar24e s & @xmath34 1.09 & 1.66 @xmath33 0.15 & 9.85 @xmath33 0.35 & 1.92 @xmath33 0.73 & 2.13 @xmath33 0.12 & 6.44 @xmath33 0.34 & 10.15 @xmath33 3.45 & 5.51 @xmath33 1.87 + doar44 & @xmath34 2.02 & 0.52 @xmath33 0.05 & 4.39 @xmath33 0.16 & @xmath34 0.56 & 0.36 @xmath33 0.05 & 3.00 @xmath33 0.15 & & 2.24 @xmath33 0.03 + dotau & 2.31 @xmath33 0.19 & 1.46 @xmath33 0.07 & 2.63 @xmath33 0.33 & 1.01 @xmath33 0.08 & 2.19 @xmath33 0.06 & 3.39 @xmath33 0.22 & 4.44 @xmath33 0.27 & 0.59 @xmath33 0.04 + drtau & 2.84 @xmath33 0.43 & 5.25 @xmath33 0.12 & 10.11 @xmath33 0.13 & 0.80 @xmath33 0.12 & 2.37 @xmath33 0.13 & 6.97 @xmath33 0.28 & 8.36 @xmath33 1.34 & 6.55 @xmath33 1.05 + ec82 & @xmath34 0.60 & @xmath34 0.36 & 2.05 @xmath33 0.18 & 0.39 @xmath33 0.02 & @xmath34 0.33 & 1.73 @xmath33 0.17 & & 2.04 @xmath33 0.02 + exlup08 & 8.69 @xmath33 2.61 & @xmath34 0.99 & 5.99 @xmath33 0.90 & 3.88 @xmath33 1.16 & 1.89 @xmath33 0.24 & 3.85 @xmath33 1.53 & 13.20 @xmath33 7.92 & 5.51 @xmath33 3.31 + exlup14 & 0.32 @xmath33 0.02 & 1.00 @xmath33 0.16 & 3.43 @xmath33 0.48 & 0.12 @xmath33 0.01 & @xmath34 0.46 & @xmath34 1.14 & 0.70 @xmath33 0.02 & 0.89 @xmath33 0.03 + fntau & @xmath34 0.05 & @xmath34 0.19 & @xmath34 0.27 & @xmath34 0.03 & 0.21 @xmath33 0.05 & @xmath34 0.42 & & 0.29 @xmath33 0.01 + fztau & & 5.94 @xmath33 0.10 & 6.44 @xmath33 0.05 & & 1.67 @xmath33 0.08 & 5.38 @xmath33 0.11 & & 9.59 @xmath33 0.68 + gqlup & @xmath34 0.57 & 0.70 @xmath33 0.04 & 4.92 @xmath33 0.09 & @xmath34 0.26 & 0.51 @xmath33 0.04 & 3.06 @xmath33 0.11 & 4.18 @xmath33 0.21 & 2.22 @xmath33 0.12 + hd36112 & & @xmath34 1.85 & @xmath34 6.55 & & @xmath34 1.81 & 9.76 @xmath33 1.36 & & 8.85 @xmath33 3.45 + hd95881 @xmath135 & & @xmath34 2.24 & @xmath34 3.00 & & @xmath34 2.19 & @xmath34 2.03 & & 4.91 @xmath33 0.20 + hd97048 @xmath135 & & & & & & & & 12.30 @xmath33 0.40 + hd98922 @xmath135 & @xmath34 0.08 & @xmath34 7.48 & @xmath34 6.05 & @xmath34 0.08 & @xmath34 7.29 & 7.29 @xmath33 2.66 & & 5.08 @xmath33 0.14 + hd101412 & & @xmath34 1.17 & @xmath34 1.27 & & @xmath34 1.14 & 1.89 @xmath33 0.25 & & + hd135344b & @xmath34 0.30 & @xmath34 0.65 & @xmath34 1.84 & @xmath34 0.14 & @xmath34 0.63 & 3.66 @xmath33 0.36 & & 1.99 @xmath33 0.79 + hd139614 & & & & & & & & 1.36 @xmath33 0.05 + hd141569 @xmath135 & & & & & & & & 1.77 @xmath33 0.14 + hd142527 & & @xmath34 2.37 & @xmath34 10.41 & & @xmath34 2.31 & 14.30 @xmath33 2.00 & & 8.58 @xmath33 4.29 + hd144432s & & & & & & & & 2.73 @xmath33 0.71 + hd150193 @xmath135 & & @xmath34 3.55 & @xmath34 7.00 & & @xmath34 3.46 & 10.40 @xmath33 0.94 & & 8.22 @xmath33 1.94 + hd163296 @xmath135 & & @xmath34 3.91 & 7.21 @xmath33 1.02 & & @xmath34 3.83 & 24.77 @xmath33 0.99 & & 18.70 @xmath33 1.60 + hd179218 @xmath135 & & @xmath34 5.27 & @xmath34 9.89 & & @xmath34 4.50 & @xmath34 4.89 & & 2.88 @xmath33 0.74 + hd190073 @xmath135 & & @xmath34 2.43 & @xmath34 1.18 & & @xmath34 2.38 & 5.77 @xmath33 0.59 & & 6.99 @xmath33 0.59 + hd244604 & & @xmath34 1.53 & @xmath34 1.66 & & @xmath34 1.50 & @xmath34 0.95 & & 0.64 @xmath33 0.01 + hd250550 @xmath135 & @xmath34 0.14 & & & 0.14 @xmath33 0.01 & & & & 5.80 @xmath33 0.35 + htlup & @xmath34 0.73 & @xmath34 1.15 & @xmath34 1.19 & 0.36 @xmath33 0.06 & 1.26 @xmath33 0.31 & 2.58 @xmath33 0.48 & 2.27 @xmath33 0.25 & 0.49 @xmath33 0.06 + imlup & @xmath34 0.18 & @xmath34 0.18 & 0.65 @xmath33 0.12 & @xmath34 0.08 & 0.34 @xmath33 0.04 & @xmath34 0.34 & 0.44 @xmath33 0.01 & 0.07 @xmath33 0.11 + irs48 & & & & & & & & 0.58 @xmath33 0.13 + lkha330 & & @xmath34 0.23 & @xmath34 1.77 &
& @xmath34 0.23 & @xmath34 0.82 & & 1.01 @xmath33 0.09 + rno90 & 2.61 @xmath33 0.34 & 7.38 @xmath33 0.26 & 23.29 @xmath33 0.35 & 1.40 @xmath33 0.18 & 2.98 @xmath33 0.25 & 19.66 @xmath33 0.35 & 12.93 @xmath33 2.07 & 6.24 @xmath33 1.00 + rulup & 4.27 @xmath33 0.60 & 3.08 @xmath33 0.16 & 5.00 @xmath33 0.23 & 1.60 @xmath33 0.22 & 6.08 @xmath33 0.16 & 7.39 @xmath33 0.21 & 6.63 @xmath33 0.73 & 4.93 @xmath33 0.54 + rylup &
& @xmath34 0.96 & @xmath34 1.52 & & @xmath34 0.93 & 2.11 @xmath33 0.30 & & 1.76 @xmath33 0.09 + rwaur & & 6.79 @xmath33 0.20 & 9.48 @xmath33 0.17 & & 2.09 @xmath33 0.17 & 8.35 @xmath33 0.25 & 8.47 @xmath33 0.38 & 5.83 @xmath33 0.25 + scra s & 6.46 @xmath33 2.00 & 8.59 @xmath33 0.87 & 23.16 @xmath33 2.33 & 3.85 @xmath33 1.19 & 13.32 @xmath33 0.82 & 40.11 @xmath33 1.69 & 13.25 @xmath33 2.92 & 4.86 @xmath33 1.07 + scra n & 8.68 @xmath33 2.69 & 8.59 @xmath33 0.87 & 23.16 @xmath33 2.33 & 3.75 @xmath33 1.16 & 13.32 @xmath33 0.82 & 40.11 @xmath33 1.69 & 8.11 @xmath33 1.78 & 5.76 @xmath33 1.27 + sr9 & & @xmath34 0.93 & 2.42 @xmath33 0.21 & & @xmath34 0.91 & 3.27 @xmath33 0.13 & & 1.31 @xmath33 0.03 + sr21 & @xmath34 0.03 & @xmath34 1.16 & @xmath34 4.60 & @xmath34 0.02 & @xmath34 1.13 & 14.42 @xmath33 0.98 & & 0.46 @xmath33 0.02 + ttau n & 12.77 @xmath33 10.22 & & & 6.14 @xmath33 4.91 & & & 32.12 @xmath33 25.70 & 9.65 @xmath33 7.72 + ttau s & 9.66 @xmath33 7.73 & & & 11.31 @xmath33 9.05 & & & 37.56 @xmath33 30.05 & + twcha & 0.49 @xmath33 0.00 & 0.62 @xmath33 0.02 & 2.42 @xmath33 0.05 & 0.10 @xmath33 0.00 & 0.19 @xmath33 0.02 & 1.41 @xmath33 0.04 & & 1.05 @xmath33 0.01 + twhya & @xmath34 0.02 & 0.20 @xmath33 0.05 & 3.96 @xmath33 0.47 & @xmath34 0.01 & 0.58 @xmath33 0.04 & 10.21 @xmath33 0.40 & & 0.64 @xmath33 0.00 + vssg1 & & 2.11 @xmath33 0.26 & 7.49 @xmath33 0.45 & & 1.20 @xmath33 0.24 & 4.94 @xmath33 0.35 & 4.83 @xmath33 0.29 & 2.88 @xmath33 0.17 + vvser & & @xmath34 1.27 & @xmath34 1.73 & & 1.88 @xmath33 0.32 & 4.48 @xmath33 0.25 & & 2.02 @xmath33 0.36 + vwcha & 2.07 @xmath33 0.11 & 1.97 @xmath33 0.08 & 7.14 @xmath33 0.08 & 1.32 @xmath33 0.07 & 1.38 @xmath33 0.07 & 9.85 @xmath33 0.12 & 6.46 @xmath33 0.30 & 2.96 @xmath33 0.14 + vzcha & 0.74 @xmath33 0.01 & 0.83 @xmath33 0.05 & 2.46 @xmath33 0.07 & 0.30 @xmath33 0.00 & 0.52 @xmath33 0.05 & 1.97 @xmath33 0.05 & 3.11 @xmath33 0.03 & 1.01 @xmath33 0.02 + waoph6 & 2.66 @xmath33 0.19 & 1.29 @xmath33 0.04 & 2.56 @xmath33 0.14 & 1.32 @xmath33 0.09 & 1.00 @xmath33 0.04 & 3.30 @xmath33 0.12 & 2.75 @xmath33 0.14 & 0.96 @xmath33 0.05 + wxcha & @xmath34 0.51 & 1.18 @xmath33 0.04 & 3.08 @xmath33 0.07 & 0.32 @xmath33 0.01 & 0.39 @xmath33 0.04 & 1.63 @xmath33 0.06 & 1.98 @xmath33 0.03 & 0.66 @xmath33 0.02 + figures [ fig : phot_corr_1 ] and [ fig : phot_corr_2 ] show the stellar photospheric correction of crires 2.9@xmath0 m spectra as explained in section [ sec : ana3 ] .
this procedure proves to be effective in correcting most science spectra , those that match well with a spectral template and where the disk gas emission lines are broader than the absorption lines from the same molecules in the stellar photosphere . in some cases ,
however , there is no unambiguous / acceptable spectral match with a template ( in cwtau , htlup , and imlup ) , or the disk emission lines are possibly as narrow as the photospheric lines ( in fntau and twhya ) ; in these cases , the residuals after photospheric correction leave larger uncertainties on potential detections of h@xmath14o and oh and on their line profiles . figures [ fig : mol eq bc ] , [ fig : mol spca ] , [ fig : mol unclear ] , [ fig : mol eq co1 - 0 ] , [ fig : mol unclear1 ] , and [ fig : no h2o oh ] show fits to crires 2.9@xmath0 m spectra made by using the co line profiles as models for h@xmath14o and oh emission , as explained in section [ sec : ana4 ] .
in addition to fitting the 2.9@xmath0 m data with the nc or bc separately ( in those disks that have two co components ) , we also combine nc and bc profiles in different fractions , a technique that we call spectral principal component analysis ( spca ) as being based on principal spectral components given by nc and bc ( figure [ fig : mol spca ] ) . in the double - component disks , we can therefore test whether h@xmath14o and oh are composed of one rather than two components , and which one dominates . to determine which model best represents the data between nc , bc , and their combination , we use the akaike information criterion @xcite , defined as aic = @xmath136 the number of free model parameters + the chisquare value .
the criterion penalizes models with a larger number of free parameters , so that a simpler model is preferred unless the more complex model decreases the aic number significantly ( we adopt the criterion of a @xmath137aic @xmath138 , corresponding to a @xmath139 higher probability of minimizing the information loss ) . in most cases ,
the chisquare value is enough to identify the best model between bc and nc .
this procedure is generally robust against the photospheric correction performed on the 2.9@xmath0 m velocity - resolved spectra obtained by crires ( section [ sec : ana3 ] ) , because the disk emission lines are typically broader than photospheric lines from the same molecules ( figures [ fig : phot stand ] , [ fig : mol comp ] , [ fig : phot_corr_1 ] , [ fig : phot_corr_2 ] ) , and because the high resolving power of crires allows to separate emission components with different velocities . in two disks ( fntau , twhya )
, however , photospheric lines have similar width as the possibly detected oh emission lines , while h@xmath14o is undetected ; their oh line fluxes and profiles are therefore highly uncertain ( see figure [ fig : mol unclear1 ] ) .
overall , both non - detections and detections of weak emission advocate for the predominance of bc over nc in h@xmath14o and oh emission at 2.9@xmath0 m . in a few cases ,
spectra are best fit with a combination of bc and nc : specifically as205 , drtau , and scran ( and only oh in cwtau ) .
these three disks are also the only ones where nc was detected in the co @xmath60 lines , possibly due to the higher s / n and dust veiling in these spectra compared to other spectra in the survey @xcite .
two other disks show tentative evidence for a weak nc , but its presence can not be firmly established due to uncertainties from the photospheric correction : rulup and scras . in all disks ,
the nc is about 210 times ( or more ) weaker in h@xmath14o and oh than in co , as compared to bc . in a few cases only ,
the oh profile is better matched by nc rather than by bc ( dftau , rno90 , vwcha ) . in single - component disks ,
oh is detected only in ec82 and twcha .
figure [ fig : lum_trends ] shows the measured h@xmath14o and oh line fluxes plotted against the stellar+accretion luminosity , taken as a proxy for the irradiation and heating of inner disks . only the oh line at 30@xmath0 m shows a steady flux increase with increasing luminosity , consistent with an increasing area of the warm disk emitting region .
hotter oh lines show instead a turning point at a luminosity of @xmath73l@xmath23 , beyond which line fluxes decrease ( similarly to what seen in the h@xmath14o at 33@xmath0 m ) .
figure [ fig : accretion_figure ] shows the measured line fluxes plotted against the accretion luminosity alone , to highlight how line fluxes overall increase with accretion luminosity .
the last two plots in the bottom right of the figure show the t - r diagram of co emission , highlighting how there is not global trend between the sequence of co gaps and the accretion luminosity and rate .
oh emission is more frequently detected than water emission in disks , and is tentatively detected at 30@xmath0 m out to r@xmath30/r@xmath140 ( figure [ fig : r_fl_corr ] ) and up to larger luminosities ( figure [ fig : lum_trends ] ) , possibly supporting a still active photo - dissociation of trace amounts of h@xmath14o gas .
oh formation through water photo - dissociation populates higher energy levels first ( e.g. the 12.6@xmath0 m line , see * ? ? ?
* ) , though , which are less often detected in disks that have a large r@xmath30 .
it is therefore unclear whether a cold oh reservoir traced at 30@xmath0 m may be formed by photo - dissociation of h@xmath14o or by gas - phase chemistry , known to be efficient at temperatures of 100300k ( 100300k , e.g. * ? ? ?
future observations of velocity - resolved h@xmath14o and oh lines at wavelengths longer then 2.9@xmath0 m will be able to test if two reservoirs of oh exist : one due to uv photo - dissociation of water ( with similar line profiles for h@xmath14o and oh , as found at 2.9@xmath0 m ) , which dominates the hotter oh lines , and one due to oh formation in cold environments where water formation and/or release to the gas phase is less efficient ( by finding larger line profiles for h@xmath14o and smaller for oh ) . | we present a new velocity - resolved survey of 2.9@xmath0 m spectra of hot h@xmath1o and oh gas emission from protoplanetary disks , obtained with crires at the vlt ( @xmath2@xmath33kms@xmath4 ) . with the addition of archival _
spitzer_-irs spectra , this is the most comprehensive spectral dataset of water vapor emission from disks ever assembled .
we provide line fluxes at 2.933@xmath0 m that probe from disk radii of @xmath5au out to the region across the water snow line . with a combined dataset for 55 disks
, we find a new correlation between h@xmath1o line fluxes and the radius of co gas emission as measured in velocity - resolved 4.7@xmath0 m spectra ( r@xmath6 ) , which probes molecular gaps in inner disks .
we find that h@xmath1o emission disappears from 2.9@xmath0 m ( hotter water ) to 33@xmath0 m ( colder water ) as r@xmath6 increases and expands out to the snow line radius .
these results suggest that the infrared water spectrum is a tracer of inside - out water depletion within the snow line .
it also helps clarifying an unsolved discrepancy between water observations and models , by finding that disks around stars of m@xmath7 m@xmath8 generally have inner gaps with depleted molecular gas content .
we measure radial trends in h@xmath1o , oh , and co line fluxes that can be used as benchmarks for models to study the chemical composition and evolution of planet - forming disk regions at 0.0520au .
we propose that jwst spectroscopy of molecular gas may be used as a probe of inner disk gas depletion , complementary to the larger gaps and holes detected by direct imaging and by alma . |
it is generally believed that active galaxies and quasars are powered by the presence of massive black holes ( bhs ) in their nuclei , and that such bhs are present in many , possibly all , quiescent galaxies as well ( see kormendy & richstone 1995 , lynden - bell 1996 and rees 1996 for reviews of this paradigm and its history ) .
evidence for this can be derived from studies of the dynamics of stars and gas in the nuclei of individual galaxies .
the high spatial resolution data that can now be obtained with the hubble space telescope ( hst ) allows the existing evidence to be strengthened considerably .
the present paper is part of a new hst study of the quiescent galaxy m32 , in which the presence of a bh has long been suspected based on the steep central rotation velocity gradient and nuclear peak in the velocity dispersion seen in ground - based data ( e.g. , tonry 1987 ; van der marel 1994a ; bender , kormendy & dehnen 1996 ) .
the main results of our project were summarized , and discussed in the context of other recent work , in van der marel ( 1997a ) . the acquisition and reduction of the stellar kinematical hst data were described in van der marel , de zeeuw & rix ( 1997b ; hereafter paper i ) . here
we present new dynamical models that we have used to interpret the combined hst and ground - based data .
two types of self - consistent dynamical models have been constructed previously to interpret the ground - based data for m32 .
dressler & richstone ( 1988 ) and richstone , bower & dressler ( 1990 ) used a method based on schwarzschild s ( 1979 ) technique , in which individual orbits are calculated and superposed , to provide a self - consistent model that fits a given set of data .
these ` maximum entropy ' ( richstone & tremaine 1988 ) models could fit the ( then available ) data only by invoking the presence of a central dark mass of @xmath13@xmath14 .
the models are general in the sense that they make no assumptions about the dynamical structure of the galaxy . however , a drawback was that only spherical geometry was considered . even though the models can be made to rotate , it remains unclear what systematic errors are introduced when they are applied to a flattened ( e3 ) galaxy like m32 .
an alternative approach has been to construct axisymmetric models with phase - space distribution functions ( dfs ) that depend only on the two classical integrals of motion , @xmath15 , where @xmath16 is the binding energy and @xmath17 is the angular momentum component along the symmetry axis , both per unit mass .
these models properly take flattening and rotation into account . to fit the m32 data
, they require the presence of a central dark mass between @xmath18 ( van der marel 1994b ; qian 1995 ; dehnen 1995 ) and @xmath19 ( bender , kormendy & dehnen 1996 ) .
the disadvantage of these models is that they have a special dynamical structure .
the velocity dispersions in the meridional plane are isotropic , @xmath20 , which might not be the case in m32 .
however , the models do fit the observed line - of - sight velocity profile ( vp ) shapes without invoking freely adjustable parameters , which provides some reason to believe that the m32 df may not be too different from the form @xmath0 . the previous work on m32 has shown that models with a bh can fit the ground - based data , but the modeling has not been general enough to demonstrate that a bh is required .
below that their data can be fit by an axisymmetric model without a bh .
] in fact , the spatial resolution of the ground - based data might have been insufficient for this to be the case ( see appendix a for a discussion of this issue ) .
it has certainly not been sufficient to rule out a cluster of dark objects ( as opposed to a central bh ) on the basis of theoretical arguments ; goodman & lee ( 1989 ) showed that this requires a resolution of @xmath21 .
our new hst data of the nuclear region of m32 were obtained with the hst faint object spectrograph ( fos ) through square apertures of @xmath22 and @xmath23 , respectively , yielding the highest spatial resolution stellar kinematical data for m32 obtained to date . the results ( see , e.g. , figure [ f : hstbhpred ] below ) show a steeper rotation curve and higher central velocity dispersion than the best ground - based m32 data .
the primary goals of our project are to determine whether these new hst data rigorously rule out models without any dark mass , and to what extent they constrain the mass and size of the dark object in m32 . to obtain constraints on the presence of a dark object that are least dependent on a priori assumptions about the df , we need to compare the hst data not only to the predictions of axisymmetric @xmath0 models , but also to the predictions of models with a fully general dynamical structure .
orbit superposition techniques provide the most straightforward approach to construct such models .
however , orbit superposition is more difficult to implement for the axisymmetric case than for the spherical case : the orbits are not planar and typically possess an additional integral of motion , so that the orbit library must sample three rather than two integrals of motion .
furthermore , a larger parameter space must be explored because of the unknown inclination angle .
this implies that larger amounts of cpu time and computer memory are required .
however , other than that , there are no reasons why such models would be infeasible .
motivated by the increased speed and memory capacity of computers , we therefore developed a technique to construct fully general axisymmetric orbit superposition models , that fit any given number of observed photometric and kinematic constraints .
independent software implementations were written by h .-
w.r , n.c . and r.v.d.m .
our technique may be viewed as the axisymmetric generalization of the spherical modeling used by richstone and collaborators , with the important additional feature that we calculate vp shapes and include an arbitrary number of gauss - hermite moments in the fit .
we take into account the error on each observational constraint to obtain an objective @xmath24 measure for the quality - of - fit .
our basic algorithm is described in rix ( 1997 ; hereafter r97 ) and summarized in de zeeuw ( 1997 ) . r97
provide an application to the spherical geometry ; cretton ( 1997 ; hereafter c97 ) present the extension to the axisymmetric case . here
we summarize the main steps of the axisymmetric algorithm briefly , and focus on the application to m32 .
the resulting models are the most general yet constructed for m32 .
the paper is organized as follows . in section [ s : massmodel ]
we discuss our parametrizations for the stellar mass density and for the potential of the dark object . in section
[ s : dfpred ] we describe the construction of models with @xmath0 dfs , and in section [ s : datacomp ] we compare the predictions of these models , both with and without bhs , to the kinematical data . in section [ s : technique ] we outline the orbit - superposition technique for constructing models with a fully general dynamical structure , and in section [ s : results ] we compare the predictions of these models to the data .
we construct models with an extended dark object in section [ s : extfit ] .
we summarize and discuss our main conclusions in section [ s : conc ] .
readers interested primarily in the results of our models may wish to skip sections [ s : massmodel ] , [ s : dfpred ] and [ s : technique ] .
we adopt a parametrized form for the axisymmetric mass density of m32 : @xmath25^{\beta } \ > [ 1+(m / c)^2]^{\gamma } , \qquad m^2 \equiv r^2 + ( z / q)^2 . \label{massdens}\ ] ] the mass density @xmath26 is related to the luminosity density @xmath27 according to @xmath28 , where @xmath29 is the average mass - to - light ratio of the stellar population ( hereafter given in solar v - band units ) . both @xmath29 and the intrinsic axial ratio @xmath30
are assumed to be constant ( as a function of radius ) .
the projected axial ratio @xmath31 is determined by the inclination @xmath32 according to @xmath33 .
the parameters @xmath29 and @xmath32 can be freely specified ; all other parameters are determined by fitting to the available m32 surface photometry .
the highest spatial resolution surface photometry available for m32 is that presented by lauer ( 1992 ) , based on pre - costar hst / wfpc images .
their measurements extend to @xmath34 from the nucleus . at larger radii ground - based data are available from kent ( 1987 ) and peletier ( 1993 ) .
figure [ f : surfbr ] shows the major axis surface brightness measurements from these sources .
the solid curve shows the surface brightness profile predicted by our model , for @xmath35 , @xmath36 , @xmath37 , @xmath38 , @xmath39 , @xmath40 , @xmath41 , @xmath42 , and an assumed distance of @xmath43 .
the factor @xmath44^{\gamma}$ ] in equation ( [ massdens ] ) ensures that the model has finite mass , and that it provides an adequate fit to the observed surface brightness profile out to @xmath45 .
apart from this factor , the model is identical to that used by van der marel ( 1994b ) and qian ( 1995 ) ( dashed curve in figure [ f : surfbr ] ) .
our model for the mass density is somewhat less general than that used by dehnen ( 1995 ) , who deprojected the surface photometry in an unparametrized manner .
his approach avoids possible biases resulting from the choice of an _ ad hoc _ parametrization ( merritt & tremblay 1994 ; gebhardt 1996 ) .
it also allows the axial ratio of m32 to vary with radius .
the observed axial ratio is very close to constant at @xmath46 in the central @xmath47 , but increases slowly to @xmath48 at @xmath49 .
even though our model does not reproduce this modest variation , overall it provides an excellent fit , and is fully adequate for a study of the nuclear dynamics .
the uncertainties in the interpretation of the kinematic data for the center of m32 are due almost entirely to our ignorance of the dynamical structure of m32 .
the uncertainties introduced by errors in the brightness profile or by the non - uniqueness of the deprojection are relatively minor ( van den bosch 1997 ) .
the effect of possible triaxiality is more difficult to assess , but we will argue in section [ ss : triaxiality ] that triaxiality is unlikely to modify any of the major conclusions of our paper .
the gravitational potential is assumed to be @xmath50 , where @xmath51 is the potential generated by the luminous matter with mass density ( [ massdens ] ) , and @xmath52 allows for the possibility of a massive dark object in the nucleus .
we assume the latter to be @xmath53 which is the potential generated by a cluster with a plummer model mass density ( e.g. , binney & tremaine 1987 ) . for @xmath54
one obtains the case of a dark nuclear point mass , i.e. , a nuclear bh .
we do not include the potential of a possible dark halo around m32 .
there are no ( strong ) observational constraints on the possible presence and characteristics of such a dark halo , and even if present , it will not affect the stellar kinematics near the nucleus of m32 .
the regular orbits in general axisymmetric potentials are characterized by three integrals of motion , the binding energy @xmath55 , the component of the angular momentum around the symmetry axis @xmath56 , and a non - classical , or effective , third integral @xmath57 ( ollongren 1962 ; richstone 1982 ; binney & tremaine 1987 ) . in any given axisymmetric potential
there is an infinity of dfs @xmath58 that generate a given axisymmetric mass density @xmath59 .
such models are difficult to construct , primarily because the third integral can not generally be expressed explicitly in terms of the phase - space coordinates .
however , for any mass density @xmath59 there is exactly one df that is even in @xmath17 , and does not depend on @xmath57 .
this unique even ` two - integral ' df , @xmath60 , provides a useful low - order approximation to any axisymmetric model , and has the convenient property that many physical quantities , including the df itself , can be calculated semi - analytically .
we study models of this type for m32 because they have successfully reproduced ground - based m32 data , and because they provide a useful guide for the interpretation of more general three - integral models , which are discussed in sections [ s : technique ] and [ s : results ] . to calculate the @xmath60 dfs for our models we have used a combination of the techniques described in qian ( 1995 ) and dehnen ( 1995 ) , even though either technique by itself could have been used to get the same result ( in fact , yet another technique to address this problem is described in c97 ) .
initially , four radial regimes are considered : @xmath61 ; @xmath62 ; @xmath63 ; and @xmath64 , with @xmath65 and @xmath66 as defined in equation ( [ massdens ] ) . in these regimes
the mass density is approximately : @xmath67 ; @xmath68^{\beta}$ ] ; @xmath69^{\gamma}$ ] ; and @xmath70 , respectively . for each of these mass densities
@xmath60 can be calculated with the technique and software of qian ( 1995 ) .
the dfs for the four regimes are then smoothly patched together in energy , to yield an approximation to the full df .
this approximation is then used as the starting point for lucy - richardson iteration as described in dehnen ( 1995 ) .
this yields the df @xmath71 , reproducing the model mass density to @xmath72 per cent rms .
the total df is the sum of the part @xmath73 that is even in @xmath17 , and the part @xmath74 that is odd in @xmath17 . in principle , @xmath75 is determined completely by the mean streaming velocities @xmath76 , but these are not determined well enough by the data to make an inversion practicable . instead , therefore
, @xmath75 can be freely specified so as to best fit the data , with the only constraint that the total df should be positive definite .
none of the main conclusions of our paper depend sensitively on the particular parametrization used for @xmath74 , so we restrict ourselves here to a simple choice ( section [ ss : bhfit ] ) .
once the complete df is known , the projected line - of - sight vps can be calculated for any particular observational setup as in qian ( 1995 ) . from the vps , predictions can be calculated for the observable kinematical quantities .
figure [ f : hstbhpred ] shows the hst / fos data presented in paper i , obtained with the apertures ` 0.1-pair ' ( @xmath22 square ) and ` 0.25-pair ' ( @xmath23 square ) .
the figure also shows the highest available spatial resolution ground - based data , obtained by van der marel ( 1994a ) with the william herschel telescope ( wht ) , and by bender , kormendy & dehnen ( 1996 ) with the canada - france hawaii telescope ( cfht ) .
the spatial resolution for these ground - based observations is roughly @xmath77 and @xmath78 fwhm , respectively .
the predictions of @xmath0 models for the wht and cfht data have already been discussed in detail by previous authors , and we therefore focus here on a comparison of @xmath0 models to the new hst data .
the curves in figure [ f : hstbhpred ] are the predictions of edge - on @xmath0 models with dark nuclear _ point masses _
( i.e. , bhs ) .
these models have an intrinsic axial ratio @xmath79 .
the dark mass only influences the kinematical predictions in the central few arcsec , and the mass - to - light ratio @xmath29 was therefore chosen to fit the normalization of the kinematical data at larger radii .
a good fit to the wht data between @xmath80 and @xmath81 is obtained with @xmath82 .
predictions were calculated for each individual observation , taking into account the aperture position , aperture size and psf for the hst data as given in paper i. connecting curves in the figure were drawn to guide the eye .
it proved sufficient to study only models with a very simple odd part @xmath74 , namely those that produce a total df in which at every @xmath83 a fraction @xmath84 of the stars has @xmath85 , and a fraction @xmath86 has @xmath87 . for each @xmath8 , the fraction @xmath88 of stars with @xmath85 in the model
was chosen to optimize the @xmath24 of the fit to the rotation curve . the displayed models _ with _ a bh fit the rotation curve well , more or less independent of @xmath8 ; models with higher @xmath8 require smaller @xmath84 .
the models differ primarily in their predictions for the velocity dispersions .
the observed trend of increasing velocity dispersion towards the nucleus is successfully reproduced by models with a nuclear point mass of @xmath89 .
the model without a bh predicts a roughly constant velocity dispersion with radius , and is strongly ruled out .
in fact , this model also fails to fit the observed rotation velocity gradient in the central arcsec . the displayed model for the no - bh case is maximally rotating ( @xmath90 ) , and it is thus not possible to improve this by choosing a more general form for the odd part of the df . for a quantitative analysis of the best - fitting @xmath8 , we define a @xmath24 statistic that measures the quality of the model fit to the observed hst velocity dispersions : @xmath91^{1/2 } .
\label{chisigdef}\ ] ] figure [ f : chisqbh ] shows the relative rms residual @xmath92 as function of @xmath8 .
the best fit to the hst dispersions is obtained for @xmath93 .
the quoted error is a formal @xmath4 error based on the assumption of gaussian statistics ( probably an underestimate , as there is some hint for systematic errors in the data in addition to random errors ) .
an alternative way of estimating @xmath8 is to model the average of the four data points within @xmath94 from the center : @xmath95 , where the @xmath4 error is based on the scatter between the data points .
figure [ f : sigcen ] shows the observed @xmath96 as a hatched region ; a solid curve shows the predicted value as a function of @xmath8 .
the predictions fall in the observed range for @xmath97 .
the inclination of m32 can not be derived from the observed photometry and is therefore a free parameter in the modeling . however , the predictions of @xmath0 models are rather insensitive to the assumed inclination ( van der marel 1994b ) .
this was verified by also calculating the predictions of inclined models with @xmath2 , which have an intrinsic axial ratio @xmath98 .
a mass - to - light ratio @xmath99 was adopted , so that at large radii one obtains the same rms projected velocity on the intermediate axis ( between the major and minor axes ) as for an edge - on model . on the major axis
the @xmath2 model then predicts a slightly higher rms velocity than the edge - on model , resulting in a @xmath100 per cent smaller best - fitting @xmath8 .
apart from this , the conclusions from the inclined models were found to be identical to those for the edge - on models ( cf .
figure [ f : chisqsurf ] below).=-2 the @xmath0 models that fit the hst data can only be correct if they also fit the ground - based wht and cfht data .
one may define similar @xmath24 quantities as for the hst data , to determine the best - fitting @xmath8 for either of these data sets .
figures [ f : chisqbh ] and [ f : sigcen ] show the relative rms residual @xmath92 of the fit to all the dispersion measurements , and the average dispersion @xmath101 of the dispersion measurements centered with @xmath94 from the nucleus . for the wht data ,
@xmath92 is minimized for @xmath102 , whereas for the cfht data it is minimized for @xmath103 .
the central velocity dispersion measured with the wht is best fit with @xmath104 , whereas for the cfht data it is best fit with @xmath105 .
these results are roughly consistent with those of previous authors .
van der marel ( 1994b ) , qian ( 1995 ) and dehnen ( 1995 ) found @xmath106 for the best fitting edge - on @xmath0 model to the ground - based wht data , while bender , kormendy & dehnen obtained a best fit to their higher spatial resolution cfht data with @xmath107 .
the latter value is somewhat higher than the one we find here , because it was chosen only to provide a good fit to the cfht rotation curve ; it does not fit the cfht velocity dispersions very well .
these results indicate that , under the assumption of an @xmath0 df , the different observations can not all be fit simultaneously with the same @xmath8 , even after accounting for the different observational setups .
the lowest spatial resolution wht data require a significantly lower @xmath8 than the highest spatial resolution hst data .
this implies that m32 has a df that is not of the form @xmath15 .
to construct more general three - integral models for m32 , we extended schwarzschild s orbit superposition algorithm . its basic structure is to calculate an orbit library that samples integral space in some complete and uniform way , to store the time - averaged intrinsic and projected properties of the orbits , and to search for the weighted superposition of orbits that best fits the observed kinematics , while reproducing the mass density @xmath59 for self - consistency .
here we summarize the main steps , with emphasis on those aspects that are unique to the m32 application .
complete descriptions of the technique are given in r97 and c97 .
we sample integral space with an @xmath108 grid .
the quantity @xmath109 is the radius of the circular orbit in the equatorial plane with energy @xmath16 .
its angular momentum , @xmath110 , is the maximum angular momentum at the given energy .
we define @xmath111 . for fixed @xmath112 , the position of a star in the meridional ( @xmath113 plane is restricted to the region bounded by the ` zero - velocity - curve ' ( zvc ) , defined by the equation @xmath114 , where @xmath115 is the ` effective gravitational potential ' ( binney & tremaine 1987 ) .
we parametrize the third integral at each @xmath112 using an angle @xmath116 , which fixes the position at which an orbit touches the zvc ( cf . figure [ f : zvcplot ] ) .
the quantity @xmath117 was sampled using 20 logarithmically spaced values between @xmath118 arcsec , and @xmath119 arcsec .
this range of radii contains all but a fraction @xmath120 of the stellar mass of m32 .
the quantity @xmath121 was sampled using an ` open ' grid ( in the same sense that numerical quadrature formulae can be open or closed , e.g. , press 1992 ) of @xmath122 values , spaced linearly between @xmath123 and @xmath124 , i.e. , @xmath125 , for @xmath126 .
the quantity @xmath116 was sampled using an ` open ' grid of 7 values , spaced linearly between 0 and @xmath127 . here
@xmath127 is the angle @xmath116 for the ` thin tube ' orbit at the given @xmath128 ( see figure [ f : zvcplot ] ) .
the special values @xmath129 and @xmath130 ( meridional plane and circular orbits ) and the special values @xmath131 and @xmath132 ( equatorial plane and thin tube orbits ) are presumed to be represented by their closest neighbors on the grid , but are not included explicitly .
an orbit was integrated for each @xmath108 combination , starting with @xmath133 from the zvc .
the integration time was 200 times the period of the circular orbit at the same energy .
this is sufficient to properly sample the phase - space trajectory for the large majority of orbits , although it exceeds the hubble time only at radii @xmath45 .
the integrations yield the ` orbital phase - space density ' for each orbit , as described in c97 .
these were binned onto : ( i ) an @xmath134 grid in the meridional plane ; ( ii ) an @xmath135 grid on the projected @xmath136 plane of the sky ; and ( iii ) several cartesian @xmath137 cubes , with @xmath138 the line - of - sight velocity .
the first two grids were chosen logarithmic in @xmath139 , with identical bins to those used for @xmath117 , and linear in @xmath140 $ ] .
the @xmath137 cubes were centered on @xmath141 , with @xmath142 square spatial cells , and 91 velocity bins of @xmath143 .
spatial cell sizes were adopted of @xmath144 , @xmath145 and @xmath78 , respectively .
the @xmath137 cubes were used to calculate , for each orbit , the predicted line - of - sight velocity histograms for all positions and setups for which kinematical data are available , taking into account the observational point - spread - functions ( psfs ) and aperture positions , orientations and sizes , as described in c97 .
construction of a model consists of finding a weighted superposition of the orbits in the library that reproduces two sets of constraints : * self - consistency constraints .
the model should reproduce the masses predicted by the luminous density @xmath59 ( section [ s : massmodel ] ) , for each cell of the meridional @xmath134 grid , for each cell of the projected @xmath135 grid , and for each aperture on the sky for which there is data . * kinematical constraints .
the model should reproduce the observed kinematics of the galaxy , including vp shapes .
we chose to express this as a set of linear constraints on the gauss - hermite moments @xmath146 of the vps ( r97 ) .
smoothness of the solutions in integral space may be enforced by adding extra regularization constraints ( see section [ ss : smoothmod ] below ) .
the quality of the fit to the combined constraints can be measured through a @xmath24 quantity ( r97 ) .
the assessment of the fit to the kinematical constraints includes the observational errors . in principle
, one would like to fit the self - consistency constraints with machine precision . in practice
this is unfeasible , because the projected mass constraints are not independent from each other ( aperture positions for different data sets partly overlap ) and from the meridional plane mass constraints .
it was found that models with no kinematical constraints could at best fit the masses with a fractional error of @xmath147 .
motivated by this , fractional ` errors ' of this size were assigned to all the masses in the self - consistency constraints .
as described in r97 and c97 , we use the nnls routine of lawson & hanson ( 1974 ) to determine the combination of non - negative orbital occupancies ( which need not be unique ) that minimizes the combined @xmath24 .
the model predictions have a finite numerical accuracy , due to , e.g. , gridding and discretization .
tests show that the numerical errors in the predicted kinematics of our method are @xmath148 for the rotation velocities and velocity dispersions , and @xmath149 in the gauss - hermite moments ( c97 ) .
numerical errors of this magnitude have only an insignificant effect on the data - model comparison ( see appendix a ) .
we have studied three - integral models with dark central _ point_-masses , i.e. , @xmath54 in equation ( [ potent ] ) .
there are then three free model parameters : the inclination @xmath32 , the mass - to - light ratio @xmath29 , and the bh mass @xmath8 .
the parameter space must be explored through separate sets of orbit libraries .
the potentials of such models are identical except for a normalization factor , and the orbits are therefore identical except for a velocity scaling
. each @xmath150 combination does require a separate nnls fit to the constraints . ] .
as for the two - integral models , we have studied only two , widely spaced , inclinations : @xmath1 and @xmath2 . for each inclination
we have sampled the physically interesting range of @xmath150 combinations . for each @xmath150 combination
we determined the orbital weights ( and hence the dynamical structure ) that best fit the data , and the corresponding goodness - of - fit quantity @xmath151 .
all available hst / fos , cfht and wht data were included as kinematical constraints on the models .
older kinematical data for m32 were not included because of their lower spatial resolution and/or poorer sky coverage . in total , each nnls fit had 1960 orbits to fit 782 constraints : 366 self - consistency constraints , and 416 kinematical constraints ( for 86 positions on the projected plane of the galaxy ) .
we focus primarily ( figures [ f : contour][f : datafitsatoh ] ) on models without additional constraints that enforce smoothness in integral - space .
this is a sufficient and conservative approach for addressing the primary question of our paper : _ which models are ruled out by the m32 data ? _
if the data can not even be fit with an arbitrarily unsmooth df , they certainly can not be fit with a smooth df .
figure [ f : contour ] shows the main result : contour plots of @xmath152 for both inclinations that were studied .
the displayed @xmath24 measures the quality of the fit to the kinematical constraints only ; the actual nnls fits were done to both the kinematical and the self - consistency constraints , but contour plots of the total @xmath24 look similar .
the overall minimum @xmath24 values are obtained for : @xmath153 and @xmath154 for @xmath1 ; and @xmath155 and @xmath156 for @xmath2 .
figure [ f : bestfits ] compares the kinematical predictions of the best - fitting edge - on and @xmath2 models to the data .
two problems with the data must be taken into account when assessing the quality of the fit .
first , the hst velocity dispersions show a scatter between some neighboring points that is much larger than the formal errors , most likely due to some unknown systematic effect .
the models can not be expected to reproduce this .
second , the cfht rotation velocities at radii @xmath157 exceed the wht measurements by an amount which can not be attributed to differences in spatial resolution , but must be due to some unknown systematic error in either of the two data sets .
the wht data have the smallest error bars , and therefore receive most weight in the nnls fit . as a result ,
the models tend to underpredict the cfht rotation curve .
these systematic problems with the data preclude the use of @xmath24 as a meaningful statistic to assess which models provide an acceptable fit : if the observations themselves are not mutually consistent , then clearly no model can be statistically consistent with all of them .
although the use of any statistical test is suspect in the presence of systematic errors , one may still assign confidence regions on the model parameters by using the _ relative likelihood _
statistic @xmath158 .
this statistic merely measures which parameter combinations provide an equally good ( or bad ) fit as the one(s ) that yield the minimum @xmath24 .
if we assume that the observational errors are normally distributed ( which , as mentioned , is likely to be an oversimplification ) , then @xmath159 follows a @xmath24 probability distribution with the number of degrees of freedom equal to the number of model parameters ( press 1992 ) . the best - fitting edge - on model in figure [ f : bestfits ] has @xmath160 while the best - fitting @xmath2 model has @xmath161 , both for @xmath162 degrees of freedom
the fact that @xmath163 , even for these optimum fits , is due primarily to the systematic errors in the data . to the eye
, the models appear to fit the data as well as could be hoped for .
the @xmath24 values do indicate that the @xmath2 model provides a significantly better fit than the edge - on model , implying that m32 is not seen edge - on .
however , the results presented here do not allow us to derive the actual inclination of m32 .
that would require a detailed study of the entire range of possible inclinations , which would be more computer - intensive .
the important conclusion in the present context is that the topology of the @xmath24 contours in figure [ f : contour ] is virtually identical for both inclinations : the allowed range for @xmath8 is therefore uninfluenced by our ignorance of the true inclination of m32 .
the @xmath159 statistic was used to assign confidence values to the contours in figure [ f : contour ] . at the @xmath164 per cent confidence level ( @xmath4 for a gaussian probability distribution ) ,
the allowed @xmath8 fall in the range @xmath165@xmath166 for @xmath1 , and in the range @xmath167@xmath168 for @xmath2 . at the @xmath169 per cent confidence level ( @xmath5 for a gaussian probability distribution )
, they fall in the ranges @xmath170@xmath171 and @xmath172@xmath173 , respectively . in reality , small numerical errors in the models might have distorted the @xmath24 contours .
we address this issue in appendix a. any numerical errors are small enough that they have no influence on our conclusion that models without a dark mass are firmly ruled out .
however , the possibility of small numerical errors does increase the confidence bands on @xmath8 .
based on the analysis in appendix a , we conclude that @xmath174 at @xmath164 per cent confidence , and @xmath175 at @xmath169 per cent confidence .
these estimates take into account both the observational errors in the data and possible numerical errors in the models , and are valid for both inclinations that were studied .
figure [ f : datafitsatoh ] compares the model predictions to the observed rotation velocities and velocity dispersions for all the models labeled in figure [ f : contour ] .
models c&g are the overall best fits for the two inclinations
. models b&f and d&h are ( approximately ) the best - fitting models for @xmath176 and @xmath177 , respectively .
the latter models are marginally ruled out at the @xmath178% confidence level ( cf . the above discussion ) ,
although to the eye they do appear to reproduce the main features of the data .
they differ from the overall best - fitting models primarily in their predictions for the hst velocity dispersions .
the differences in the predictions for the ground - based data are smaller ( and invisible to the eye in figure [ f : datafitsatoh ] ) , but nonetheless more statistically significant because of the smaller error bars for these data .
models a&e , the best fits without a central dark mass , are indisputably ruled out .
the main problem for these models is to fit the central peak in the velocity dispersion .
they come rather close to fitting the wht observations , and predict a central dispersion of @xmath179
. however , the models without a dark mass fail to reproduce the higher central dispersion of @xmath180 measured with the cfht ( although still only marginally ) , and do nt even come close to reproducing the hst dispersions , which exceed @xmath181 in the central @xmath94 . the top row in figure [ f : intspace ] illustrates the orbital occupancies in integral space for the best - fitting edge - on model ( model c ) .
only a small fraction of the orbits in the library is used to fit the constraints , while the remainder receives zero weight .
this yields an equilibrium solution of the collisionless boltzmann equation , but is not physically plausible .
smoothness of the solutions in integral space can be enforced by adding linear regularization constraints to the problem ( zhao 1996 ; c97 ) .
we have explored this only in an _ ad hoc _ way , merely to be able to assess the effect of smoothness constraints on the resulting fit to the data .
a model is defined as a set of masses @xmath182 in integral space . for each point that is not on the boundary of the @xmath108 grid
, we measure the smoothness of the model ( press 1992 ; eq . [
18.5.10 ] ) through the second order divided differences ( in each of the three variables , assuming for simplicity that the distances between adjacent grid points are equal in all directions ) of the function @xmath183 .
the function @xmath184 is a rough approximation to the energy dependence of the model , obtained , e.g. , by studying the spherical isotropic limit of the given mass density .
the regularization constraints are then that the divided differences should equal @xmath185 , where the ` error ' @xmath186 determines the amount of smoothing .
models with @xmath187 have no smoothing , while models with @xmath188 force @xmath183 to be a linear function on the @xmath108 grid .
the second and third rows in figure [ f : intspace ] show the integral space for model c with the addition of either a modest ( @xmath189 ) or a large ( @xmath190 ) amount of regularization in the nnls fit , respectively . as the bottom panels show
, the price paid for the increased smoothness is a somewhat poorer fit to the data .
however , the fits are still quite good .
this demonstrates that the good fits to the data shown in figure [ f : bestfits ] are not primarily the result of the use of implausible distributions in integral space .
these distributions result from the numerical properties of the problem , but there also exist smooth solutions which provide similar fits .
figure [ f : moments ] illustrates the dynamical structure of the edge - on models a d and the @xmath2 models e h .
by contrast to the models with a bh , the models a&e without a bh invoke a large amount of radial motion in the central arcsec to produce a peak in the observed velocity dispersions ( cf .
binney & mamon 1982 ) .
the maximum allowed radial anisotropy in the models is determined by the observed rotation velocities in m32 , because dynamical models predict lower values of @xmath191 when they are more radially anisotropic ( richstone , bower & dressler 1990 ; de bruijne , van der marel & de zeeuw 1996 ) . figure [ f : datafitsatoh ] shows that the allowed radial anisotropy is by far insufficient to fit the observed peak in the velocity dispersion profile without invoking a bh . the models b d and f h , which represent the best fits for different potentials , all have a similar dynamical structure in that they are dominated by azimuthal motion .
this is most pronounced for the intrinsically flatter @xmath2 models .
it is not known what physical process could have produced this particular velocity ellipsoid shape in m32 .
the rightmost panels in figure [ f : moments ] show the velocity moments for @xmath0 models with the same gravitational potentials as models c&g .
the @xmath0 models are similar to the best - fitting three - integral models , in that they have an excess of azimuthal motion .
this is why they have been so successful in fitting ground - based data , including available vp shape parameters , and it shows that they provide a useful low - order approximation to the dynamical structure of m32 .
however , the @xmath0 models have @xmath192 by definition , in contrast to the inequality between @xmath193 and @xmath194 seen in the best - fitting three - integral models .
this explains the finding of section [ ss : dfform ] that @xmath0 models can not successfully reproduce _ all _ observed features of the kinematical data .
the second velocity moments @xmath193 and @xmath194 can be combined with the mixed moment @xmath195 to determine the tilt of the velocity ellipsoid in the meridional plane .
for all the models a d and e h we found @xmath195 to be small , and the velocity ellipsoids are more closely aligned with spherical coordinate axes than with cylindrical coordinate axes ( but they are not perfectly aligned with either ) .
this is not uncommon in three - integral models for axisymmetric systems ( e.g. , dejonghe & de zeeuw 1988 ; dehnen & gerhard 1993 ; de zeeuw , evans & schwarzschild 1996 ) .
figure [ f : vsig ] shows the quantity @xmath196 in the equatorial plane , for the models c and g. the inclined , and intrinsically flatter , model g rotates faster than the edge - on model c. the predictions for oblate isotropic rotator models ( @xmath197 ) are shown for comparison .
the edge - on model c rotates faster than an oblate isotropic rotator model for radii @xmath198 ; the inclined model g rotates slower than the oblate isotropic rotator model .
we have not attempted to derive confidence bands on the dynamical structure of m32 .
this is a much more difficult problem than the derivation of confidence bands on the model parameters @xmath199 , and is beyond the scope of the present paper .
the results in the previous section demonstrate that m32 must have a massive dark object in its nucleus . to obtain a limit on the size of this dark object ,
we have studied models in which it has a finite size @xmath200 ( cf .
. [ [ potent ] ] ) .
searching the parameter space of three - integral models with different @xmath201 is extremely computer - intensive .
we have therefore restricted ourselves to @xmath0 models with extended dark objects .
this is not likely to bias our conclusions , because the best - fitting three - integral models found in section [ s : results ] are similar to two - integral models .
figure [ f : hstextpr ] shows the predictions of edge - on @xmath0 models with extended nuclear dark objects , for two representative values of @xmath8 . as in the models of section [ ss : bhfit ] ,
the fraction @xmath84 of stars with @xmath85 in each model was chosen to best fit the rotation curve . by
adjusting @xmath84 , adequate fits to the rotation velocities can be obtained for most relevant models .
hence , only the velocity dispersions are shown in the figure .
the predicted dispersions are typically constant or decreasing towards the center within the scale radius @xmath201 .
the hst data show a much higher velocity dispersion in the center than further out .
hence , the extension of any possible dark nuclear cluster can not be large ; figure [ f : hstextpr ] suggests @xmath202 .
figure [ f : chisqsurf ] shows a contour plot of the quantity @xmath203 ( eq . [ [ chisigdef ] ] ) , measuring the quality of the model fit to the hst dispersion measurements , both for the edge - on case and for @xmath2 .
the best fit is obtained for @xmath204 ( the point mass case discussed previously ) .
the formal @xmath169 per cent confidence level ( assuming gaussian formal errors ) rules out all models with @xmath205 , independent of the inclination .
the models with the largest @xmath201 must have a total mass of at least @xmath206 . at the distance of m32 , @xmath207 .
hence , the upper limit on the scale radius corresponds to @xmath208 .
combined with a total mass in the cluster of @xmath206 , this implies a central mass density of at least @xmath209 .
the half - mass radius of a plummer model is @xmath210 .
hence , there must be @xmath211 inside @xmath212 .
the total v - band luminosity inside this radius than ours . ] is @xmath213 , implying a luminous mass of @xmath214 .
hence , the ratio of the total mass to luminosity inside this radius must be @xmath215 .
the observed kinematics constrain only the amount of mass in the system , not whether this mass is luminous or dark .
one can therefore fit the data equally well with models in which the average mass - to - light ratio @xmath29 of the stellar population increases towards the nucleus , and in which there is no dark mass .
we have not explicitly constructed such models , but it is clear from the preceding discussion that in such models @xmath29 must rise from @xmath216 in the main body of the galaxy to @xmath217 at @xmath218 .
such a drastic variation in mass - to - light ratio would imply a strong change in the stellar population , accompanied by broad - band color gradients .
the size of these gradients depends on the actual stellar population mix , which is unknown .
however , one may use the properties of main sequence stars as a guideline . for these
, a change in @xmath29 from @xmath219 to @xmath220 implies color changes @xmath221 , @xmath222 , and @xmath223 ( using the tables of stellar properties in allen 1973 ) .
such variations between @xmath94 and @xmath224 should have been obvious in photometric observations .
however , neither subarcsec resolution ground - based imaging ( lugger 1992 ) nor pre - refurbishment hst imaging ( crane 1993 ) have revealed any significant color gradients in the central arcsec of m32 .
post - refurbishment hst observations ( lauer et al . , private communication ) also do not show strong color gradients .
thus , the nuclear mass concentration in m32 can not be due merely to a change in the mix of ordinary stars in the nuclear region.=-2 the absence of observed color gradients does not exclude the possibility of nuclear concentrations of brown dwarfs , white dwarfs , neutron stars , or stellar - mass bhs in m32 .
however , at high densities , such clusters of dark objects are not stable over a hubble time .
this was discussed by goodman & lee ( 1989 ) , and their arguments were updated and extended in van der marel ( 1997a ) .
the latter paper shows that the new hst limit on the density of dark material in m32 rules out all but the most implausible clusters , leaving a single massive bh as the most likely interpretation of the data .
the kinematical predictions of our models depend on the assumed plummer form of the extended dark object
. if it is a cluster of collapsed objects , this distributed dark mass may itself be cusped ( e.g. , gerhard 1994 ) . however , the limit on @xmath201 results from the fact that the dispersion of stars with mass density ( [ massdens ] ) in a plummer potential does not have a ( strong ) central peak .
this property is common to many alternative types of models , such as those of king ( e.g. , binney & tremaine 1987 ) , hernquist ( 1990 ) and jaffe ( 1983 ) .
these all produce constant or decreasing dispersions inside their scale radius , as does the plummer model . hence
, the upper bound on @xmath201 derived here is likely to be generic to most plausible density profiles for the extended dark object .
in addition , king , jaffe and hernquist models are more centrally condensed than plummer models , and would therefore require dark clusters of even higher central densities to fit the data .
the main bottlenecks in proving the presence of nuclear bhs in quiescent galaxies from stellar kinematical data have long been : ( i ) the restricted spatial resolution of ground - based data ; and ( ii ) lack of sufficiently general dynamical models to rule out constant mass - to - light ratio models beyond doubt .
the hst now provides spectra of superior spatial resolution . to fully exploit the potential of these new data it is imperative to improve the modeling techniques that have been used in the past decade
the situation is considerably more complicated than for gas disks in ( active ) galaxies , where the assumption of simple circular orbits is often adequate .
interpretation of stellar kinematical data for flattened elliptical galaxies ideally requires axisymmetric ( or even better , triaxial ) dynamical models with completely general three - integral distribution functions .
such models have not previously been constructed for any stellar kinematical bh candidate galaxy .
we therefore developed a technique for the construction of such axisymmetric models , and used it to interpret our hst data for m32 . to guide the construction and interpretation of the three - integral models we first compared the new hst data to the predictions of @xmath0 models , which have been used extensively to interpret ground - based m32 data .
such models have the advantage that the df can be calculated semi - analytically , but have the disadvantage of having a special dynamical structure , with @xmath225 everywhere .
there is no a priori reason why any galaxy should have this property .
however , the fact that @xmath0 models fit the observed vp shapes inferred from ground - based data to within @xmath226% ( in terms of deviations from a gaussian ) , suggested that the m32 df might in fact be close to the form @xmath0 .
we find here that @xmath0 models for m32 can also fit the new hst data , and that this requires the presence of a nuclear dark mass , as was the case for the ground - based data .
however , the best fitting dark mass of @xmath227@xmath228 is larger than the @xmath229@xmath230 that best fits the @xmath231 spatial resolution data from the cfht , and is even more different from the @xmath232@xmath233 that best - fits the @xmath234 spatial resolution data from the wht .
thus , under the assumption of an @xmath0 df , the different data sets can not be fit with the same @xmath8 .
this indicates that the m32 df is not of the form @xmath0 , although it might be close to it . to obtain a model - independent estimate of the best - fitting @xmath8 , and to firmly rule out models without any dark mass ,
it is necessary to study more general three - integral models .
we have made such models for m32 , both with and without central bhs , and for various possible values of the average mass - to - light ratio @xmath29 of the stellar population .
the models were constructed to fit all available kinematical hst , cfht and wht data , and the acceptability of each model was assessed through the @xmath24 of its fit to the data .
the models demonstrate explicitly for the first time that there is no axisymmetric constant mass - to - light ratio model that can fit the kinematical data without invoking the presence of a nuclear dark mass , independent of the dynamical structure of m32 .
a nuclear dark point mass of @xmath235 is required ( with @xmath4 and @xmath5 error bars of @xmath6 and @xmath236 , respectively , which includes the possible effect of small numerical errors in the models ) .
this mass is similar to that quoted by most previous papers , but the confidence on the detection of a nuclear dark mass in m32 is now much higher .
constant mass - to - light ratio models still come very close to fitting the ground - based data , and only the new hst data make the case for a nuclear dark mass clear - cut .
the inclination of m32 can not be inferred from the available surface photometry , and is therefore a free parameter in the modeling .
ideally one would like to construct dynamical models for all possible inclinations ( which would be very computer - intensive ) and determine the inclination that best fits the kinematical data .
here we have taken the more modest approach of constructing models for only two representative inclinations : @xmath1 ( edge - on ) and @xmath2 .
the intrinsic axial ratios for these inclinations are @xmath79 and @xmath98 , respectively .
the three - integral @xmath2 models provide a better fit than the edge - on models , which suggests that m32 is not seen edge - on .
however , the allowed range for @xmath8 does not depend sensitively on the assumed inclination : models with no central dark mass are firmly ruled out for both inclinations .
so even though a more detailed study of the full inclination range for m32 would improve our knowledge of the true inclination and intrinsic axial ratio of m32 , it would probably not change significantly the constraints on the central dark mass .
the best - fitting three - integral models are similar to @xmath0 models in that they have an excess of azimuthal motion .
this is why they have been so successful in fitting ground - based data , including available vp shape parameters , and it confirms that they provide a useful low - order approximation to the dynamical structure of m32 .
however , @xmath0 models do have @xmath237 .
this does not reproduce the inequality between @xmath238 and @xmath239 , nor the modest tilt of the velocity ellipsoid indicated by the small @xmath195 term , seen in the best - fitting three - integral models .
this is why @xmath0 models can not successfully explain _ all _ observed features of the kinematical data . to constrain the size of the dark object in m32
we have constructed @xmath0 models with an extended dark nuclear object .
these show that the hst data put an upper limit of @xmath240 on the half - mass radius of the nuclear dark object , implying a central density exceeding @xmath11 .
this limit on the density of dark material in m32 essentially rules out nuclear clusters of planets , brown dwarfs , white dwarfs , neutron stars , or smaller mass bhs ( van der marel 1997a ) .
the absence of color gradients in the central arcsec of m32 implies that the nuclear mass concentration can also not be attributed to a stellar population gradient .
a single massive nuclear bh therefore provides the most plausible interpretation of the data .
axisymmetric dynamical models with a nuclear bh provide an excellent fit to all available kinematical data for m32 .
however , to be physically meaningful , the models must also be dynamically stable . in van der marel , sigurdsson & hernquist ( 1997c ) we presented n - body simulations of the @xmath0 models for m32 .
the models were found to be completely stable , both for @xmath1 and for @xmath2 .
this shows that dynamical stability is not a problem for the models , and that the inclination of m32 can not be meaningfully constrained through stability arguments .
we have not evolved n - body models for the best - fitting three - integral models , but we expect these models to be stable as well , given their similarity to @xmath0 models .
the two - body relaxation time in m32 can be estimated as in , e.g. , binney & tremaine ( 1987 ; eq . [ 8 - 71 ] ) .
using the relevant quantities for our best - fitting dynamical model , we find for solar mass stars in the central cusp ( @xmath241 ) that @xmath242 .
the time scale for ` resonant relaxation ' ( rauch & tremaine 1996 ) is of the same order .
the central cusp must therefore be evolving secularly over a hubble time .
however , the diffusion of stars in phase space is slow enough that one may assume the evolution to be through a sequence of quasi - equilibrium models .
this justifies our approach of modeling m32 as a collisionless equilibrium system .
studies of the secular evolution of the m32 cusp will be interesting , but will not change the need for a nuclear dark object .
in fact , the process of dynamical relaxation supports the presence of a dark object : without a dark object the relaxation would proceed at a much more rapid rate that is difficult to reconcile with observations ( lauer 1992 ) .
one remaining uncertainty in our dynamical modeling is the possibility of triaxiality .
after the step from spherical models to axisymmetric models , triaxial models are the obvious next step .
however , there are several reasons to believe that for m32 this additional step will be less important .
first , m32 is known not to be spherical , but there is no reason why it can not be axisymmetric .
there is no significant isophote twisting in m32 , and no minor axis rotation .
this does not mean that m32 can not be triaxial ( we might be observing it from one of the principal planes ) , but it also does not mean that m32 needs to be triaxial .
second , spherical constant mass - to - light ratio models ( without a nuclear dark mass ) for ground - based m32 data failed to fit by only a few @xmath243 , and it was quite conceivable that axisymmetry could fix this ( which it did , cf . figure [ f : drfit ] below ) .
however , axisymmetric constant mass - to - light ratio models for the new hst data fail to fit the nuclear velocity dispersion by @xmath244 km / s , and this can not likely be fixed through triaxiality .
third , theoretical arguments suggest that strongly triaxial models with density cusps as steep as in m32 may not be stable , owing to the fact that regular box - orbits are replaced by boxlets and irregular orbits that may not be able to sustain a triaxial shape ( binney & gerhard 1985 ; merritt & fridman 1996 ; merritt & valluri 1996 ; see also the review by de zeeuw 1996 ) .
rapidly - rotating low - luminosity elliptical galaxies like m32 always have steep power - law cusps ( faber 1997 ) , and may therefore be axisymmetric as a class ( de zeeuw & carollo 1996 ) .
this is consistent with statistical studies of their intrinsic shapes ( e.g. , merritt & tremblay 1996 ) .
so , apart from the fact that triaxiality is unlikely to remove the need for a central dark object in m32 , it may even be so that m32 can not be significantly triaxial .
the growth of a black hole into a stellar system is adiabatic if it occurs over a time scale that is ` long ' ( see sigurdsson , quinlan , & hernquist 1995 for a quantitative discussion ) compared to the typical orbital period of the stars . for the case of m32
, the black hole formation can be considered adiabatic if it took at least @xmath12 .
young ( 1980 ) studied the adiabatic growth of bhs in spherical isothermal models with central density @xmath245 and core radius @xmath246 .
the bh growth leaves the mass density at large radii unchanged , but induces a central cusp @xmath247 for @xmath248 .
the form of the density profile at intermediate radii is determined by the dimensionless parameter @xmath249 $ ] , which measures the ratio of the bh mass to the initial core mass .
lauer ( 1992 ) showed that the shape of the m32 brightness profile measured with hst can be well fit with @xmath250 .
the radial and density normalization implied by the data are then @xmath251 and @xmath252 .
photometric _ model therefore implies that @xmath253 . although this result depends somewhat on the assumed isothermality of the initial distribution ( quinlan , hernquist & sigurdsson 1995 ) , it is quite remarkable that our best - fitting _ dynamical _ models have exactly @xmath254 , for both inclinations that we studied .
the m32 data are therefore fully consistent with the presence of a bh that grew adiabatically into a pre - existing core .
this is similar to the situation for m87 ( cf .
young 1978 ; harms 1994 ) .
lee & goodman ( 1989 ) extended young s calculations to the case of rotating models .
for the value of @xmath255 implied by the photometry , their models predict a profile of @xmath256 that is approximately flat with radius ( with amplitude fixed by the axial ratio of the system ) . however , this result depends very sensitively on the assumed rotation law of the initial model .
the radial variations in @xmath257 seen in our best - fitting models ( figure [ f : vsig ] ) are probably equally consistent with the adiabatic growth hypothesis .
a star of mass @xmath258 and radius @xmath259 on a circular orbit of radius @xmath260 will be tidally disrupted if @xmath261 ( e.g. , binney & petit 1988 ) .
thus , disruption of a solar type star by the bh in m32 will occur inside @xmath262 arcsec
. a disruption event will be highly luminous , but is not predicted to occur more often than once every @xmath263 ( rees 1988 ) .
the minimum pericenter distance for a star with given @xmath128 in a kepler potential is @xmath264 , where as before , @xmath117 is the radius of the circular orbit at the given energy and @xmath265 .
the kinematical data for m32 only meaningfully constrain the df for energies with @xmath266 . for @xmath267 , only stars with @xmath268 have @xmath269 . the data
do not constrain variations in the df over such a small range in @xmath121 , and our dynamical models therefore can not address the existence and properties of the so - called ` loss cone ' ( frank & rees 1976 ; lightman & shapiro 1977 ) . for the @xmath121-grid
that we have employed , all solar type stars on orbits with @xmath270 have @xmath271 .
even giants with @xmath272 have @xmath273 .
this justifies our neglect of tidal disruption in the orbit calculations .
an interesting question is why bhs in quiescent galaxies are nt more luminous ( e.g. , kormendy & richstone 1995 ) . for m32 ,
the total x - ray luminosity is @xmath274 ( eskridge , white & davis 1996 ) , the far infrared luminosity is @xmath275 ( knapp 1989 ) , and for the @xmath276 radio emission @xmath277 ( roberts 1991 ) .
part or all of the observed x - ray emission may be due to low - mass x - ray binaries , so the total luminosity due to accretion onto the bh in m32 is @xmath278 . by contrast
, the eddington luminosity of the bh is @xmath279 . for a canonical mass - loss rate of @xmath280 ( faber & gallagher 1976 ) ,
the stars that are bound to the bh in m32 shed @xmath281 of gas as a result of normal stellar evolution .
if a fraction @xmath282 of this gas is steadily accreted with efficiency @xmath201 , it produces a luminosity @xmath283 .
thus either the accretion fraction @xmath282 or the accretion efficiency @xmath201 must be very small in m32 .
thin disk accretion with @xmath284 requires @xmath285 , which is possible ( the accretion fraction is difficult to predict theoretically , because it depends on the hydrodynamics of the stellar winds that shed the gas ) , but may be implausibly low .
instead , it appears more likely that @xmath201 is small , since there is a family of ` advection dominated ' accretion solutions that naturally predict such low efficiencies .
models of this type successfully explain the ` micro - activity ' of the bh ( sgr a@xmath286 ) in our own galaxy ( narayan , yi & mahadevan 1995 ) . in a typical accretion model of this type ( narayan & yi 1995 , their fig .
11 ) , @xmath287 suffices to explain the upper bound on @xmath288 for m32 .
future observations of m32 will include spectra with the new long - slit hst spectrograph stis .
these will provide significantly better sky coverage than our fos data , but the spatial resolution will be similar .
the high - resolution hst data can be complemented with that from fully two - dimensional ground - based spectrographs , such as oasis on the cfht and sauron on the wht .
these combined data will yield improved constraints on the bh mass , on the orbital structure and inclination of m32 , and on possible deviations from axisymmetry .
we dedicate this paper , and its companions r97 and c97 , to the memory of martin schwarzschild , who pioneered the modeling technique employed here .
martin s sense of purpose , his exceptional clarity of thinking , his transparent personal integrity and , most of all , his genuine warm interest and support , remain a great source of inspiration to us .
support for this work was provided by nasa through grant number # go-05847.01 - 94a , and through a hubble fellowship # hf-1065.01 - 94a awarded to rpvdm , both from the space telescope science institute which is operated by the association of universities for research in astronomy , incorporated , under nasa contract nas5 - 26555 .
nc acknowledges financial support from the swiss government ( etat du valais ) and nuffic , and the hospitality of steward observatory , the mpa garching and geneva observatory .
tdz is grateful for the generous hospitality of the institute for advanced study .
both he and nc received financial support from the leids kerkhoven bosscha fonds .
in this appendix we discuss the topology of the @xmath152 contours for the edge - on orbit - superposition models .
the top panels of figure [ f : subdatachi ] show the @xmath24 contours when only ( subsets of ) the ground - based wht data are included in the fit .
these panels can be compared to figure [ f : subdatachi]d , which shows the contours for the case in which all wht , cfht and hst data are included .
figure [ f : subdatachi]a shows the @xmath24 contours when only the major axis @xmath289 and @xmath290 wht measurements are fit .
binney & mamon ( 1982 ) showed that a large range of gravitational potentials can fit any given observed velocity dispersion profile .
the valley seen in the @xmath24 contours is a consequence of this : it outlines a one - parameter family of models that can fit the data with different velocity dispersion anisotropy . for
a non - rotating spherical system , only models that require negative second velocity moments are ruled out . for a rotating system like m32 , the observed rotation rate sets additional limits on the allowed radial anisotropy .
for the case of the major axis @xmath289 and @xmath290 wht measurements , a no - bh model is just marginally acceptable at @xmath169 per cent confidence , cf . figure [ f : subdatachi]a . for the lower - spatial resolution
major axis @xmath289 and @xmath290 measurements of dressler & richstone ( 1988 ) such a model is entirely acceptable .
figure [ f : drfit ] compares the predictions of the best - fit axisymmetric orbit - superposition model without a bh to their data .
richstone , bower & dressler ( 1990 ) concluded that these data could not be fit by any _ spherical _
model without a bh .
this is because spherical models allow less rotation , and therefore failed to fit the observed rotation velocities .
this underscores the importance of making axisymmetric models for flattened galaxies like m32 .
vp shape measurements provide independent constraints on the velocity dispersion anisotropy .
figure [ f : subdatachi]b shows the @xmath24 contours for edge - on orbit - superposition models when not only the wht major axis @xmath289 and @xmath290 measurements are fit , but also the major axis vp shape measurements . with the inclusion of the vp shapes , models without a bh
are ruled out .
figure [ f : subdatachi]c shows the @xmath24 contours when also the wht measurements along other position angles are included , which contracts the allowed @xmath8 range to @xmath291@xmath292 at the formal @xmath169 per cent confidence level .
the wht data by themselves therefore rule out axisymmetric models without a bh .
however , the models without a bh still come very close to fitting the data , and , e.g. , fail to fit the central velocity dispersion by only 12@xmath243 ( cf .
figure [ f : datafitsatoh ] ) .
so one can not make a particularly strong claim for a bh on the basis of the wht data alone , because it is conceivable that the fit could be improved with , e.g. , only a minor amount of triaxiality .
the same holds for the cfht data , but the new hst data do make the case for a bh in m32 clear - cut .
the contours for the case in which all the available wht , cfht and hst kinematical data are included in the fit ( figure [ f : subdatachi]d ) show one global @xmath24 minimum , and a second local minimum .
the presence of a global minimum does not necessarily imply that the combined data constrain a single best - fit potential .
it might be that there is a small range of potentials that all fit equally well , but that such a range of constant @xmath24 would not be evident due to the finite numerical accuracy of our technique . in figure
[ f : subdatachi]e we show explicitly how the topology of the @xmath24 contours might have been influenced by the possibility of small numerical errors in our models . it was obtained from figure [ f : subdatachi]d by recalculating the @xmath24 contours after adding a random error @xmath293 $ ] km / s and @xmath294 $ ] ( cf . section [ s : technique ] ) to the prediction for each data point , for each @xmath150 combination .
the results show that numerical errors can indeed influence the @xmath24 contours near the @xmath24 minimum .
thus , the second minimum in figure [ f : subdatachi]d might be the result of numerical inaccuracies in our technique .
however , the numerical errors are small enough that they only have a negligible effect on the overall @xmath24 topology .
in particular , models without a dark mass remain firmly ruled out .
to assess the possible effect of numerical errors on the confidence bands for @xmath8 , we constructed 100 figures like figure [ f : subdatachi]e using different random realizations .
for each we determined the position of the @xmath24 minimum , and the minimum and maximum @xmath8 for which there is an @xmath29 such that the model with @xmath199 falls within the @xmath169 per cent confidence region .
the results are plotted in figure [ f : subdatachi]f .
all allowed @xmath8 values fall in the range @xmath295@xmath296 .
thus , @xmath297 at @xmath169 per cent confidence .
similar experiments show that @xmath298 at @xmath164 per cent confidence .
experiments for @xmath2 produced similar results , and mass ranges that were either the same or slightly smaller .
thus , we conclude that the @xmath4 and @xmath5 errors on the estimated @xmath299 , are @xmath300 and @xmath236 , respectively . | axisymmetric dynamical models are constructed for the e3 galaxy m32 to interpret high spatial resolution stellar kinematical data obtained with the hubble space telescope ( hst ) .
models are studied with two - integral , @xmath0 , phase - space distribution functions , and with fully general three - integral distribution functions .
the latter are built using an extension of schwarzschild s approach : individual orbits in the axisymmetric potential are calculated numerically , and populated using non - negative least - squares fitting so as to reproduce all available kinematical data , including line - of - sight velocity profile shapes .
the details of this method are described in companion papers by rix et al . and cretton et al
. models are constructed for inclinations @xmath1 ( edge - on ) and @xmath2 .
no model without a nuclear dark object can fit the combined ground - based and hst data , independent of the dynamical structure of m32 .
models with a nuclear dark object of mass @xmath3 ( with @xmath4 and @xmath5 error bars of @xmath6 and @xmath7 , respectively ) do provide an excellent fit .
the inclined models provide the best fit , but the inferred @xmath8 does not depend sensitively on the assumed inclination
. the models that best fit the data are not two - integral models , but like two - integral models they are azimuthally anisotropic .
two - integral models therefore provide useful low - order approximations to the dynamical structure of m32 .
we use them to show that an extended dark object can fit the data only if its half - mass radius is @xmath9 ( @xmath10 ) , implying a central dark matter density exceeding @xmath11 .
the inferred @xmath8 is consistent with that suggested previously by ground - based kinematical data .
however , radially anisotropic axisymmetric constant mass - to - light ratio models are now ruled out for the first time , and the limit on the dark matter density implied by the hst data is now stringent enough to rule out most plausible alternatives to a massive black hole .
thus , the evidence for a massive black hole in the quiescent galaxy m32 is now very compelling .
the dynamically inferred @xmath8 is identical to that suggested by existing models for hst photometry of m32 that assume adiabatic growth ( over a time scale exceeding @xmath12 ) of a black hole into a pre - existing core .
the low activity of the nucleus of m32 implies either that only a very small fraction of the gas that is shed by evolving stars is accreted onto the black hole , or alternatively , that accretion proceeds at very low efficiency , e.g. in an advection - dominated mode . # |
high - transition - temperature ( high-@xmath7 ) superconductivity occurs in lamellar copper oxides when holes @xcite or electrons @xcite are doped into the cuo@xmath5 planes . for the parent compounds of hole - doped materials such as la@xmath5cuo@xmath8 and yba@xmath5cu@xmath6o@xmath9 , the cu@xmath10 spins order at relatively high temperatures ( @xmath11300 k and 420 k , respectively ) in a simple antiferromagnetic ( af ) collinear structure that doubles the crystallographic unit cell in the cuo@xmath5 planes @xcite .
although the parent compound of electron - doped copper oxides such as nd@xmath1cuo@xmath2 also has af spin structures doubling the cuo@xmath5 unit cell , the cu@xmath10 moments order in three phases with two different af noncollinear spin structures as shown in figs .
1(a ) and 1(b ) @xcite .
these noncollinear spin structures appear in nd@xmath1cuo@xmath2 because of the presence of the magnetic exchange interaction between cu@xmath10 and nd@xmath12 @xcite .
compared to the hole - doped la@xmath13sr@xmath14cuo@xmath8 , the long - range af order in electron - doped nd@xmath13ce@xmath14cuo@xmath8 persists to much larger @xmath15 ( @xmath16 ) @xcite , and coexists with superconductivity for even the highest @xmath7 (= 25 k ) material ( @xmath17 ) @xcite .
in contrast , superconductivity in la@xmath13sr@xmath14cuo@xmath8 emerges from a spin - glass regime and occurs over a wider doping concentration .
the close proximity of af order and superconductivity raises an interesting question concerning the role of long - range magnetic order in the superconductivity of copper oxides .
theoretically , it was predicted that , when an applied field creates vortices in these superconductors , af order would be induced in the core of each vortex @xcite . for underdoped superconducting la@xmath13sr@xmath18cuo@xmath8 ,
neutron scattering experiments show that a @xmath0-axis aligned magnetic field ( @xmath19-axis ) not only suppresses superconductivity but also enhances the static incommensurate spin density wave order , thus suggesting that such order competes directly with superconductivity @xcite .
although muon spin resonance ( @xmath20sr ) @xcite and nuclear magnetic resonance ( nmr ) experiments in underdoped yba@xmath5cu@xmath6o@xmath21 @xcite also suggest enhanced af order originating from regions near the vortex core , neutron scattering experiments failed to confirm any enhancement of the static long - range order in yba@xmath5cu@xmath6o@xmath22 for fields up to 7-t @xcite .
therefore , in spite of intensive effort @xcite , the nature of the superconductivity - suppressed ground state in high-@xmath7 superconductors is still unknown .
the major difficulty in studying the ground state of hole - doped high-@xmath7 superconductors is the enormous upper critical fields @xmath23 ( @xmath24-t ) required to completely suppress superconductivity . fortunately , electron - doped materials generally have @xmath23 , for magnetic fields aligned along the @xmath0-axis , less than 10-t @xcite , a value easily reachable in neutron scattering experiments . while recent experiments by matsuda _
_ @xcite found that a 10-t @xmath0-axis aligned field has no effect on the af order in the superconducting nd@xmath25ce@xmath26cuo@xmath8 , we showed that such fields in nd@xmath3ce@xmath4cuo@xmath8 enhance the af moment and induce a direct quantum phase transition from the superconducting state to an anomalously conducting antiferromagnetically ordered state at @xmath23 @xcite .
the induced af moments scale approximately linearly with the applied field , saturate at @xmath23 , and then decrease for higher fields , indicating that the field - induced af order competes directly with superconductivity @xcite .
although electron - doped nd@xmath13ce@xmath14cuo@xmath8 offers a unique opportunity for studying the superconductivity - suppressed ground state of high-@xmath7 copper oxides , the system is somewhat more complicated than hole - doped materials such as la@xmath13sr@xmath18cuo@xmath8 and yba@xmath5cu@xmath6o@xmath27 for three reasons .
first , it contains two magnetic ions ( rare - earth nd@xmath12 and cu@xmath10 ) , and the ordered cu sublattice induces the long - range af ordering of nd ions @xcite .
the effect of an applied field on rare - earth nd@xmath12 magnetic moments and their ordering is unknown .
second , for even the highest @xmath7 (= 25 k ) nd@xmath13ce@xmath14cuo@xmath8 ( @xmath17 ) , superconductivity coexists with the long - range residual af order , and the nature of their coexistence is unclear @xcite .
finally , superconductivity in nd@xmath13ce@xmath14cuo@xmath8 can only be achieved by annealing the as - grown samples at high temperatures @xcite .
the annealing process not only induces superconductivity in nd@xmath13ce@xmath14cuo@xmath8 , but also produces structural superlattice reflections of unknown origin @xcite . to understand the effect of a magnetic field on superconductivity in nd@xmath3ce@xmath4cuo@xmath8 @xcite
, one must first determine its influence on the residual af order without the complication of superconductivity .
since the residual af order in superconducting nd@xmath13ce@xmath14cuo@xmath8 has the same magnetic structure as that of the insulating nd@xmath5cuo@xmath8 at low temperatures @xcite , investigating the field effect on af orders in nd@xmath1cuo@xmath2 will resolve this issue .
second , the effect of ce doping in nd@xmath1cuo@xmath2 can be studied by performing magnetic field experiments in as - grown nonsuperconducting nd@xmath3ce@xmath4cuo@xmath8 . finally , to resolve the nature of the coexisting superconducting and af orders in nd@xmath13ce@xmath14cuo@xmath8 , we also need to understand the microscopic origin of the superlattice reflections and the effect of a magnetic field on these reflections .
this article describes experiments designed to understand the effect of a @xmath0-axis aligned magnetic field in all noncollinear spin structure phases of nd@xmath1cuo@xmath2 and in residual af order of as - grown nonsuperconducting and superconducting nd@xmath3ce@xmath4cuo@xmath8 . for nd@xmath1cuo@xmath2
, previous work showed that a magnetic field applied parallel to the cuo@xmath5 planes transforms the spins from the noncollinear to collinear af structure @xcite .
we find that a field applied perpendicular to the cuo@xmath5 planes only induces a canting of the af moment , and does not change the noncollinear nature of spin structures in all phases of nd@xmath1cuo@xmath2 .
for nonsuperconducting nd@xmath3ce@xmath4cuo@xmath8 , a 7-t @xmath0-axis aligned field does not enhance the af moment at low temperature .
finally , we confirm that the annealing process necessary to make superconducting nd@xmath3ce@xmath4cuo@xmath2 also induces epitaxial , three - dimensional ordered cubic ( nd , ce)@xmath5o@xmath6 ( space group @xmath28 , and lattice parameter @xmath29 ) as an impurity phase @xcite . in addition , the annealing process makes a series of quasi two - dimensional superlattice reflections associated with lattice distortions of nd@xmath3ce@xmath4cuo@xmath2 . while these quasi two - dimensional superlattice reflections have no field - induced effect , we determine the field - induced effect in the impurity phase and show that such effect can not account for the anisotropy in field - induced intensity between @xmath19-axis and @xmath30-plane . to further demonstrate that field - induced scattering is an intrinsic property of the superconductor , we probed bragg reflections exclusively from nd@xmath3ce@xmath4cuo@xmath2 by performing experiments using a horizontal field magnet .
the results confirm that af signal arises from the suppression of superconductivity by the @xmath0-axis aligned field in nd@xmath3ce@xmath4cuo@xmath8 @xcite .
our experiments were performed on the bt-2 and bt-9 triple - axis spectrometers at the nist center for neutron research and on the e4 two - axis diffractometer at the berlin neutron scattering center , hahn - meitner - institute ( hmi ) .
we measure momentum transfer @xmath31 in units of @xmath32 and specify the reciprocal space positions in reciprocal lattice units ( rlu ) @xmath33 appropriate for the tetragonal unit cells of nd@xmath1cuo@xmath2 ( space group @xmath34 , @xmath35 and @xmath36 ) and nd@xmath3ce@xmath4cuo@xmath8 ( space group @xmath34 , @xmath37 and @xmath38 ) , where @xmath39 and @xmath0 are in - plane and out - of - plane lattice parameters , respectively .
the schematic diagrams show the non - collinear spin structures in ( a ) type - i ( @xmath40 k ) and type - iii ( @xmath41 k ) phases , and ( b ) type - ii ( @xmath42 k ) phase .
the closed and open circles in ( c)-(e ) represent the scattering intensity at @xmath43 and 7-t , respectively , for fields aligned along the @xmath0-axis ( @xmath19-axis ) . ] for nist experiments , the collimations were , proceeding from the reactor to the detector , 40@xmath44 - 46@xmath44-sample-40@xmath44 - 80@xmath44 ( full - width at half - maximum ) , and the final neutron energy was fixed at @xmath45 mev .
the monochromator , analyzer and filters were all pyrolytic graphite .
we aligned the cuo@xmath5 planes in the horizontal @xmath46 $ ] scattering plane and applied the vertical magnetic field along the @xmath0-axis ( @xmath19-axis ) . in this geometry , we can access reciprocal space at any @xmath47 . to determine anisotropy of the field - induced effect , we also performed experiments in the @xmath48 scattering plane where the applied vertical fields are along the @xmath49 $ ] direction ( @xmath30-plane ) . for e4 measurements at hmi
, we used 40@xmath44 - 40@xmath44-sample-40@xmath44 collimation with fixed incident neutron energy of @xmath50 mev .
a pyrolytic graphite filter was placed in front of the sample to eliminate higher - order contamination .
the hm-2 4-t horizontal field magnet was used to apply a @xmath0-axis aligned field while probing the @xmath51 modulation of the scattering . although these measurements are crucial in determining the field - induced magnetic structure , the highly restricted access angles of the magnet limit the regions of reciprocal space that can be probed . in horizontal field measurements on nd@xmath3ce@xmath4cuo@xmath8 ,
the crystal was aligned in the @xmath48 zone and the applied field was along the @xmath0-axis .
we grew a single crystal of nd@xmath1cuo@xmath2 ( @xmath52 mm ) and crystals of nd@xmath3ce@xmath4cuo@xmath8 using the traveling solvent floating zone technique @xcite .
the nd@xmath1cuo@xmath2 crystal used in the experiments is as - grown .
we also performed experiments on as - grown nonsuperconducting nd@xmath3ce@xmath4cuo@xmath8 and superconducting nd@xmath3ce@xmath4cuo@xmath8 .
superconductivity in nd@xmath3ce@xmath4cuo@xmath8 was obtained after annealing the samples in a flowing ar / o@xmath5 gas mixture with a partial oxygen pressure of @xmath53 atm at 1000@xmath54c for 100 h. magnetic susceptibility measured on small pieces of crystals ( @xmath55 mg ) cut from the samples used for neutron experiments show the onset of bulk superconductivity at @xmath5625 k with a transition width of 3 k. with a @xmath0-axis aligned field of 4 oe , the zero - field cooled data show complete screening of flux . in the field
cooled case , the crystal expels 18% of the flux , indicating that the bulk superconductivity has at least 18% of the volume fraction .
the susceptibility of nd@xmath13ce@xmath18cuo@xmath2 in the cuo@xmath5 planes is several times bigger than that perpendicular to them .
the large magnetic anisotropy means that a @xmath0-axis aligned field acting on the magnetic moments ( nd and cu ) produces a large torque on the sample . to prevent the samples from rotating under the influence of a @xmath19-axis field , they were clamped on solid aluminum brackets . for experiments at nist
, the bracket was inserted inside a he filled aluminum can mounted on a standard 7-t split - coil superconducting magnet .
for hmi experiments , the sample assembly was mounted on a mini - goniometer and inserted directly to the sample chamber of the hm2 magnet . in an attempt to determine the homogeneity of the annealed superconducting and as - grown nonsuperconducting nd@xmath3ce@xmath4cuo@xmath8 , we performed neutron diffraction measurements on both samples .
the results confirm that superlattice reflections exist only in superconducting samples @xcite , and are resolution limited ( indicating a correlation length larger than 300 ) in the cuo@xmath5 plane and relatively broad along the @xmath0-axis ( see sections iv - v ) .
although the sharpness of the superlattice reflections and susceptibility measurements suggest that the crystal is homogeneous with bulk superconductivity , it is not clear how the residual af order , superstructure , and superconductivity coexist microscopically in the material .
before describing the field effect on the long - range magnetic order of nd@xmath1cuo@xmath2 , we briefly review its zero - field behavior . as shown in refs .
@xcite , the cu spins in nd@xmath1cuo@xmath2 first order into the noncollinear type - i spin structure below @xmath57 k ( fig .
1a ) . on further cooling ,
the cu spins re - orient into type - ii ( at @xmath58 k ) and type - iii ( @xmath59 k ) phases . in the type - ii phase ( fig .
1b ) , all the cu spins rotate by 90@xmath54 about the @xmath0-axis from the type - i phase .
they rotate back to their original direction below @xmath60 in the type - iii phase ( fig .
1a ) . the closed circles in figs .
1c , 1d , and 1e show the temperature dependence of the scattering at @xmath61 , @xmath62 , and @xmath63 , respectively .
clear af phase transitions are seen at @xmath64 , @xmath65 , and @xmath60 as marked by the arrows in fig .
1c , confirming previous work @xcite .
the large intensity increase of the @xmath61 peak below @xmath1120 k is associated with staggered moments on nd sites induced by cu - nd coupling .
magnetic structure factor calculations indicate that the @xmath63 reflection has vanishing intensity in the type - i / iii phases and becomes finite in the type - ii phase .
the large intensity jumps of @xmath63 at @xmath65 and @xmath60 shown in fig .
1e clearly bear this out .
since nd@xmath1cuo@xmath2 only has af phase transitions at zero - field , the intensity of the nuclear bragg peak @xmath62 has no magnetic contributions and hence is essentially temperature independent ( fig .
1d ) . to estimate the magnetic moments of nd and cu in different phases of nd@xmath1cuo@xmath2
, we normalized the intensity of af peaks at @xmath63 and @xmath61 to that of the weak @xmath62 or strong @xmath66 nuclear bragg peak .
the estimated nd and cu moments differ dramatically depending on the chosen bragg peaks , dure primarily to extinction but also nd absorption of the large crystal ( table i ) .
in particular , the intensities of the strong peaks such as @xmath66 are severely extinction limited , which overestimates the magnitude of the ordered moment .
.the magnitude of magnetic moments calculated by normalizing the af intensity at @xmath63 and @xmath61 to that of the weak @xmath62 or strong @xmath66 nuclear bragg reflection . from the powder diffraction measurements on nd@xmath5cuo@xmath8 @xcite
, the ordered cu moment was estimated to be 0.46 @xmath67 at 80 k and nd moment was 0.46 @xmath67 at 5 k. in computing the cu and nd moments , we assumed that the nd moment does not contribute to magnetic scattering above 50 k and the cu moment does not change below 50 k. [ cols="^,^,^,^,^",options="header " , ] to estimate the fractional volume of the cubic ( nd , ce)@xmath5o@xmath6 in our superconducting nd@xmath3ce@xmath4cuo@xmath8 , we aligned the crystal in the @xmath68 and @xmath48 zones and performed @xmath0-axis scans along the @xmath69 $ ] and @xmath70 $ ] directions , respectively , at room temperature .
figure 7 plots the outcome of the experiment . along the @xmath69 $ ] direction ( fig .
7a ) , sharp resolution - limited bragg peaks corresponding to the cubic @xmath71 and @xmath72 reflections are observed at @xmath73 and @xmath74 in the tetragonal miller indexes of nd@xmath3ce@xmath4cuo@xmath8 , respectively ( table ii ) . along the @xmath70 $ ] direction , the cubic ( nd ,
ce)@xmath5o@xmath6 @xmath75 , @xmath76 , and @xmath77 peaks are observed at the expected places ( fig .
the observation of sharp bragg peaks from ( nd , ce)@xmath5o@xmath6 along the @xmath0-axis and in the cuo@xmath5 plane indicates that cubic ( nd , ce)@xmath5o@xmath6 forms three - dimensional long - range order in the matrix of nd@xmath3ce@xmath4cuo@xmath8 . by comparing the large ( nd , ce)@xmath5o@xmath6 @xmath78 bragg peak ( @xmath118700 counts / minute ) with the very weak @xmath79 ( @xmath1121160 counts / minute ) reflection of nd@xmath3ce@xmath4cuo@xmath8
, we estimate that ( nd , ce)@xmath5o@xmath6 has a volume fraction 1.0@xmath80 . alternatively ,
if we use the very strong @xmath66 ( @xmath111.08@xmath81 counts / minute ) reflection of nd@xmath3ce@xmath4cuo@xmath8 , we find a volume fraction of 2.0@xmath82 for ( nd , ce)@xmath5o@xmath6 .
however , the very strong fundamental peaks are severely extinction limited , and this overestimates the ( nd , ce)@xmath5o@xmath6 volume fraction ( just as the ordered cu moment is overestimated in table i using the @xmath66 reflection ) .
therefore , the estimate using the weak nd@xmath3ce@xmath4cuo@xmath8 structural peaks is more reliable . for the cubic ( nd , ce)@xmath5o@xmath6 with the @xmath28 space group symmetry , structure factor calculations show vanishing intensity at @xmath83 and equivalent positions . although the absence of a sharp bragg peak along the @xmath69 $ ] direction at @xmath84 confirms the structure factor calculation , the @xmath83 peak that is sharp along the @xmath85 $ ] direction ( see fig .
@xcite and fig .
6a , section v ) but diffusive along the @xmath69 $ ] direction ( fig .
7a ) is reminiscent of the superlattice reflection seen at @xmath63 @xcite . since these diffuse superlattice reflections at ( 1/2,0,0 ) and ( 1/2,1/2,0 )
have no magnetic field dependence at 5 k and are not related to the cubic ( nd , ce)@xmath5o@xmath6 , they must be associated with the formation of a quasi two - dimensional lattice distortion necessary for nd@xmath3ce@xmath4cuo@xmath8 to become superconducting .
work is currently underway to determine the microscopic origin of the lattice distortion .
o@xmath6 impurity phase and @xmath0-axis modulation of the structural superlattice reflections in superconducting nd@xmath3ce@xmath4cuo@xmath8 .
( a ) @xmath51-scan along the @xmath69 $ ] direction .
in addition to the well - marked cubic ( nd , ce)@xmath5o@xmath6 peaks , a broad diffusive peak with a full - width at half maximum of @xmath86 rlu is observed at @xmath87 .
such diffusive peak is similar to the superlattice reflections reported in ref .
@xcite at @xmath63 .
we also note that part of the peak intensity at @xmath87 arises from the small detector angles and the neutron absorption of the long nd@xmath3ce@xmath4cuo@xmath8 crystal .
( b ) @xmath51-scan along the @xmath70 $ ] direction with ( nd , ce)@xmath5o@xmath6 peaks marked by arrows . since no broad diffusive peak is found at @xmath87 , most of the intensity at @xmath88 in fig .
6e is due to ( nd , ce)@xmath5o@xmath6 . ]
the identification of the epitaxial cubic ( nd , ce)@xmath5o@xmath6 with lattice parameter close to that of the superconducting nd@xmath3ce@xmath4cuo@xmath8 raises the important question concerning the possible role of this impurity phase in the observed field - induced effect @xcite , as the rare - earth magnetic ion nd@xmath12 in ( nd , ce)@xmath5o@xmath6 will be polarized by the applied field . in general , the rare - earth oxides such as nd@xmath5o@xmath6 and er@xmath5o@xmath6 have the bixbyite structure with 32 rare - earth ions in a cubic unit cell and order antiferromagnetically at low temperature @xcite . since scattering at all half integer positions [ except @xmath61 from nd@xmath3ce@xmath4cuo@xmath8 ] is temperature independent above 5 k @xcite ,
it is safe to assume that the ( nd , ce)@xmath5o@xmath6 impurity is in the paramagnetic state at this temperature . in the paramagnetic state of ( nd , ce)@xmath5o@xmath6 ,
a field will induce a net moment given by a brillouin function , and the field - induced moment should saturate in the high field limit .
this is in clear contrast to our observation where the scattering first increases with field , and then decreases at higher fields at 5 k @xcite . of course , at sufficiently low temperatures
where ( nd , ce)@xmath5o@xmath6 and/or nd in nd@xmath3ce@xmath4cuo@xmath8 ( @xmath90 k ) spontaneously order @xcite , an applied field will rotate the ordered af moment along the field direction and therefore suppress the af intensity .
the results we report in ref . @xcite carefully avoided the regime of spontaneous magnetic order for both ( nd , ce)@xmath5o@xmath6 and superconducting nd@xmath3ce@xmath4cuo@xmath8 @xcite by measuring spectra above 5 k. since the cubic ( nd , ce)@xmath5o@xmath6 impurity phase has almost the same lattice parameter as nd@xmath3ce@xmath4cuo@xmath8 in the @xmath39-@xmath91 plane , measurements at @xmath87 could be ambiguous as the scattering could originate from either ( nd , ce)@xmath5o@xmath6 or nd@xmath3ce@xmath4cuo@xmath8 . the experimental resolution of this ambiguity is straightforward , measurements simply need to be made at finite @xmath51 , where the nd@xmath3ce@xmath4cuo@xmath8 peaks are not coincident with ( nd , ce)@xmath5o@xmath6 . to accomplish this , we aligned the crystal in the @xmath48 zone inside the hm2 4-t horizontal field magnet at hmi . in this geometry
, we can probe the @xmath51-dependence of the scattering while keeping the field along the @xmath0-axis .
figure 8 summarizes the outcome of the experiment . at zero - field and 5 k ,
the @xmath92 $ ] scan shows well - defined peaks associated with the residual af order of nd@xmath3ce@xmath4cuo@xmath8 at @xmath93 and @xmath94 .
in addition , we find the @xmath95 reflection of the cubic ( nd , ce)@xmath5o@xmath6 and the @xmath96 powder peak of the aluminum sample holder ( fig .
8a and table ii ) .
when @xmath0-axis aligned fields are applied , the residual af @xmath93 peak enhances systematically with increasing field ( fig .
8b ) while the ( nd , ce)@xmath5o@xmath6 @xmath95 ( figs .
8a and c ) and the aluminum @xmath96 reflections ( fig .
8a ) are not affected .
figures 8d and e show the @xmath97 $ ] scan and temperature dependence of the scattering at the @xmath93 , respectively .
clear field - induced enhancements are observed below @xmath7 , consistent with the data in the @xmath47 plane @xcite .
we note that a 2-t field parallel to the cuo@xmath5 along the @xmath49 $ ] direction induces a spin - flop transition and suppresses the intensity at @xmath93 ( see fig . 4 in ref .
therefore , the @xmath93 peak shows an induced af component when the field is along the @xmath0-axis and superconductivity is strongly suppressed , but not when it is in the @xmath39-@xmath91 plane and superconductivity is only weakly affected @xcite .
we also note that the qualitatively different behavior observed for @xmath30-plane versus @xmath19-axis for @xmath93 directly violates the cubic symmetry of ( nd , ce)@xmath5o@xmath6 .
-axis field on the integrated intensity of af bragg reflections and the epitaxy nd@xmath5o@xmath6 at @xmath98 k in the @xmath48 scattering plane using the hm-2 horizontal field magnet at hmi .
the filled circles represent 0-t data while the open circles are identical scans at 4-t .
( a ) the @xmath92 $ ] scan at 0-t and 4-t with superconducting nd@xmath3ce@xmath4cuo@xmath8 , ( nd , ce)@xmath5o@xmath6 , and aluminum peaks marked by the arrows .
( b ) detailed scans along the @xmath92 $ ] direction around @xmath93 reflection at various fields .
( c ) detailed scans around the cubic ( nd , ce)@xmath5o@xmath6 @xmath95 peak at 0-t and 4-t . there is no observable field - induced effect at 4 t. ( d ) the @xmath97 $ ] scan around @xmath93 reflection at 0 t and 4 t. ( e ) the temperature dependence of the scattering at @xmath93 at 0-t and 4-t .
the solid and dotted lines in ( b ) and ( d ) are gaussian fits . ]
figure 9 compares the temperature and field dependence of the field - induced scattering at @xmath61 ( figs .
9a and c ) and @xmath93 ( figs .
9b and d ) .
the remarkable similarity of the field response in these reflections suggests that they must originate from the same physical process . considering that a 7-t @xmath0-axis aligned field has no effect on @xmath61 in as - grown nonsuperconducting nd@xmath3ce@xmath4cuo@xmath8 ( see section iv )
, we conclude that field - induced af order at @xmath61 and @xmath93 in the figure can only result from the suppression of superconductivity .
furthermore , the new data indicate that the field - induced enhancement forms three - dimensional long - range af order in superconducting nd@xmath3ce@xmath4cuo@xmath8 .
ce@xmath4cuo@xmath8 at @xmath61 and @xmath93 . while the data at @xmath61 are from ref .
@xcite , the results at @xmath93 are new .
the temperature dependence of the difference between 4 t and 0 t at ( a ) @xmath61 and ( b ) @xmath93 .
the field dependence of the integrated intensity at ( c ) @xmath61 and ( d ) @xmath93 . ]
although our horizontal field measurements conclusively demonstrate that a @xmath0-axis aligned magnetic field enhances the residual af order in superconducting nd@xmath3ce@xmath4cuo@xmath8 , it is also important to determine the field - induced effect of the ( nd , ce)@xmath5o@xmath6 impurity phase . because ( nd , ce)@xmath5o@xmath6 forms a three - dimensional long - range ordered cubic lattice ( fig .
7 ) , its field - induced effect should be isotropic for fields along ( nd , ce)@xmath5o@xmath6 @xmath99 $ ] and @xmath100 $ ] directions . in the notation of nd@xmath3ce@xmath4cuo@xmath8 miller indexes ( table ii ) , these are along @xmath49 $ ] ( @xmath30-plane ) and @xmath101 $ ] ( @xmath19-axis ) directions , respectively . as superconductivity
is strongly suppressed for a @xmath19-axis field but much less affected by the same field in the @xmath102-plane , measurements of field directional anisotropy will establish the influence of ( nd , ce)@xmath5o@xmath6 to the observed field effect in nd@xmath3ce@xmath4cuo@xmath8 @xcite .
figure 10 summaries the outcome of such experiment on bt-2 using the @xmath103 crystal of superconducting nd@xmath3ce@xmath4cuo@xmath8 in two different field geometries .
we first describe measurements in the @xmath48 zone , where the applied vertical field is along the @xmath49 $ ] direction of nd@xmath3ce@xmath4cuo@xmath8 and @xmath99 $ ] direction of ( nd , ce)@xmath5o@xmath6 . in this geometry
, we can probe the field - induced effect on @xmath63 and @xmath104 without much affecting the superconductivity .
while scattering at @xmath63 may originate from either nd@xmath3ce@xmath4cuo@xmath8 or ( nd , ce)@xmath5o@xmath6 , @xmath104 is exclusively associated with the @xmath105 reflection of ( nd , ce)@xmath5o@xmath6 ( table ii ) . figures 10a and b show the outcome of the experiment along the @xmath106 $ ] and @xmath107 $ ] directions at 5 k. clear field - induced effects are seen at @xmath104 ( fig .
10b ) , indicating that ( nd , ce)@xmath5o@xmath6 can indeed be polarized by the applied field .
similar measurements on @xmath104 with field aligned along the @xmath49 $ ] direction of ( nd , ce)@xmath5o@xmath6 show a weaker field - induced effect , thus suggesting that its easy axis is along the @xmath99 $ ] direction .
if we ignore the contribution of the superlattice structure to @xmath63 and assume that the scattering there is due entirely to ( nd , ce)@xmath5o@xmath6 ( table ii ) , its integrated intensity should be identical to that at @xmath104 because of the cubic symmetry of ( nd , ce)@xmath5o@xmath6 . by normalizing the zero field intensity at @xmath104 to that at @xmath63
, we can compare the field - induced effect at these two equivalent positions for ( nd , ce)@xmath5o@xmath6 . since the field - induced effect at @xmath63 ( fig .
10a ) is only about 25% larger that at @xmath104 ( fig .
10b ) , we conclude that the field - induced intensity at @xmath63 is mostly due to the polarization of ( nd , ce)@xmath5o@xmath6 in this field geometry .
ce@xmath4cuo@xmath8 at @xmath63 and @xmath104 in the @xmath30-plane geometry with that at @xmath63 and @xmath108 in the @xmath19-axis geometry .
scans through ( a ) @xmath63 and ( b ) @xmath104 at 5 k for field along the @xmath49 $ ] axis . since @xmath104 from ( nd
, ce)@xmath5o@xmath6 is very close to the strong @xmath105 nuclear bragg peak from nd@xmath3ce@xmath4cuo@xmath8 , the radial scan in the inset of ( b ) shows sloped background .
we determine the background scattering at @xmath104 by fitting a gaussian with fixed width to @xmath105 ( dashed line in the inset ) .
this is confirmed by the transverse scan across @xmath104 .
similar scans through ( c ) @xmath63 and ( d ) @xmath108 for field along the @xmath101 $ ] direction .
the ratios of integrated intensities of @xmath104 and @xmath63 between 6.9 t and 0 t ( @xmath109(6.9 t)@xmath110(0 t ) ) are 2.7 and 3.3 , respectively , for the @xmath30-plane field . in comparison ,
the ratio is 5.7 for both @xmath63 and @xmath108 for the @xmath0-axis aligned field . in ref . @xcite and fig .
6 , the ratios of @xmath109(7 t)@xmath110(0 t ) at @xmath63 are 6.8 and 5.5 , respectively , in the @xmath19-axis geometry . ] in figures 10c and d , we plot the field - induced effect for the @xmath19-axis experiment at two equivalent reflections @xmath63 and @xmath108 . while the magnitude of field - induced effect is consistent with ref .
@xcite , they are twice as large as that of figs . 10a and b. since the cubic symmetry of ( nd , ce)@xmath5o@xmath6 requires the same induced effect for fields along @xmath99 $ ] and @xmath100 $ ] , the observation of a much larger field - induced effect in @xmath19-axis geometry means the excess field - induced intensity must originate from the suppression of superconductivity .
finally , we remark that one concern raised @xcite was that the finite @xmath51 behavior we observe might not be intrinsic to superconducting nd@xmath3ce@xmath4cuo@xmath8 , but rather it is somehow induced by the magnetic coupling to the ( nd , ce)@xmath5o@xmath6 impurity phase . however , ( nd , ce)@xmath5o@xmath6 has a very weak exchange interaction and thus orders at very low temperature .
it is difficult to see how ( nd , ce)@xmath5o@xmath6 could dominate the nd@xmath3ce@xmath4cuo@xmath8 physics , particularly when it only constitutes a small ( @xmath1110@xmath111 ) volume fraction of the crystals in superconducting nd@xmath3ce@xmath4cuo@xmath8 . on the other hand ,
the nd magnetic structure in nd@xmath3ce@xmath4cuo@xmath8 has the same symmetry as the cu spin configuration and thus is maximally coupled to the cu spins @xcite . even in this case
, the perturbation of the nd order parameter by the cu spins is small .
therefore , while the ( nd , ce)@xmath5o@xmath6 ordering may be induced by being in contact with bulk nd@xmath3ce@xmath4cuo@xmath8 , it is highly improbable that the field - induced magnetic scattering at @xmath93 could be induced by the ( nd , ce)@xmath5o@xmath6 impurity .
we have investigated the effect of a @xmath19-axis field in all phases of nd@xmath1cuo@xmath2 and in as - grown nonsuperconducting and superconducting nd@xmath3ce@xmath4cuo@xmath8 . at zero - field ,
cu spins in nd@xmath1cuo@xmath2 form noncollinear structures because of the coupling between cu@xmath10 and nd@xmath12 .
such a magnetic interaction also creates a small in - plane spin - wave gap @xmath112 at @xmath113 . for a magnetic field
aligned parallel to the cuo@xmath5 plane , cu spins transform from a noncollinear to collinear structure in a spin - flop phase transition with a critical field less than 2-t @xcite .
such a spin - flop transition occurs because when the magnetic field associated with the zeeman energy ( @xmath114 ) equals to @xmath112 , the net magnetic exchange interaction vanishes and with it the noncollinear spin structure @xcite . for a 7-t @xmath19-axis field ,
our data clearly indicate that the noncollinear af spin structures in nd@xmath1cuo@xmath2 are essentially unaffected for temperatures above 5 k. as a consequence , the zero - field in - plane spin - wave gap @xmath112 and the magnetic exchange interaction must also remain unchanged in the field .
the large increase in the cu(nd ) fm moments suggests that the applied @xmath0-axis field only induces a canting of the af order .
these results contrast significantly with that of superconducting nd@xmath3ce@xmath4cuo@xmath8 , where the applied field induces a static , long - range ordered af state @xcite .
we demonstrate that the annealing process necessary for superconductivity in nd@xmath3ce@xmath4cuo@xmath8 also induces structural superlattice reflections at @xmath63 and @xmath83 positions .
in addition , we confirm the presence of the cubic ( nd , ce)@xmath5o@xmath6 as an impurity phase in superconducting nd@xmath3ce@xmath4cuo@xmath8 following the annealing process @xcite .
although the lattice parameter of the cubic ( nd , ce)@xmath5o@xmath6 is very close to the in - plane lattice parameter of nd@xmath3ce@xmath4cuo@xmath8 , most of the structural superlattice reflections in the @xmath47 plane are quasi two - dimensional and can not be associated with the three - dimensional cubic ( nd , ce)@xmath5o@xmath6 . by probing the @xmath51-dependence of the scattering with a @xmath0-axis aligned field ,
we show that the residual af order in superconducting nd@xmath3ce@xmath4cuo@xmath8 enhances with increasing field .
such behavior is different from the field effect on as - grown nonsuperconducting nd@xmath3ce@xmath4cuo@xmath8 , where the long - range noncollinear af order is essentially unaffected by a 7-t @xmath0-axis aligned field . by studying the anisotropy of field - induced effect ,
we determine the effect of the magnetic field on the cubic ( nd , ce)@xmath5o@xmath6 and confirm that the results with the @xmath0-axis field in the @xmath46 $ ] plane are inconsistent with the impurity phase . combining these results with horizontal field experiments , we conclude that af order is induced in nd@xmath3ce@xmath4cuo@xmath8 upon suppression of superconductivity by a @xmath0-axis aligned magnetic field .
we are grateful to y. ando , henry fu , s. a. kivelson , d .- h .
lee , d. mandrus , h. a. mook , and s .- c .
zhang for helpful conversations .
we also thank s. larochelle and p. k. mang for initially alerting us to the existence of the secondary nd@xmath5o@xmath6 phase .
this work was supported by u.s .
nsf dmr-0139882 and doe under contract no .
de - ac05 - 00or22725 .
e - mail : [email protected] j. g. bednorz and k. a. muller , z. phys .
b * 64 * , 189 ( 1987 ) .
y. tokura , h. takagi , and s. uchida , nature ( london ) * 337 * , 345 ( 1989 ) .
h. takagi , s. uchida , and y. tokura , phys .
lett . * 62 * , 1197 ( 1989 ) .
d. vaknin , s. k. sinha , d. e. moncton , d. c. johnston , j. m. newsam , c. r. safinya , and h. e. king , phys .
lett . * 58 * , 2802 ( 1987 ) .
j. m. tranquada , d. e. cox , w. kunnmann , h. moudden , g. shirane , m. suenaga , p. zolliker , d. vaknin , s. k. sinha , m. s. alvarez , a. j. jacobson , and d. c. jonhston , phys .
* 60 * , 156 ( 1988 ) .
s. skanthakumar , h. zhang , t. w. clinton , w - h .
li , j. w. lynn , z. fisk , and s - w .
cheong , physica * c 160 * , 124 ( 1989 ) . j. w. lynn and s. skanthakumar , in _
handbook on the physics and chemistry of rare earths _ , edited by k. a. gschneidner , jr .
, l. eyring , and m. b. maple , vol 31 ( 2001 elservier science b. v. ) p 315 .
s. skanthakumar , j. w. lynn , j. l. peng , and z. y. li , j. appl . phys .
* 73 * , 6326 ( 1993 ) .
s. skanthakumar , j. w. lynn , j. l. peng , and y. z. li , phys . rev .
b * 47 * , 6173 ( 1993 ) .
r. sachidanandam , t. yilidirim , and a. b. harris , phys .
b * 56 * , 260 ( 1997 ) .
d. petitgrand , s. v. maleyev , ph .
bourges , and a. s. ivanov , phys .
b * 59 * , 1079 ( 1999 ) .
k. yamada , k. kurahashi , y. endoh , r. j. birgeneau , and g. shirane , j. phys . and
solids * 60 * , 1025 ( 1999 ) .
t. uefuji , t. kubo , k. yamada , m. fujita , k. kurahashi , i. watanabe , and k. nagamine , physica c * 357 - 360 * , 208 ( 2001 ) ; t. uefuji , k. kurahashi , m. fujita , m. matsuda , and k. yamada , physica c * 378 - 381 * , 273 ( 2002 ) .
s. c. zhang , science * 275 * , 1089 ( 1997 ) .
d. arovas , a. j. berlinsky , c. kallin , and s. c. zhang , phys .
* 79 * , 2871 ( 1997 ) .
s. katano , m. sato , k. yamada , t. suzuki , and t. fukase , phys .
b * 62 * , r14677 ( 2000 ) . b. lake , g. aeppli , k. n. clausen , d. f. mcmorrow , k. lefmann , n. e. hussey , n. mangkorntong , n. nohara , h. takagi , t. e. mason , science * 291 * , 1759 ( 2001 ) . b. lake , h. m. ronnow , n. b. christensen , g. aeppli , k. lefmann , d. f. mcmorrow , p. vorderwisch , p. smeibidl , n. mangkorntong , t. sasagawa , m. nohara , h. takagi , and t. e. mason , nature ( london ) * 415 * , 299 ( 2002 ) .
b. khaykovich , y. s. lee , r. erwin , s .- h .
lee , s. wakimoto , k. j. thomas , m. a. kastner , and r. j. birgeneau , phys .
b * 66 * , 014528 ( 2002 ) .
r. i. miller , r. f. kiefl , j. h. brewer , j. e. sonier , j. chakhalian , s. dunsiger , g. d. morris , a. n. price , d. a. bonn , w. h. hardy , and r. liang , phys .
* 88 * , 137002 ( 2002 ) . v. f. mitrovic , e. e. sigmund , m. eschrig , h. n. bachman , w. p. halperin , a. p. reyes , p. kuhns , and w. g. moulton , nature ( london ) * 413 * , 501 ( 2001 ) .
p. dai , h. a. mook , g. aeppli , s. m. hayden , and f. do@xmath115an , nature ( london ) * 406 * , 965 ( 2000 ) . h. a. mook , p. dai , s. m. hayden , a. hiess , j. w. lynn , s .- h .
lee , and f. do@xmath115an , phys .
b * 66 * , 144513 ( 2002 ) .
y. hidaka and m. suzuki , nature ( london ) * 338 * , 635 ( 1989 ) .
p. fournier , p. mohanty , e. maiser , s. darzens , t. venkatesan , c. j. lobb , g. czjzek , r. a. webb , and r. l. greene , phys .
* 81 * , 4720 ( 1998 ) .
r. w. hill , c. proust , l. taillefer , p. fournier , and r. l. greene , nature * 414 * , 711 ( 2001 ) .
y. wang , y. ono , y. onose , g. gu , y. ando , y. tokura , s. uchida , and n. p. ong , science * 299 * , 86 ( 2003 ) .
m. matsuda , s. katano , t. uefuji , m. fujita , and k. yamada , phys .
b * 66 * , 172509 ( 2002 ) . h. j. kang , p. dai , j. w. lynn , m. matsuura , j. r. thompson , s .- c .
d. n. argyriou , y. onose , and y. tokura , nature ( london ) * 423 * , 522 ( 2003 ) .
j. w. lynn , s. skanthakumar , i. w. sumarlin , w - h .
li , r. n. shelton , j. l. peng , z. fisk , and s - w .
cheong , phys .
b * 41 * , 2569 ( 1990 ) . y. onose , y. taguchi , k. ishizaka , and y. tokura , phys .
lett . * 87 * , 217001 ( 2001 ) .
k. kurahashi , h. matsushita , m. fujita , and k. yamada , j. phys .
. jpn . * 71 * , 910 ( 2002 ) .
p. k. mang , s. larochelle , and m. greven , unpublished .
von g. brauer and h. gradinger , z. anorg .
allgem . chem . *
276 * , 209 ( 1954 ) .
r. m. moon , w. c. koehler , h. r. child , and l. j. raubenheimer , phys .
* 176 * , 722 ( 1968 ) .
m. greven ( private communication ) . | we have measured the effect of a @xmath0-axis aligned magnetic field on the long - range magnetic order of insulating nd@xmath1cuo@xmath2 , as - grown nonsuperconducting and superconducting nd@xmath3ce@xmath4cuo@xmath2 .
on cooling from room temperature , nd@xmath1cuo@xmath2 goes through a series of antiferromagnetic ( af ) phase transitions with different noncollinear spin structures . in all phases of nd@xmath1cuo@xmath2
, we find that the applied @xmath0-axis field induces a canting of the af order but does not alter the basic zero - field noncollinear spin structures .
similar behavior is also found in as - grown nonsuperconducting nd@xmath3ce@xmath4cuo@xmath2 .
these results contrast dramatically with those of superconducting nd@xmath3ce@xmath4cuo@xmath2 , where a @xmath0-axis aligned magnetic field induces a static , anomalously conducting , long - range ordered af state .
we confirm that the annealing process necessary to make superconducting nd@xmath3ce@xmath4cuo@xmath2 also induces epitaxial , three - dimensional long - range ordered cubic ( nd , ce)@xmath5o@xmath6 as an impurity phase . in addition , the annealing process makes a series of quasi two - dimensional superlattice reflections associated with lattice distortions of nd@xmath3ce@xmath4cuo@xmath2 in the cuo@xmath5 plane .
while the application of a magnetic field will induce a net moment in the impurity phase , we determine its magnitude and eliminate this as a possibility for the observed magnetic field - induced effect in superconducting nd@xmath3ce@xmath4cuo@xmath2 . |
given @xmath10 , a sequence @xmath11 of zeros and ones is called a _ @xmath12-expansion _ of @xmath0 , if @xmath13 clearly , a real number @xmath0 has a @xmath12-expansion if and only if @xmath14 $ ] .
non - integer base expansions pioneered by rnyi @xcite and parry @xcite obtained great attention from different branches of mathematics , such as number theory , dynamical system , measure theory , combinatorics , et al . in 1990s erds and jo @xcite discovered that there exist infinitely many reals having a continuum of expansions , and later sidorov @xcite showed that this property is generic which turns out to be quite different from integer base expansions . surprisingly , erds et al .
@xcite also showed that there exist infinitely many reals having a unique expansion .
after that there are many works devoted to the investigations of unique expansions ( cf .
@xcite ) . on the other hand ,
let @xmath15 be the set of _ univoque bases _
@xmath10 such that @xmath16 has a unique @xmath12-expansion .
erds et al .
@xcite showed that @xmath15 is a lebesgue null set and of first category .
later , darczy and ktai @xcite proved that @xmath15 has full hausdorff dimension .
recently , komornik and loreti @xcite investigated the topological properties of @xmath15 and showed that its closure @xmath17 is a cantor set . in general ,
for a real number @xmath18 we consider the set @xmath19 of univoque bases @xmath10 such that @xmath0 has a unique @xmath12-expansion , i.e. , @xmath20 clearly , for @xmath21 we have @xmath22 since @xmath23 always has a unique @xmath12-expansion @xmath24 for each @xmath10 . avoiding this trivial case we will assume @xmath25 throughout the paper .
l et al .
@xcite showed that for any @xmath26 the set @xmath19 is a lebesgue null set but has full hausdorff dimension .
when @xmath27 , komornik and loreti @xcite considered the smallest base of @xmath28 , denoted by @xmath5 , which is called the _ komornik - loreti constant _ in @xcite .
later , allouche and cosnard @xcite showed that @xmath5 is a transcendental number .
in this paper we consider the infinimum base @xmath8 of @xmath19 , i.e. , @xmath29 then by @xcite we have @xmath30 . recall that @xmath31 is the golden ratio .
now we state our main results for @xmath8 .
[ t11 ] let @xmath25 .
then @xmath4 if and only if @xmath32 furthermore , @xmath6 if and only if @xmath33 in the following theorem we show that @xmath8 is indeed the smallest base of @xmath19 when @xmath4 . [ t12 ] if @xmath4 , then @xmath34 .
we point out that in theorem [ t52 ] we determine the explicit value of @xmath8 when @xmath9 ( see figure [ fig:1 ] for the graph of @xmath8 with @xmath35 $ ] ) . with @xmath36\approx[1.0507 , 2]$].,title="fig:",width=377 ] +
the rest of the paper is arranged as follows . in section [ sec : unique expansions ] we recall some properties of unique expansions . the proof of theorem [ t11 ] will be given respectively in section [ sec : proof of thm11 ] for the case @xmath37 and in section [ sec : proof of thm12 ] for the case @xmath26 . in section [ sec : algorithm ] we determine the explicit value of @xmath8 when @xmath9 , and prove theorem [ t12 ] .
finally , we end the paper with some questions .
in this section we recall some results of unique expansions . for @xmath10 let @xmath38 be the set of @xmath39 having a unique @xmath12-expansion , and let @xmath40 be the set of corresponding @xmath12-expansions .
recall from @xcite that @xmath41 is the classical thue - morse sequence beginning with @xmath42 [ p21 ] the classical thue - morse sequence @xmath41 satisfies @xmath43 for @xmath44 let @xmath45 be the appropriate root of the equation @xmath46 then @xmath47 , @xmath48 , @xmath49 , et al . , and @xmath50 strictly increases to @xmath5 as @xmath51 . for @xmath10 we denote by @xmath52 the quasi - greedy @xmath12-expansion of @xmath16 , i.e. , the lexicographically largest @xmath12-expansion of @xmath16 with infinitely many non - zero elements ( cf .
@xcite ) . then by ( [ e21 ] ) and proposition [ p21 ] one can verify that for any @xmath53 we have @xmath54 here , for a word @xmath55 we denote by @xmath56 if @xmath57 , and we denote by @xmath58 if @xmath59 . furthermore , @xmath60 stands for the _ reflection _ of @xmath61 . for a positive integer @xmath53 we denote by @xmath62 the concatenations of @xmath61 to itself @xmath63 times , and by @xmath64 the concatenations of @xmath61 to itself infinitely many times . the following description of the set @xmath40 with @xmath65 was essentially established by glendinning and sidorov @xcite . [ p22 ]
let @xmath66 $ ] with @xmath67 .
then @xmath40 contains all sequences of the form @xmath68 and their reflections , where @xmath69 stands for all possible non - negative integers . by using proposition [ p22 ]
we can write down the sets @xmath40 for @xmath66 $ ] and @xmath70 . *
if @xmath71=(1,(1+\sqrt{5})/2]$ ] , then by proposition [ p22 ] we have @xmath72 * if @xmath73 $ ] , then by proposition [ p22 ] it gives that @xmath74 * if @xmath75 $ ] , then by proposition [ p22 ] we obtain @xmath76 the following monotonicity property of @xmath40 is well - known ( see , e.g. , @xcite ) .
[ p23 ] * @xmath77 for any @xmath78 . *
let @xmath10 and @xmath79 .
then @xmath80 if and only if @xmath81 . here and throughout the paper
we will use lexicographcial order @xmath82 or @xmath83 between sequences .
in this section we will consider @xmath8 for @xmath37 , and prove theorem [ t11 ] for @xmath37 . for @xmath84
we define @xmath85 .
note by proposition [ p22 ] that @xmath40 does not change for any @xmath86 $ ] . then @xmath87 [ l31 ] let @xmath88 . then @xmath89 if and only if @xmath90 .
if @xmath91 , then @xmath92 for some @xmath93 .
so , there exist @xmath86 $ ] and @xmath94 such that @xmath95 this implies that @xmath96 .
on the other hand , suppose that @xmath97 .
then @xmath98 for any @xmath99 .
this implies that @xmath100 .
note that @xmath50 increases to @xmath5 as @xmath51 .
then by lemma [ l31 ] it follows that @xmath9 if and only if @xmath101 .
so , in order to prove theorem [ t11 ] it suffices to show that the union of all @xmath102 covers the interval @xmath103 .
first we construct a sequence @xmath104 in @xmath103 , where @xmath105 here the second equality holds by proposition [ p21 ] .
we will show in lemma [ l33 ] that @xmath106 and the sequence @xmath104 strictly decreases to @xmath16 .
[ l32 ] for @xmath107 the set @xmath108 contains all sequences of the form @xmath109 note by proposition [ p22 ] that @xmath108 contains the sequences @xmath110 where @xmath111 .
observe that @xmath112 , and by proposition [ p21 ] we have @xmath113 therefore , by ( [ e32 ] ) it follows that @xmath108 contains the sequences @xmath114 [ l33 ] @xmath106 for any @xmath84 .
furthermore , @xmath115 strictly decreases to @xmath16 as @xmath51 . by taking @xmath116 in lemma [ l32 ] and using proposition [ p21 ]
it follows that @xmath106 .
then we only need to prove the monotonicity . by ( [ e21 ] )
it follows that @xmath117 for any @xmath84 .
note that @xmath118 as @xmath51 .
then @xmath119 so , it suffices to show that @xmath120 for any @xmath84 .
observe that @xmath121 .
furthermore , by lemma [ l32 ] we have @xmath122 and in a similar way as in the proof of lemma [ l32 ] one can verify that @xmath123 .
then by propositions [ p21 ] and [ p23 ] ( b ) it follows that @xmath124 now we prove that the union of all @xmath102 covers @xmath103 .
[ l34 ] @xmath125 by proposition [ p22 ] it follows that @xmath126 .
then by ( [ e31 ] ) we have @xmath127\right\}}&={\left\{(1^{\infty})_q : q\in(1 , q_1]\right\}}\\ & = [ ( 1^{\infty})_{q_1 } , { \infty})=[z_1,{\infty } ) .
\end{split}\ ] ] [ l35 ] for @xmath107 we have @xmath128 fix @xmath107 . by lemma [ l32 ]
it follows that @xmath108 contains the sequences @xmath129 note that @xmath130 then by proposition [ p23 ] ( b ) and lemma [ l32 ] it follows that @xmath131 for any @xmath86 $ ] .
observe by ( [ e21 ] ) and proposition [ p21 ] that @xmath132 then by ( [ e31 ] ) and ( [ eq:33 ] ) it follows that @xmath133\right\}}=\bigcup_{k=1}^{\infty}\big [ ( { { \mathbf{c}}}_k)_{q_n } , ( { { \mathbf{c}}}_k)_{q_{n-1}}\big).\end{aligned}\ ] ] in the following we will show that the union on the right indeed equals @xmath134 . observe that @xmath135 and @xmath136 increases to @xmath137 then by ( [ eq:33 ] ) it suffices to prove that @xmath138 for any @xmath139 by proposition [ p21 ] and ( [ e21])([e22 ] ) it follows that @xmath140 on the other hand , observe that @xmath141 this implies that @xmath142 whence , @xmath143 where the second inequality follows by using @xmath107 and @xmath144 .
therefore , by ( [ e34 ] ) we conclude that @xmath145 .
[ proof of theorem [ t11 ] for @xmath37 ] by lemmas [ l34 ] and [ l35 ] it follows that @xmath147 note by lemma [ l33 ] that @xmath115 strictly decreases to @xmath16 as @xmath51 .
then @xmath148 hence , by lemma [ l31 ] we conclude that @xmath9 for any @xmath37 .
in this section we will consider @xmath8 for @xmath26 , and finish the proof of theorems [ t11 ] .
recall from ( [ e31 ] ) that for @xmath84 we have @xmath149 [ l41 ] @xmath150 . by proposition [ p22 ]
it follows that @xmath151 furthermore , for any @xmath71 $ ] we have @xmath152 therefore , the lemma follows by the definition of @xmath153 .
[ l42 ] @xmath154 note by proposition [ p23 ] ( a ) that @xmath155 for any @xmath156 .
then by proposition [ p22 ] we have @xmath157 for any @xmath107 and @xmath158 .
observe that @xmath50 strictly increases to @xmath5 as @xmath51 .
then @xmath159 note by proposition [ p23 ] ( b ) that the sequence @xmath160 strictly decreases to @xmath161 as @xmath162 , and for @xmath163 we have @xmath164 . then @xmath165 so , to complete the proof it suffices to show that @xmath166 we will prove this by induction on @xmath167 .
first we consider @xmath168 .
then by ( [ e21 ] ) and ( [ e22 ] ) we obtain @xmath169 now we assume that ( [ e41 ] ) holds for some @xmath170
. then @xmath171 by induction this establishes ( [ e41 ] ) . by lemmas [ l31 ] , [ l41 ] and [ l42 ] it follows that @xmath9 for any @xmath172.\ ] ] in the following lemma we show that the inclusion in lemma [ l42 ] is indeed an equality . [ l43 ] for any @xmath84 we have @xmath173=\emptyset.\ ] ] by lemma [ l41 ] we may assume @xmath107 .
suppose on the contrary that there exists @xmath174\ ] ] for some @xmath175 , where @xmath94 and @xmath86 $ ] . by ( [ e42 ] )
it follows that @xmath176 note that @xmath157 . then by proposition [ p23 ] ( b )
it follows that @xmath177 on the other hand , note that @xmath178 again , by proposition [ p23 ] ( b ) we obtain that @xmath179 observe by proposition [ p22 ] that @xmath180 is the lexicographically smallest sequence in @xmath108 beginning with @xmath181 .
moreover , @xmath182 is the lexicographically largest sequence in @xmath108 starting at @xmath183 . hence , by ( [ e43 ] ) and ( [ e44 ] ) it follows that @xmath184 then by ( [ e22 ] ) this implies @xmath185 leading to a contradiction with ( [ e42 ] ) . by lemmas
[ l31 ] and [ l43 ] it follows that @xmath186 for any @xmath187 $ ] .
this yields that @xmath188
if @xmath189 so , to complete the proof it suffices to show that @xmath190 for any @xmath191 suppose on the contrary that there exists @xmath192 for some @xmath175 , where @xmath193 . then by proposition [ p23 ] ( b ) it gives that @xmath194 note by proposition [ p22 ] that @xmath180 is the lexicographically smallest sequence in @xmath195 beginning with @xmath181 .
furthermore , any sequence in @xmath195 starting at @xmath183 can not exceed @xmath196 .
then by ( [ e46 ] ) it follows that @xmath197 therefore , @xmath198 leading to a contradiction with ( [ e45 ] ) .
in this section we will determine the explicit value of @xmath8 when @xmath9 , and prove theorem [ t12 ] .
recall from ( [ e21 ] ) that @xmath199 , and @xmath200 strictly increases to @xmath5 as @xmath162 .
therefore , @xmath201,\ ] ] where the unions on the right are pairwise disjoint . by lemma [ l51 ]
it follows that @xmath202 $ ] .
let @xmath207 $ ] satisfy ( [ e52 ] ) .
note that @xmath208 .
then by proposition [ p22 ] it follows that @xmath209 .
this implies that @xmath210 has a unique @xmath211-expansion , i.e. , @xmath212 .
so , @xmath213 . in the following it suffices to prove that @xmath214 .
take @xmath215 $ ] , and assume that @xmath216 is the unique @xmath12-expansion of @xmath0 . then by proposition [ p22 ]
it follows that @xmath217 . furthermore , @xmath218 . by the definition of @xmath205 we have @xmath219 hence , by using @xmath220
we conclude that @xmath221
. therefore , @xmath222 .
another application of theorem [ t52 ] is to consider @xmath8 for @xmath233 . by lemma [ l35 ]
it follows that @xmath234 for @xmath144 let @xmath235 then @xmath236 strictly increases to @xmath237 as @xmath162 .
observe that @xmath238 then the sequence @xmath239 forms a partition of @xmath240 .
fix @xmath249 and take @xmath250 . then @xmath251 . by theorem [ t52 ]
it follows that @xmath248 $ ] is the appropriate root of @xmath252 where @xmath205 is the lexicographically smallest sequence in @xmath253 such that @xmath254 .
so , it suffices to prove that @xmath255 .
observe that @xmath256 then by proposition [ p23 ] ( b ) it follows that @xmath257 . by proposition [ p22 ]
the smallest sequence in @xmath258 which is larger than @xmath259 is @xmath260 .
furthermore , @xmath261 .
therefore , by the definition of @xmath205 we have @xmath262 as required . * by theorem [ t11 ] it follows that @xmath265 for any @xmath266 can we determine the exact value of @xmath8 when @xmath265 ? * by theorem [ t12 ] we know that @xmath34 if @xmath4 .
is it true that @xmath34 when @xmath265 ?
allouche and j. shallit , the ubiquitous prouhet - thue - morse sequence , sequences and their applications ( singapore , 1998 ) , springer ser . _ discrete math .
, springer _ , london , ( 1999 ) , 116 . | given a positive real number @xmath0 , we consider the smallest base @xmath1 for which there exists a unique sequence @xmath2 of zeros and ones such that @xmath3 in this paper we give complete characterizations of those @xmath0 s for which @xmath4 , where @xmath5 is the komornik - loreti constant . furthermore , we show that @xmath6 if and only if @xmath7 finally , we determine the explicit value of @xmath8 if @xmath9 . |
this work is concerned with the numerical solution of the steady - state equation of radiative transfer ( ert ) with isotropic physical coefficients and scattering kernel @xcite : @xmath0 where @xmath1 ( @xmath2 ) is a bounded domain with smooth boundary @xmath3 , @xmath4 is the unit sphere in @xmath5 , and @xmath6 ( @xmath7 being the unit outer normal vector at @xmath8 ) is the incoming part of the phase space boundary . for the only reason of simplifying the presentation , we have assumed that there is no incoming source on the boundary .
moreover , we have assumed that the internal source @xmath9 is only a function of the spatial variable .
in fact , this is not needed either for our algorithm to work ; see more discussions in section [ sec : concl ] .
the equation of radiative transfer is a popular model for describing the propagation of particles in complex media .
it appears in many fields of science and technology , ranging from classical fields such as nuclear engineering @xcite , astrophysics @xcite , and remote sensing @xcite , to modern applications such as biomedical optics @xcite , radiation therapy and treatment planning @xcite , and imaging in random media @xcite .
the coefficients @xmath10 and @xmath11 have different physical meanings in different applications . in general , the coefficient @xmath11 measures the strength of the scattering of the underlying medium at @xmath12 , while @xmath13 measures the strength of the physical absorption of the medium .
the coefficient @xmath10 measures the total absorption at @xmath12 due to both the physical absorption and absorption caused by scattering , that is the loss of particles from the current traveling direction into other directions due to scattering .
numerical methods for solving the equation of radiative transfer has been extensively studied , see for instance @xcite and references therein for an overview . besides monte carlo type of methods that are based on stochastic representation of the ert @xcite , many different deterministic discretization schemes have been proposed @xcite and numerous iterative schemes , as well as preconditioning strategies , have been developed to solve the discretized systems ; see for instance @xcite and references therein .
there are many challenging issues in the numerical solutions of the equation of radiative transfer .
one of such challenges is the high - dimensionality involved .
the ert is posed in phase space , meaning that the main unknown in the equation , in steady state , depends on both the spatial variable and the angular variable . in the spatial three - dimensional case
, the unknown @xmath14 depends on five variables , three in the spatial domain and two in the angular domain .
this poses significant challenges in terms of both solution speed and storage . in this work ,
we propose a new method to solve the ert in isotropic media , that is , media whose physical coefficients and the scattering kernel do not depend on the angular variable @xmath15 , i.e. , the media absorb and scatter particles in the same manner for all directions .
our method is based on the observation that when the underlying medium is isotropic , the angularly averaged ert solution , @xmath16 , satisfies a fredholm integral equation of the second type .
this integral equation can be solved , using a fast multiple method , for @xmath16 .
once this is done , we can plug @xmath16 into the ert to solve for @xmath14 itself .
the rest of this paper is organized as follows . in section [ sec : integral ] , we re - formulate the ert into a fredholm integral equation of the second type for the unknown @xmath16 .
we then propose in section [ sec : fmm ] a numerical procedure for solving the ert based on this integral formulation and implement an interpolation - based fast multipole method @xcite to solve the integral equation .
important issues on the implementation of our method are discussed in section [ sec : impl ] . in section [ sec : num ] we present some numerical tests for the algorithm that we developed .
concluding remarks are then offered in section [ sec : concl ] .
our algorithm is based on the integral formulation of the ert .
this is a well - developed subject .
we refer to @xcite for more details .
to present the formulation , let us first introduce a function @xmath17 defined as @xmath18 we can then rewrite the equation of radiative transfer , using the method of characteristics , into the following integral form @xcite : @xmath19 here @xmath20 is the distance it takes for a particle to go from @xmath12 to reach the domain boundary @xmath3 in the @xmath21 direction : @xmath22 the integral formulation in is classical and has been used to derive many theoretical results and numerical methods on the ert @xcite .
the most crucial step of our algorithm is to integrate the integral formulation again over @xmath4 to obtain an integral equation for the local density @xmath23 : @xmath24 the result is a fredholm integral equation of the second type .
it reads @xmath25 where the linear integral operator @xmath26 is defined as @xmath27 to simplify the expression for @xmath26 , let @xmath28 , and define the function @xmath29 @xmath30 which is nothing but the total absorption along the line segment between @xmath12 and @xmath31 .
we can then express the integral operator @xmath26 as @xmath32 where the integral kernel @xmath33 is defined as @xmath34 with @xmath35 the surface area of the unit sphere @xmath4 .
@xmath36 when @xmath37 and @xmath38 when @xmath39 . in the case where @xmath40 and @xmath41 are independent of the spatial variable
, the integral kernel @xmath33 simplifies to @xmath42 the algorithm we propose here is based on the integral formulation of the ert for the variable @xmath43 that we derived in .
we need the following result on the operator @xmath26 .
the proof is standard .
let @xmath40 and @xmath41 be bounded such that @xmath44 @xmath45 .
then the linear operator @xmath46 , defined in , is compact . for any @xmath47
, we define @xmath48 since @xmath40 and @xmath41 are bounded , we conclude that @xmath49 is bounded , by boundedness of @xmath50 , and therefore @xmath51 .
therefore , the operator @xmath52 defined as @xmath53 is a hilbert - schmidt integral operator and hence a compact operator .
let @xmath54 be a sufficiently large ball that contains @xmath50 , that is , @xmath55 .
for any @xmath56 , we have that @xmath57 where the last step comes from the young s convolution theorem .
this implies that , when @xmath58 , we have @xmath59 therefore , @xmath60 as @xmath58 . since @xmath52 is compact for each @xmath47 ,
we conclude , by for instance ( * ? ? ?
* chapter 3 , theorem 5 ) , that @xmath26 is compact . from ( [ eq : ert intgrl u ] )
, we can obtain that @xmath61 where @xmath62 .
the operator @xmath63 is a fredholm operator , and by fredholm alternative theorem and the fact that the ert admits only the zero solution when @xmath64 , see for instance @xcite , we conclude that there is a unique solution to .
let us finish this section with the following important observation .
the kernel for the volume integral equation that we derived here takes the same form in the cases of homogeneous ( i.e. @xmath40 and @xmath41 do not depend on spatial variable ) and inhomogeneous ( i.e. @xmath40 and @xmath41 depend on spatial variable ) media .
this means that the algorithm that we present in the next sections work for both homogeneous and inhomogeneous media , even though in the case of homogeneous media some simplifications can be made to reduce the computational costs of the algorithm .
this is quite different for integral formulations of many other problems , such as the helmholtz or the laplace equation where only homogeneous problems can be done with explicit kernels ( that are mostly the corresponding green functions ) @xcite .
our strategy of solving the ert is to first solve for @xmath43 and then solve for @xmath14 from @xmath43 .
the main solution procedure is as follows .
* algorithm i : general solution procedure * 1 .
evaluate the source function @xmath65 analytically , or by : 1 . solving the following scattering - free transport equation for @xmath66 : @xmath67 2 .
evaluating @xmath68 .
2 . use a krylov subspace method , such as the gmres algorithm or the minres algorithm @xcite , to solve the integral equation for @xmath43 .
3 . recover the ert solution @xmath14 by 1 .
evaluating the source @xmath69 ; 2 . solving the following scattering - free transport equation for @xmath14 : @xmath70 the solution of the scattering - free transport equations in the first and last steps can be done efficiently with a fast sweeping method such as that in @xcite or even analytically in special cases .
therefore , our focus here will be on the solution of the integral equation in the second step .
let us remark that one feature of the above method for solving the ert is that it _ does not _ require an explicit discretization over the angular variable .
it is clear that the main computational cost of the algorithm is on the solution of the integral equation which involves only the spatial variable .
therefore , besides the solution of the scattering - free transport equation , the computational complexity of the algorithm _ does not _ scale with the size of the angular discretization . in many applications , the main quantities of interests
is the local density @xmath23 , not @xmath71 . in these cases ,
the @xmath72 step of algorithm i is not necessary .
the computational complexity of the algorithm therefore is completely independent of the angular discretization .
for the same reason , the storage requirement of the algorithm also depends only on the spatial discretization .
there are many existing methods for the discretization of the integral equation with weakly singular kernel @xmath73 ; see for instance @xcite and references therein . here
we assume that we have a spatial discretization , consisting of @xmath74 nodes , of the integral equation that gives us the following approximation to the integral equation @xmath75 with @xmath76 the weight for the @xmath77-th point . since @xmath33 is singular at @xmath78 , we set @xmath79 in the above summation and use the weight @xmath80 to control the self - contribution of @xmath81 to the summation .
to solve the integral equation @xmath82 with a gmres or minres algorithm , we need to be able to evaluate matrix - vector product of the form @xmath83 for different vectors @xmath43 .
therefore , the main computational cost will be determined by the computational cost of the evaluation of @xmath84 , that is the summation .
direct evaluation of such a summation takes @xmath85 operations in general . in this work
, we use the fast multipole method ( fmm ) , originally developed by greengard and rokhlin @xcite , to accelerate the evaluation of this matrix - vector product . for the simplicity of implementation
, we use an interpolation - based fmm that was proposed by fong and darve in @xcite .
other efficient implementations of fmm , see for instance @xcite and references therein , may also be applied to our problem here .
this will be a future work .
the fmm method in @xcite , based on chebyshev interpolation , works as follows .
let @xmath86 be the first - kind chebyshev polynomial of degree @xmath87 defined on @xmath88 $ ] .
define the function @xmath89 with the conventions @xmath90 and @xmath91 .
then a two - variable kernel @xmath73 , assuming to be smoothed enough , on @xmath88^d\times [ -1 , 1]^d$ ] can be approximated by the following interpolation formula @xcite @xmath92 where @xmath93 , @xmath94 being the set of chebyshev interpolation notes which are simply taken as the set of the @xmath95-dimensional tensor product of the chebyshev nodes of @xmath96 .
the same approximation can be constructed when the kernel @xmath33 is defined on any regular domains by a linear transform .
if we now plug the approximation into the summation , we have , after a slight re - arrangement , the following formula @xmath97 this formula allows us to evaluate @xmath84 efficiently in three steps by simply following the order of the summations : ( i ) evaluate @xmath98 , @xmath99 ; ( ii ) evaluate @xmath100 , @xmath101 ; and then ( iii ) evaluate @xmath102 , @xmath103 .
if the computational cost of the evaluations of the interpolation polynomial @xmath104 and the kernel @xmath33 do not scale with @xmath105 and @xmath74 , then the costs of the three steps scale as @xmath106 , @xmath107 and @xmath106 respectively . therefore , the total cost scales as @xmath108 when @xmath109 is sufficiently small . in our implementation of the fong - darve fmm algorithm @xcite , we follow the standard multilevel approach with tree structures
the only specialty for our implementation is the related to the evaluation of the kernel @xmath110 for the pair @xmath111 which we describe in the next section .
we now briefly comment on some important issues on the implementation of the algorithm we described in the previous section .
[ [ validity - of - low - rank - approximation . ] ] validity of low rank approximation .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + due to boundedness of the exponential factor @xmath29 at @xmath78 and the fact that @xmath29 decays as a function of @xmath112 , our kernel @xmath113 in should admit the same if not better low - rank approximation as the kernel @xmath114 which has been well - studied in the fast multipole method community @xcite .
this justifies the chebyshev interpolation in .
[ [ the - computational - cost . ] ] the computational cost .
+ + + + + + + + + + + + + + + + + + + + + + + the mostly computationally expensive step of the fmm algorithm is step ( ii ) of evaluating where we have to evaluate the integral kernel @xmath33 for different @xmath115 pairs .
each evaluation of the kernel requires the evaluation of a line integral of the total absorption coefficient @xmath40 along the line that connects @xmath12 and @xmath31 .
if the integral can be analytically computed , for instance when @xmath40 is constant in which case the integral is simply @xmath116 , this evaluation is relatively cheap . otherwise , these evaluations have to be done numerically with selected quadrature rules . in many practical applications ,
the total absorption coefficient @xmath40 consists of a constant background with localized perturbations . in this case
, we can think of @xmath40 as a function with the periodic boundary condition .
we can therefore accelerate the evaluation of the line integrals with the technique of fast fourier transform ( fft ) .
assume that @xmath40 is sufficiently smooth to allow for the @xmath117 term fourier representation : @xmath118 where @xmath119 is assumed to ensure that @xmath40 is real - valued .
it is then straightforward to verify that the line integral of @xmath40 from @xmath12 to @xmath31 is given by @xmath120 where @xmath121 , @xmath122 , and @xmath123 for a given set of chebyshev interpolation points , we have a fixed number of pairs of @xmath115 for which we need to evaluate @xmath113 . in our implementation
, we cache all these kernel evaluations .
these kernel evaluations will be reused without any extra calculations during the gmres iterations ; see for instance the numerical results in tab . [
tab : con coeff ] and tab .
[ tab : var coeff ] of section [ sec : num ] .
[ [ fmm - approximation - accuracy . ] ] fmm approximation accuracy .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + the accuracy of solution to the ert with our numerical procedure depends mainly on two factors : the resolution of the spatial discretization , and the accuracy of the fast multipole approximation of the summation , the latter relying on the order of the chebyshev polynomial used .
increasing the order of the polynomial will increase the accuracy of the approximation in general .
however , that will also increase the computational cost of the algorithm , due to the increase of cost in evaluating @xmath104 for instance .
we have to therefore balance between accuracy and cost .
compared to existing kernels that have been studied the fmm community , our kernel here decays faster when the total absorption @xmath40 is large .
we therefore have to use more chebyshev interpolation points in general to ensure accuracy of the approximation .
let @xmath124 be the numerical solution with a direct evaluation of the summation in and @xmath125 the fmm - accelerated numerical solution .
when @xmath124 and @xmath125 solutions are computed on the same mesh , the finer mesh will produce larger error @xmath126 when the same number of interpolation points is used .
this is because finer mesh provides structures that are harder to capture with the same interpolation polynomial .
moreover , the accuracy of approximating @xmath124 by @xmath125 depends on the total absorption coefficient @xmath40 since the larger @xmath40 is , the faster the exponential decay is in the integral kernel .
therefore , for the same order of interpolation , the larger @xmath40 is , the worse the approximation is .
we observe these phenomenon in our numerical experiments ; see for instance , the simulations in section [ sec : num ] .
0.5 cm we now present some numerical simulations to demonstrate the performance of our algorithm .
we focus on the comparison between the algorithm for solving with a regular gmres solver and with our algorithm , i.e. , a gmres solver with fmm accelerations on the evaluation of .
our main purpose is to demonstrate that the computational complexity of the fmm - accelerated gmres algorithm indeed scales linearly with respect to the size of the spatial discretization , while maintains desired accuracy .
this means that the main cost of our algorithm for solving the ert is independent of the angular discretization . in all the simulations , we nodimensionalize the transport equation .
all the simulations are done in the fixed square domain @xmath127 with physical absorption coefficient @xmath128 .
we vary the scattering coefficient to test the performance of the algorithm in different regimes .
the larger the scattering coefficient @xmath40 is , the more diffusive the solution of the ert behaves , since the size of the domain and the physical absorption coefficient are fixed .
however , as we will see , the performance of our algorithm does not change dramatically from the low scattering transport regime to the highly scattering diffusive regime .
we introduce four time measures : ( i ) @xmath129 denotes the time cost of direct summation for ; ( ii ) @xmath130 denotes the time cost of fmm evaluation of ; ( iii ) @xmath131 denotes the time cost of the gmres algorithm with direct summation for solving ; and ( iv ) @xmath132 denotes the time cost of the gmres algorithm with fmm - acceleration for solving . note that in our computations , we have cached all the line integrals needed when setting up the algorithm .
therefore , the @xmath133 ( resp .
@xmath134 ) does not include @xmath130 ( resp .
@xmath129 ) .
all the computational times shown below are based on a dell optiplex 745 pentium d 3.4ghz desktop with 16 gb ram . to measure the accuracy of the fmm - accelerated calculation , with respect to the solution of the regular discretization
, we use the relative @xmath135 error @xmath136 where @xmath124 and @xmath125 are respectively the solutions with the direct gmres algorithm and the fmm - accelerated gmres algorithm .
l*6cl @xmath74 & @xmath137 & @xmath138 & @xmath139 & @xmath140 & @xmath141 & relative error ' '' '' ' '' '' + ' '' '' 1,024 & 4 & 1.20e @xmath142 01 & 3.51e @xmath142 02 & 1.18e @xmath143 00 & 3.71e @xmath143 00 & 2.17e @xmath142 04 + 4,096 & 4 & 6.65e @xmath142 01 & 1.25e @xmath142 01 & 4.02e @xmath143 01 & 6.23e @xmath143 01 & 3.39e @xmath142 04 + 16,384 & 4 & 3.25e @xmath143 00 & 4.98e @xmath142 01 & 1.09e @xmath143 03 & 9.93e @xmath143 02 & 3.90e @xmath142 04 + 65,536 & 4 & 1.50e @xmath143 01 & 2.04e @xmath143 00 & & & + 262,144 & 4 & 6.29e @xmath143 01 & 1.00e @xmath143 01 & & &
+ + 1,024 & 6 & 2.38e @xmath142 01 & 4.95e @xmath142 02 & 1.18e @xmath143 00 & 3.71e @xmath143 00 & 2.06e @xmath142 06 + 4,096 & 6 & 1.38e @xmath143 00 & 1.61e @xmath142 01 & 4.02e @xmath143 01 & 6.23e @xmath143 01 & 3.02e @xmath142 06 + 16,384 & 6 & 8.23e @xmath143 00 & 8.83e @xmath142 01 & 1.09e @xmath143 03 & 9.93e @xmath143 02 & 3.18e @xmath142 06 + 65,536 & 6 & 3.59e @xmath143 01 & 2.89e @xmath143 00 & & & + 262,144 & 6 & 1.59e @xmath143 02 & 1.23e @xmath143 01 & & & + + 1024 & 9 & 5.43e @xmath142 01 & 1.32e @xmath142 01 & 1.18e @xmath143 00 & 3.71e @xmath143 00 & 9.90e @xmath142 16 + 4096 & 9 & 3.64e @xmath143 00 & 5.29e @xmath142 01 & 4.02e @xmath143 01 & 6.23e @xmath143 01 & 4.24e @xmath142 09 + 16,384 & 9 & 2.17e @xmath143 01 & 2.80e @xmath143 00 & 1.09e @xmath143 03 & 9.93e @xmath143 02 & 4.47e @xmath142 09 + 65,536 & 9 & 1.10e @xmath143 02 & 1.09e @xmath143 01 & & & + 262,144 & 9 & 5.76e @xmath143 02 & 4.87e @xmath143 01 & & & + [ [ experiment - i . ] ] experiment i. + + + + + + + + + + + + + in the first set of numerical experiments , we perform simulations with a fixed scattering coefficient @xmath144 and total absorption coefficient @xmath145 ( which means the physical absorption is @xmath146 ) .
the source function we used is a ring source illustrated in the left plot of fig .
[ fig : sources ] . in tab .
[ tab : con coeff ] we show comparisons in three groups with increasing number of chebyshev interpolation points : @xmath147 , @xmath148 and @xmath149 .
we first note that , with reasonable relative approximation accuracy ( on the order of @xmath150 with @xmath147 ) , the ffm - gmres algorithm outperforms the regular gmres algorithm dramatically .
this trend is kept when we increase the accuracy of the fmm approximation by increasing , @xmath137 , the number of chebyshev interpolation points .
when the spatial discretization is too fine , it takes the regular gmres algorithm too much time to finish the calculations .
however , the ffm - accelerated gmres can still solve the system in relatively short time .
l*6cl @xmath74 & n & @xmath138 & @xmath139 & @xmath140 & @xmath141 & relative error ' '' '' ' '' '' + ' '' '' 1,024 & 4 & 2.01e @xmath142 01 & 3.07e @xmath142 02 & 8.46e @xmath142 01 & 1.94e @xmath143 00 & 2.16e @xmath142 04 + 4,096 & 4 & 1.12e @xmath143 00 & 9.55e @xmath142 02 & 2.88e @xmath143 01 & 3.26e @xmath143 01 & 3.23e @xmath142 04 + 16,384 & 4 & 5.26e @xmath143 00 & 3.52e @xmath142 01 & 7.78e @xmath143 02 & 5.20e @xmath143 02 & 3.69e @xmath142 04 + 65,536 & 4 & 2.35e @xmath143 01 & 1.62e @xmath143 00 & & & + 262,144 & 4 & 1.01e @xmath143 02 & 6.64e @xmath143 00 & & &
+ + 1,024 & 6 & 3.19e @xmath142 01 & 2.15e @xmath142 02 & 8.46e @xmath142 01 & 1.94e @xmath143 00 & 1.73e @xmath142 05 + 4,096 & 6 & 2.37e @xmath143 00 & 1.27e @xmath142 01 & 2.88e @xmath143 01 & 3.26e @xmath143 01 & 1.36e @xmath142 05 + 16,384 & 6 & 1.30e @xmath143 01 & 5.23e @xmath142 01 & 7.78e @xmath143 02 & 5.20e @xmath143 02 & 6.83e @xmath142 06 + 65,536 & 6 & 6.24e @xmath143 01 & 2.22e @xmath143 00 & & & + 262,144 & 6 & 2.89e @xmath143 02 & 9.79e @xmath143 00 & & & + + 1024 & 9 & 7.40e @xmath142 01 & 8.72e @xmath142 02 & 8.46e @xmath142 01 & 1.94e @xmath143 00 & 2.71e @xmath142 15 + 4096 & 9 & 7.01e @xmath143 00 & 4.44e @xmath142 01 & 2.88e @xmath143 01 & 3.26e @xmath143 01 & 4.84e @xmath142 06 + 16,384 & 9 & 4.18e @xmath143 01 & 2.32e @xmath143 00 & 7.78e @xmath143 02 & 5.20e @xmath143 02 & 3.09e @xmath142 06 + 65,536 & 9 & 2.10e @xmath143 02 & 1.05e @xmath143 01 & & & + 262,144 & 9 & 1.02e @xmath143 03 & 4.85e @xmath143 01 & & & + [ [ experiment - ii . ] ] experiment ii .
+ + + + + + + + + + + + + + in the second set of numerical experiments , we repeat the simulations in experiment i for an inhomogeneous medium . the coefficients are given as @xmath151 we again use the a ring source illustrated in the left plot of fig .
[ fig : sources ] . in tab .
[ tab : var coeff ] we show comparison in three groups with increasing number of chebyshev interpolation points .
the first noticeable difference between tab .
[ tab : var coeff ] and tab .
[ tab : con coeff ] is that the time it takes to evaluate the matrix - vector multiplication is now considerably more expensive .
this is mainly due to the fact that for variable coefficient , we need to evaluate the integrals by numerical quadrature rules , while in the constant coefficient case the kernels are given analytically for any pair @xmath115 . in our implementation , we cached all
the line integrals so that they can be used repeatedly during gmres iterations .
this is the reason why the solution costs @xmath133 for variable coefficient cases in tab .
[ tab : var coeff ] is very similar to the corresponding constant coefficient cases in tab [ tab : con coeff ] .
the overall computational costs again scale linearly with respect to the spatial discretization .
l*6cl @xmath74 & n & @xmath40 & @xmath41 & relative error ' '' '' ' '' '' + 1024 & 4 & 2.2 & 2.0 & 8.39e @xmath142 05 + 1024 & 4 & 5.2 & 5.0 & 1.93e @xmath142 04 + 1024 & 4 & 10.2 & 10.0 & 4.13e @xmath142 04 + + 1024 & 6 & 2.2 & 2.0 & 8.85e @xmath14207 + 1024 & 6 & 5.2 & 5.0 & 1.82e @xmath14206 + 1024 & 6 & 10.2 & 10.0 & 5.15e @xmath14206 + + 1024 & 9 & 2.2 & 2.0 & 5.02e @xmath14216 + 1024 & 9 & 5.2 & 5.0 & 7.91e @xmath14216 + 1024 & 9 & 10.2 & 10.0 & 1.17e @xmath14215 + l*6cl @xmath74 & n & @xmath40 & @xmath41 & relative error ' '' '' ' '' '' + 4096 & 4 & 2.2 & 2.0 & 1.15e @xmath142 04 + 4096 & 4 & 5.2 & 5.0 & 3.07e @xmath142 04 + 4096 & 4 & 10.2 & 10.0 & 9.08e @xmath142 04 + + 4096 & 6 & 2.2 & 2.0 & 1.27e @xmath142 06 + 4096 & 6 & 5.2 & 5.0 & 2.79e @xmath142 06 + 4096 & 6 & 10.2 & 10.0 & 9.62e @xmath142 06 + + 4096 & 9 & 2.2 & 2.0 & 1.39e @xmath142 08 + 4096 & 9 & 5.2 & 5.0 & 2.46e @xmath142 08 + 4096 & 9 & 10.2 & 10.0 & 6.86e @xmath142 08
+ l*6cl @xmath74 & n & @xmath40 & @xmath41 & relative error ' '' '' ' '' '' + 16384 & 4 & 2.2 & 2.0 & 1.26e @xmath142 04 + 16384 & 4 & 5.2 & 5.0 & 3.60e @xmath142 04 + 16384 & 4 & 10.2 & 10.0 & 1.20e @xmath142 03 + + 16384 & 6 & 2.2 & 2.0 & 1.30e @xmath142 06 + 16384 & 6 & 5.2 & 5.0 & 2.92e @xmath142 06 + 16384 & 6 & 10.2 & 10.0 & 1.04e @xmath142 05 + + 16384 & 9 & 2.2 & 2.0 & 1.82e @xmath142 08 + 16384 & 9 & 5.2 & 5.0 & 3.20e @xmath142 08 + 16384 & 9 & 10.2 & 10.0 & 9.71e @xmath142 08 + . from top to bottom : @xmath144 , @xmath152 and @xmath153 . shown
are ( from left to right ) : the solution @xmath124 and the error @xmath126 with @xmath147 , @xmath148 , and @xmath149 . ] except that @xmath154 . ] except that @xmath155 . ]
[ [ experiment - iii . ] ] experiment iii .
+ + + + + + + + + + + + + + + in the third set of numerical experiments , we study the dependence of the computational cost of the algorithm on the scattering coefficient of the ert . we perform simulations using the source function that is illustrated in the right plot of fig .
[ fig : sources ] .
the results are summarized in tab .
[ tab : periodic conv para-1024 ] , tab .
[ tab : periodic conv para-4096 ] and tab .
[ tab : periodic conv para-16384 ] for different scattering coefficients , @xmath156 , @xmath157 , and @xmath158 , with different levels of spatial discretizations .
the solution by the gmres algorithm with direct summation @xmath124 as well as the error @xmath126 ( with @xmath125 being the solution with a fmm - accelerated gmres algorithm ) are shown in fig .
[ fig : periodic 1024 ] , fig .
[ fig : periodic 4096 ] , and fig .
[ fig : periodic 16384 ] respectively for the domain with @xmath159 , @xmath160 and @xmath161 cells , using different numbers of chebyshev interpolation points .
the results show that the error of the fmm approximation does not change dramatically with respect to the change of the scattering coefficient .
that is , the algorithm we developed works in both diffusive regimes and transport regimes , as long as the medium is isotropic .
l*6cl @xmath74 & n & @xmath40 & @xmath41 & relative error ' '' '' ' '' '' + 1024 & 4 & 2.2 & 2.0 & 9.53e @xmath142 05 + 1024 & 4 & 5.2 & 5.0 & 2.13e @xmath142 04 + 1024 & 4 & 10.2 & 10.0 & 4.50e @xmath142 04 + + 1024 & 6 & 2.2 & 2.0 & 1.19e @xmath14206 + 1024 & 6 & 5.2 & 5.0 & 2.09e @xmath14206 + 1024 & 6 & 10.2 & 10.0 & 5.53e @xmath14206 + + 1024 & 9 & 2.2 & 2.0 & 7.71e @xmath14216 + 1024 & 9 & 5.2 & 5.0 & 8.47e @xmath14216 + 1024 & 9 & 10.2 & 10.0 & 1.10e @xmath14215 + l*6cl @xmath74 & n & @xmath40 & @xmath41 & relative error ' '' '' ' '' '' + 4096 & 4 & 2.2 & 2.0 & 1.24e @xmath142 04 + 4096 & 4 & 5.2 & 5.0 & 3.28e @xmath142 04 + 4096 & 4 & 10.2 & 10.0 & 9.58e @xmath142 04 + + 4096 & 6 & 2.2 & 2.0 & 1.48e @xmath142 06 + 4096 & 6 & 5.2 & 5.0 & 2.99e @xmath142 06 + 4096 & 6 & 10.2 & 10.0 & 1.01e @xmath142 05 + + 4096 & 9 & 2.2 & 2.0 & 1.85e @xmath142 08 + 4096 & 9 & 5.2 & 5.0 & 2.94e @xmath142 08 + 4096 & 9 & 10.2 & 10.0 & 7.68e @xmath142 08 + [ [ experiment - iv . ] ] experiment iv .
+ + + + + + + + + + + + + + we repeat here the numerical simulations in experiment iii with a different source function , the source function in the left plot of fig . [
fig : sources ] .
the relative error of the fmm - accelerated solutions are summarized in tab .
[ tab : ring conv para-1024 ] , tab .
[ tab : ring conv para-4096 ] , and tab .
[ tab : ring conv para-16384 ] .
the results are very similar respectively to those showed in tab .
[ tab : periodic conv para-1024 ] , tab .
[ tab : periodic conv para-4096 ] , and tab .
[ tab : periodic conv para-16384 ] .
this shows again that the performance of the algorithm does not depend on the strength of the scattering of the underlying medium .
overall , in either diffusive or transport regime , we can achieve very good accuracy with only a few chebyshev interpolation points in each direction .
the solution by the gmres algorithm with the direct summation , @xmath162 and the error @xmath163 are show in fig .
[ fig : ring 1024 ] and fig .
[ fig : ring 4096 ] . l*6cl @xmath74 & n & @xmath40 & @xmath41 & relative error ' '' '' ' '' '' + 16384 & 16 & 2.2 & 2.0 & 1.35e @xmath14204 + 16384 & 16 & 5.2 & 5.0 & 3.82e @xmath14204 + 16384 & 16 & 10.2 & 10.0 & 1.20e @xmath14203 + + 16384 & 36 & 2.2 & 2.0 & 1.51e @xmath14206 + 16384 & 36 & 5.2 & 5.0 & 3.14e @xmath14206 + 16384 & 36 & 10.2 & 10.0 & 1.09e @xmath14205 + + 16384 & 64 & 2.2 & 2.0 & 2.19e @xmath14208 + 16384 & 64 & 5.2 & 5.0 & 3.60e @xmath14208 + 16384 & 64 & 10.2 & 10.0 & 1.05e @xmath14207 + with the source in the left plot of fig .
[ fig : sources ] . from top to bottom : @xmath144 , @xmath152 and @xmath153 . shown
are ( from left to right ) : the solution @xmath124 and the error @xmath126 with @xmath147 , @xmath148 , and @xmath149 . ] except that @xmath154 . ]
to summarize , we presented in this work a fast numerical method for solving the equation of radiative transfer in isotropic media .
the main idea of the method is to reformulate the ert into an integral equation of the second type and then use the fast multipole technique to accelerate the solution of such an integral equation .
our numerical tests show that the algorithmic cost indeed scales linearly with respect to the size of the spatial component of the problem .
there are a few features of the method we proposed here .
first , with the integral formulation , we avoid angular discretization of the ert in the most expensive part of the solution process .
this in principle allows us to handle large problems that would be hard to handle in , for instance , the discrete ordinate formulation , with limited ram .
second , the kernel in our integral formulation of the ert takes the same form for homogeneous and inhomogeneous media .
therefore , the algorithm we developed does not need to be modified going from homogeneous media problems to inhomogeneous media problems .
this is quite different from existing fast multipole based methods .
that said , in homogeneous media , the setup of our algorithm is relatively computationally inexpensive since the kernel in the corresponding integral equation is explicitly given .
in inhomogeneous media , the setup requires the evaluation of the kernel for different @xmath164 pairs that involves line integrals of the total absorption coefficients between @xmath12 and @xmath31 .
this evaluation is more expensive than the homogeneous media case , but is still relatively low . in many practically relevant problems
, we have coefficients that can be treated as periodic functions .
fast fourier transform type of techniques can be used to accelerate the setup process of the algorithm . in our implementation of the fmm algorithm
, we cached all the calculations that involve the evaluation of the line integrals .
this does not cause major storage problem since the number of chebyshev interpolation nodes used in the implementation is always relatively small .
let us also emphasize that , even though our formulation requires that the underlying medium to be isotropic , the internal and boundary source functions need not to be isotropic at all .
in fact , the only thing that would have changed for the algorithm with an anisotropic source is the evaluation of @xmath165 .
in addition , as we have seen from our numerical tests , the fmm approximation with a very small number of chebyshev interpolation nodes already give relatively accuracy approximations to the true numerical solutions .
this suggests that we can probably use the algorithm with small numbers of chebyshev interpolation points as a preconditioning strategy for a general transport solver for more complicated problems .
we are currently exploring in this direction .
to the best of our knowledge , what we proposed is the first algorithm for solving the ert within the frame work of the fast multipole method .
our contribution is mainly on the introduction of the idea , not on the implementation of fast multipole methods .
indeed , our implementation is rather primitive which we believe can be greatly improved , either by refining the current strategy or by exploring other approaches @xcite .
the study we have in this short paper is by no means enough to draw conclusions on every aspect of the algorithm , for instance how the algorithm benchmarks with existing methods .
however , numerical simulations we have performed show that this is a promising method that is worth careful further investigated .
we hope that this work can motivate more studies in this direction .
this work is partially supported by the national science foundation through grants dms-1321018 and dms-1620473 . , _ a kinetic theory for nonanalog monte carlo particle transport algorithms : exponential transform with angular biasing in planar - geometry anisotropically scattering media _ , j. comp .
phys . , 145 ( 1998 ) , pp . | we propose in this work a fast numerical algorithm for solving the equation of radiative transfer ( ert ) in isotropic media .
the algorithm has two steps . in the first step ,
we derive an integral equation for the angularly averaged ert solution by taking advantage of the isotropy of the scattering kernel , and solve the integral equation with a fast multipole method ( fmm ) . in the second step ,
we solve a scattering - free transport equation to recover the original ert solution .
numerical simulations are presented to demonstrate the performance of the algorithm for both homogeneous and inhomogeneous media .
equation of radiative transfer , integral equation , fast algorithm , fast multipole method , preconditioning . 65f08 , 65n22 , 65n99 , 65r20 , 45k05 |
we consider a wireless communication network where users are able to harvest energy from the nature using rechargeable batteries .
such energy harvesting capabilities will make sustainable and environmentally friendly deployment of wireless communication networks possible .
while energy - efficient scheduling policies have been well - investigated in traditional battery powered ( un - rechargeable ) systems @xcite , energy - efficient scheduling in energy harvesting networks with nodes that have rechargeable batteries has only recently been considered @xcite .
references @xcite consider a single - user communication system with an energy harvesting transmitter , and develop a packet scheduling scheme that minimizes the time by which all of the packets are delivered to the receiver . in this paper
, we consider a multi - user extension of the work in @xcite .
in particular , we consider a wireless broadcast channel with an energy harvesting transmitter . as shown in fig .
[ fig : bc ] , we consider a broadcast channel with one transmitter and two receivers , where the transmitter node has three queues .
the data queues store the data arrivals intended for the individual receivers , while the energy queue stores the energy harvested from the environment .
our objective is to adaptively change the transmission rates that go to both users according to the instantaneous data and energy queue sizes , such that the total _ transmission completion time _ is minimized . in this paper , we focus on finding the optimum _ off - line _ schedule , by assuming that the energy arrival profile at the transmitter is known ahead of time in an off - line manner , i.e. , the energy harvesting times and the corresponding harvested energy amounts are known at time @xmath0 .
we assume that there are a total of @xmath1 bits that need to be delivered to receiver 1 , and @xmath2 bits that need to be delivered to receiver 2 , available at the transmitter at time @xmath0 . as shown in fig .
[ fig : bc_system ] , energy arrives ( is harvested ) at points in time marked with @xmath3 ; in particular , @xmath4 denotes the amount of energy harvested at time @xmath5 .
our goal is to develop a method of transmission to minimize the time , @xmath6 , by which all of the data packets are delivered to their respective receivers .
the optimal packet scheduling problem in a single - user energy harvesting communication system is investigated in @xcite . in @xcite
, we prove that the optimal scheduling policy has a `` majorization '' structure , in that , the transmit power is kept constant between energy harvests , the sequence of transmit powers increases monotonically , and only changes at some of the energy harvesting instances ; when the transmit power changes , the energy constraint is tight , i.e. , at the times when the transmit power changes , the total consumed energy equals the total harvested energy . in @xcite , we develop an algorithm to obtain the optimal off - line scheduling policy based on these properties .
reference @xcite extends @xcite to the case where rechargeable batteries have finite sizes .
we extend @xcite in @xcite to a fading channel . references @xcite investigate two related problems .
the first problem is to maximize the throughput ( number of bits transmitted ) with a given deadline constraint , and the second problem is to minimize the transmission completion time with a given number of bits to transmit .
these two problems are `` dual '' to each other in the sense that , with a given energy arrival profile , if the maximum number of bits that can be sent by a deadline is @xmath7 in the first problem , then the minimum time to transmit @xmath7 bits in the second problem must be the deadline in the first problem , and the optimal transmission policies for these two problems must be identical . in this paper
, we will follow this `` dual problems '' approach .
we will first attack and solve the first problem to determine the structural properties of the optimal solution .
we will then utilize these structural properties to develop an iterative algorithm for the second problem .
our iterative approach has the goal of reducing the two - user broadcast problem into a single - user problem as much as possible , and utilizing the single - user solution in @xcite .
the second problem is also considered in the independent work @xcite which uses convex optimization techniques to reduce the problem into local sub - problems that consider only two energy arrival epochs at a time .
we first analyze the structural properties of the optimal policy for the first problem where our goal is to maximize the number of bits delivered to both users under a given deadline constraint . to that end , we first determine the _ maximum departure region _ with a given deadline constraint @xmath6 .
the maximum departure region is defined as the set of all @xmath8 that can be transmitted to users reliably with a given deadline @xmath6 . in order to do that
, we consider the problem of maximizing @xmath9 under the energy causality constraints for the transmitter , for all @xmath10 . varying @xmath11 , @xmath12 traces the boundary of the maximum departure region .
we prove that the optimal _ total _ transmit power policy is independent of the values of @xmath11 , @xmath12 , and it has the same `` majorization '' structure as the single - user non - fading solution . as for the way of splitting the total transmit power between the two users , we prove that there exists a _ cut - off _ power level for the stronger user , i.e. , only the power above this _ cut - off _ power level is allocated to the weaker user .
we then consider the second problem , where our goal is to minimize the time , @xmath6 , by which a given @xmath13 number of bits are delivered to their intended receivers .
as discussed , since the second problem is `` dual '' to the first problem , the optimal transmission policy in this problem has the same structural properties as in the first problem .
therefore , in the second problem as well , there exists a _ cut - off _ power level .
the problem then becomes that of finding an optimal _ cut - off _ power such that the transmission times for both users become identical and minimized . with these optimal structural properties , we develop an iterative algorithm that finds the optimal schedule efficiently .
in particular , we first use the fact that the optimum total transmit power has the same structural properties as the single - user problem , to obtain the first optimal total power , @xmath14 , i.e. , the optimal total power in the first epoch .
then , given the fact that there exists a _ cut - off _ power level , @xmath15 , for the stronger user , the optimal transmit strategy depends on whether @xmath14 is smaller or larger than @xmath15 , which , at this point , is unknown .
therefore , we have two cases to consider .
if @xmath15 is smaller than @xmath14 , then the stronger user will always have a constant , @xmath15 , portion of the total transmit power .
this reduces the problem to a single - user problem for the second user , together with a fixed - point equation in a single variable ( @xmath15 ) to be solved to ensure that the transmissions to both users end at the same time . on the other hand ,
if @xmath15 is larger than @xmath14 , this means that all of @xmath14 must be spent to transmit to the first ( stronger ) user . in this case
, the number of bits delivered to the first user in this time duration can be subtracted from the total number of bits to be delivered to the first user , and the problem can be started anew with the updated number of bits @xmath16 after the first epoch .
therefore , in both cases , the broadcast channel problem is essentially reduced to single - user problems , and the approach in @xcite is utilized recursively to solve the overall problem .
the system model is as shown in figs .
[ fig : bc ] and [ fig : bc_system ] .
the transmitter has an energy queue and two data queues ( fig .
[ fig : bc ] ) .
the physical layer is modeled as an awgn broadcast channel , where the received signals at the first and second receivers are @xmath17 where @xmath18 is the transmit signal , and @xmath19 is a gaussian noise with zero - mean and unit - variance , and @xmath20 is a gaussian noise with zero - mean and variance @xmath21 , where @xmath22 .
therefore , the second user is the _ degraded _ ( weaker ) user in our broadcast channel . assuming that the transmitter transmits with power @xmath23 , the capacity region for this two - user awgn broadcast channel is @xcite @xmath24 where @xmath25 is the fraction of the total power spent for the message transmitted to the first user .
let us denote @xmath26 for future use .
then , the capacity region is @xmath27 , @xmath28 .
this capacity region is shown in fig .
[ fig : bc_capacity ] . working on the boundary of the capacity region
, we have @xmath29 as shown in fig .
[ fig : bc ] , the transmitter has @xmath1 bits to transmit to the first user , and @xmath2 bits to transmit to the second user .
energy is harvested at times @xmath5 with amounts @xmath4 .
our goal is to select a transmission policy that minimizes the time , @xmath6 , by which all of the bits are delivered to their intended receivers .
the transmitter adapts its transmit power and the portions of the total transmit power used to transmit signals to the two users according to the available energy level and the remaining number of bits .
the energy consumed must satisfy the causality constraints , i.e. , at any given time @xmath30 , the total amount of energy consumed up to time @xmath30 must be less than or equal to the total amount of energy harvested up to time @xmath30 . before we proceed to give a formal definition of the optimization problem and propose the solution ,
we start with the `` dual '' problem of this transmission completion time minimization problem , i.e. , instead of trying to find the minimal @xmath6 , we aim to identify the maximum number of bits the transmitter can deliver to both users by any fixed time @xmath6 . as we will observe in the next section , solving the `` dual '' problem enables us to identify the optimal structural properties for both problems , and these properties eventually help us reduce the original problem into simple scenarios , which can be solved efficiently .
in this section , our goal is to characterize the maximum departure region for a given deadline @xmath6 . we define it as follows . for any fixed transmission duration
@xmath6 , the maximum departure region , denoted as @xmath31 , is the union of @xmath8 under any feasible rate allocation policy over the duration @xmath33 , i.e. , @xmath34 , subject to the energy constraint @xmath35 , for @xmath36 .
we call any policy which achieves the boundary of @xmath31 to be optimal . in the single - user scenario in @xcite
, we first examined the structural properties of the optimal policy .
based on these properties , we developed an algorithm to find the optimal scheduling policy . in this broadcast scenario also , we will first analyze the structural properties of the optimal policy , and then obtain the optimal solution based on these structural properties .
the following lemma which was proved for a single - user problem in @xcite was also proved for the broadcast problem in @xcite .
[ lemma : const ] under the optimal policy , the transmission rate remains constant between energy harvests , i.e. , the rate only potentially changes at an energy harvesting epoch .
we prove this using the strict convexity of @xmath37 .
if the transmission rate for any user changes between two energy harvesting epochs , then , we can always equalize the transmission rate over that duration without contradicting with the energy constraints . based on the convexity of @xmath37 , after equalization of rates , the energy consumed over that duration decreases , and
the saved energy can be allocated to both users to increase the departures .
therefore , changing rates between energy harvests is sub - optimal .
therefore , in the following , we only consider policies where the rates are constant between any two consecutive energy arrivals .
we denote the rates that go to both users as @xmath38 over the duration @xmath39 . with this property , an illustration of the maximum departure region
is shown in fig .
[ fig : trajectory ] .
@xmath31 is a convex region .
proving the convexity of @xmath31 is equivalent to proving that , given any two achievable points @xmath13 and @xmath40 in @xmath31 , any point on the line between these two points is also achievable , i.e. , in @xmath31 .
assume that @xmath13 and @xmath40 can be achieved with rate allocation policies @xmath41 and @xmath42 , respectively .
consider the policy @xmath43 , where @xmath44 . then
, the energy consumed up to @xmath45 is @xmath46 therefore , the energy causality constraint is satisfied for any @xmath47 $ ] , and the new policy is energy - feasible .
any point on the line between @xmath13 and @xmath40 can be achieved . when @xmath48 , the inequality in ( [ eqn : g ] ) is strict .
therefore , we save some amount of energy under the new policy , which can be used to increase the throughput for both users .
this implies that @xmath31 is strictly convex . in order to simplify the notation , in this section , for any given @xmath6
, we assume that there are @xmath49 energy arrival epochs ( excluding @xmath0 ) over @xmath50 .
we denote the last energy arrival epoch before @xmath6 as @xmath51 , and @xmath52 , with @xmath53 , as shown in fig .
[ fig : bc_energy_rate ] .
since @xmath31 is a strictly convex region , its boundary can be characterized by solving the following optimization problem for all @xmath10 , @xmath54 where @xmath55 is the length of the duration between two consecutive energy arrival instances @xmath56 and @xmath57 , i.e. , @xmath58 , and @xmath59 and @xmath60 denote the rate sequences @xmath61 and @xmath62 for users 1 and 2 , respectively .
the problem in ( [ prob1 ] ) is a convex optimization problem with a convex cost function and a convex constraint set , therefore , the unique global solution should satisfy the extended kkt conditions .
the lagrangian is @xmath63 taking the derivatives with respect to @xmath61 and @xmath62 , and setting them to zero , we have @xmath64 where @xmath65 if @xmath66 , and @xmath67 if @xmath68 . based on these kkt optimality conditions , we first prove an important property of the optimal policy .
[ lemma4 ] the optimal total transmit power of the transmitter is independent of the values of @xmath69 , and it is the same as the single - user optimal transmit power .
specifically , @xmath70 i.e. , at @xmath71 , @xmath72 switches to @xmath73
. based on the expression of @xmath74 in ( [ dfn : g ] ) and the kkt conditions in ( [ cond1])-([cond2 ] ) , we have @xmath75 where ( [ p_ineq ] ) becomes an equality when @xmath76 .
therefore , when @xmath68 , ( [ eqn : g1])-([eqn : g3 ] ) imply @xmath77 when @xmath76 , we must have @xmath66 . otherwise ,
if @xmath78 , we can always let the weaker user transmit with some power over this duration without contradicting with any energy constraints .
since there is no interference from the stronger user , the departure from the weaker user can be improved , thus it contradicts with the optimality of the policy .
therefore , when @xmath76 , @xmath65 , and ( [ eqn : g1])-([eqn : g3 ] ) imply @xmath79 therefore , we can express @xmath74 in the following way : @xmath80 plotting these two curves in fig .
[ fig : power_curve ] , we note that the optimal transmit power , @xmath81 , is always the curve on the top . if @xmath82 for some @xmath83 , then , we have @xmath84 where the first inequality follows from the kkt condition that @xmath85 for @xmath86 .
therefore , we conclude that there exists an integer @xmath83 , @xmath87 , such that , when @xmath88 , @xmath76 ; and when @xmath89 , @xmath68 .
furthermore , ( [ g1])-([g2 ] ) imply that , the energy constraint at @xmath90 must be tight .
otherwise , @xmath91 , and ( [ g2 ] ) implies @xmath92 which contradicts with ( [ g1 ] ) .
therefore , in the following , when we consider the energy constraints , we only need to consider two segments @xmath93 and @xmath94 separately . when @xmath95 , based on ( [ g1 ] ) , if @xmath96 , we have @xmath97 .
starting from @xmath98 , @xmath99 remains a constant until an energy constraint becomes tight .
therefore , between any two consecutive epochs , when the energy constraints are tight , the power level remains constant .
similar arguments hold when @xmath100 .
thus , the corresponding power level is @xmath101 where @xmath102 and @xmath103 are two consecutive epochs with tight energy constraint . finally , we need to determine the epochs when the energy constraint becomes tight .
another observation is that @xmath104 must monotonically increase in @xmath105 , as shown in fig .
[ fig : power_curve ] .
this is because both of these two curves monotonically increase , and the maximum value of these two curves should monotonically increase also .
therefore , based on the monotonicity of the transmit power , we conclude that @xmath106 this completes the proof .
since the power can be obtained directly irrespective of the values of @xmath11 , @xmath12 , the optimization problem in ( [ prob1 ] ) is separable over each duration @xmath39 .
specifically , for @xmath107 , the local optimization becomes @xmath108 we relax the power constraint to be an inequality to make the constraint set convex .
thus , this becomes a convex optimization problem .
this does not affect the solution since the objective function is always maximized on the boundary of its constraint set , i.e. , the capacity region defined by the transmit power @xmath72 .
when @xmath109 , the solution to ( [ prob2 ] ) can be expressed as @xmath110 in this scenario , all of the power @xmath72 is allocated to the first user . when @xmath111 , we have @xmath112 in this scenario , a constant amount of power , @xmath113 , is allocated to the first user , and the remaining power is allocated to the second user . when @xmath114 , we have @xmath115 in this scenario , all of the @xmath72 is allocated to the second user .
let us define a constant power level as @xmath116 based on the solution of the local optimization problem in ( [ prob2 ] ) , we establish another important property of the optimal policy as follows .
[ lemma : cutoff ] for fixed @xmath11 , @xmath12 , under the optimal power policy , there exists a constant _ cut - off _ power level , @xmath15 , for the first user . if the total power level is below this _ cut - off _ power level , then , all the power is allocated to the first user ; if the total power level is higher than this level , then , all the power above this _ cut - off _ level is allocated to the second user . in the proof of lemma [ lemma4 ]
, we note that the optimal power @xmath72 monotonically increases in @xmath105 . combining lemma [ lemma4 ] and lemma [ lemma : cutoff ]
, we illustrate the structure of the optimal policy in fig .
[ fig : power ] .
moreover , the optimal way of splitting the power in each epoch is such that both users shares of the power monotonically increase in time . in particular
, the second user s share is monotonically increasing in time .
hence , the path followed in the @xmath13 plane is such that it changes direction to get closer to the second user s departure axis as shown in fig .
[ fig : trajectory ] .
the dotted trajectory can not be optimal , since the path does not get closer to the second user s departure axis in the middle ( second ) power epoch .
[ lemma2 ] under the optimal policy , the transmission rate for the first user , @xmath117 , is either a constant sequence ( zero or a positive constant ) , or an increasing sequence .
moreover , before @xmath61 achieves its final constant value , @xmath76 ; and when @xmath61 becomes a constant , @xmath62 monotonically increases in @xmath105 . based on lemma [ lemma4 ]
, we observe that for fixed @xmath6 , @xmath11 and @xmath12 , the optimal _ total _ power allocation is unique , i.e. , does not depend on @xmath11 and @xmath12 .
however , the way the total power is split between the two users depends on @xmath11 , @xmath12 .
in fact , the _ cut - off _ power level @xmath15 varies depending on the value of @xmath118 .
therefore , for different values of @xmath118 , the optimal policy achieves different boundary points on the maximum departure region , and varying the value of @xmath118 traces the boundary of this region . in this section
, we characterized the maximum departure region for any given time @xmath6 .
we proved that the optimal total transmit power is the same as in the single - user case , and there exists a cut - off power for splitting the total transmit power to both users . in the next section
, we will use these structural properties to solve the transmission completion minimization problem .
in this section , our goal is to minimize the transmission completion time of both users for a given @xmath13 .
the optimization problem can be formulated as @xmath119 where @xmath120 is the number of energy arrival epochs ( excluding @xmath0 ) over @xmath50 , and @xmath121 .
since @xmath122 depends on @xmath6 , the optimization problem in ( [ opt_prob3 ] ) is not a convex optimization problem in general .
therefore , we can not solve it using standard convex optimization tools .
we first note that this is exactly the `` dual '' problem of maximizing the departure region for fixed @xmath6 .
they are `` dual '' in the sense that , if the minimal transmission completion time for @xmath13 is @xmath6 , then @xmath13 must lie on the boundary of @xmath31 , and the transmission policy should be exactly the same for some @xmath123 .
this is based on the fact the @xmath124 for any @xmath125 .
assume @xmath13 does not lie on the boundary of @xmath31 .
then , either @xmath13 can not be achieved by @xmath6 or @xmath13 is strictly inside @xmath31 and hence @xmath13 can be achieved by @xmath126 .
therefore , if @xmath13 does not lie on the boundary of @xmath31 , then @xmath6 can not be the minimum transmission completion time .
we have the following lemma . [ lemma3 ]
when @xmath127 , under the optimal policy , the transmissions to both users must be finished at the same time .
this lemma can be proved based on corollary [ lemma2 ] .
if the transmission completion time for both users is not the same , then over the last duration , we transmit only to one of the users , while the transmission rate to the other user is zero .
this contradicts with the monotonicity of the transmission rates for both users .
therefore , under the optimal policy , the transmitter must finish transmitting to both users at the same time .
this lemma is proved in @xcite also , by using a different approach .
the authors prove it in @xcite mainly based on the convexity of the capacity region of the broadcast channel . for fixed @xmath13 ,
let us denote the transmission completion time for the first and second user , by @xmath128 and @xmath129 , respectively .
we note that @xmath128 and @xmath129 depend on the selection of the _ cut - off _ power level , @xmath130 . in particular , @xmath128 is monotonically decreasing in @xmath15 , and @xmath129 is monotonically increasing in @xmath15 . based on lemma
[ lemma3 ] , the problem of optimal selection of @xmath15 , can be viewed as solving a _
fixed point _ equation .
in particular , @xmath15 must be chosen such that , the resulting @xmath128 equals @xmath129 .
therefore , we propose the following algorithm to solve the transmission completion time , @xmath6 , minimization problem .
our basic idea is to try to identify the _ cut - off _ power level @xmath15 in an efficient way . since the power allocation is similar to the single - user case ( c.f .
lemma [ lemma4 ] ) , our approach to find @xmath6 will be similar to the method in @xcite .
first , we aim to identify @xmath14 , the first total transmit power starting from @xmath0 in the system .
this is exactly the same as identification of @xmath14 in the corresponding single - user problem . for this , as in @xcite
, we treat the energy arrivals as if they have arrived at time @xmath0 , and obtain a lower bound for the transmission completion time as in @xcite . in order to do that ,
first , we compute the amount of energy required to finish @xmath13 by @xmath131 .
this is equal to @xmath132 , denoted as @xmath133 .
then , we compare @xmath133 with @xmath134 .
if @xmath134 is greater than @xmath133 , this implies that the transmitter can finish the transmission before @xmath131 with @xmath134 , and future energy arrivals are not needed . in this case
, the minimum transmission completion time is the solution of the following equation @xmath135 if @xmath133 is greater than @xmath134 , this implies that the final transmission completion time is greater than @xmath131 , and some of the future energy arrivals must be utilized to complete the transmission .
we calculate the amount of energy required to finish @xmath13 by @xmath136 , @xmath137 , , and denote them as @xmath138 , @xmath139 , , and compare these with @xmath140 , , until the first @xmath141 that becomes smaller than @xmath142 .
we denote the corresponding time index as @xmath143 .
then , we assume that we can use @xmath144 to transmit @xmath13 at a constant rate . and
, the corresponding transmission completion time is the solution of the following equation @xmath145 we denote the solution to this equation as @xmath146 , and the corresponding power as @xmath147 . from our analysis , we know that the solution to this equation is the minimum possible transmission completion time we can achieve . then , we check whether this constant power @xmath147 is feasible , when the actual energy arrival times are imposed .
if it is feasible , it gives us the minimal transmission completion time ; otherwise , we get @xmath14 by selecting the minimal slope according to ( [ eqn : p_min ] ) .
that is to say , we draw all of the lines from @xmath0 to the corner points of the energy arrival instances before @xmath148 , and choose the line with the smallest slope .
we denote by @xmath149 the corresponding duration associated with @xmath14 .
this is shown in fig .
[ fig : bc_algr ] . once @xmath14 is selected , we know that it is the optimal total transmit power in our broadcast channel problem .
next , we need to divide this total power between the signals transmitted to the two users . based on lemma [ lemma : cutoff ] and corollary [ lemma2 ] , if the _ cut - off _ power level @xmath15 is higher than @xmath14 , then , the transmitter spends all @xmath14 for the stronger user ; otherwise , the first user finishes its transmission with a constant power @xmath15 .
we will first determine whether @xmath15 lies in @xmath150 $ ] or it is higher than @xmath14 .
assume @xmath151 .
therefore , the transmission completion time for the first ( stronger ) user is @xmath152 once @xmath15 is fixed , we can obtain the minimum transmission completion time for the second user , @xmath129 , by subtracting the energy consumed by the first user , and treating @xmath14 as an interference for the second user .
this reduces the problem to the single - user problem for the second user with fading , where the fading level is @xmath153 over @xmath154 , and @xmath21 afterwards .
the single - user problem with fading is studied in @xcite . since obtaining the minimal transmission completion time
is not as straightforward for the fading channel , a more approachable way is to calculate the maximum number of bits departed from the second user by @xmath128 , denoted as @xmath155 . in order to do that
, we first identify the optimal power allocation policy with fixed deadline @xmath128 .
this can be done according to lemma [ lemma4 ] .
assume that the optimal power allocation gives us @xmath156 .
then , we allocate @xmath14 to the first user over the whole duration , and allocate the remaining power to the second user . based on ( [ eqn : rate2 ] )
, we calculate the transmission rate for the second user over each duration , and obtain @xmath155 according to @xmath157 we observe that , given @xmath15 , @xmath155 is a monotonically increasing function of @xmath128 .
moreover , given @xmath128 , @xmath155 is a monotonically decreasing function of @xmath15 .
if @xmath155 is smaller than @xmath2 , it implies that @xmath158 , and we need to decrease the rate for the first user to make @xmath128 and @xmath129 equal .
based on lemma [ lemma : cutoff ] , this also implies that the transmission power for the first user is a constant @xmath159 .
in particular , @xmath15 is the unique solution of @xmath160 note that @xmath161 is a continuous , strictly monotonically decreasing function of @xmath15 , hence the solution for @xmath15 in ( [ eq ] ) is unique .
since @xmath128 is a decreasing function of @xmath15 and @xmath161 is a decreasing function of @xmath15 , we can use the bisection method to solve ( [ eq ] ) . in this case , the minimum transmission completion time is @xmath162 . if @xmath155 is larger than @xmath2 , that implies @xmath163 , and we need to increase the power allocated for the first user to make @xmath128 and @xmath129 equal , i.e. , @xmath164 . therefore , from lemma [ lemma : cutoff ] , over the duration @xmath165 , the optimal policy is to allocate the entire @xmath14 to the first user only .
we allocate @xmath14 to the first user , calculate the number of bits departed for the first user , and remove them from @xmath1 .
this simply reduces the problem to that of transmitting @xmath166 bits starting at time @xmath167 , where @xmath168 .
the process is illustrated in fig .
[ fig : search_pc ] .
then , the minimum transmission completion time is @xmath169 where @xmath170 is the number of recursions needed to get @xmath15 . in both scenarios ,
we reduce the problem into a simple form , and obtain the final optimal policy . before we proceed to prove the optimality of the algorithm
, we introduce the following lemma first , which is useful in the proof of the optimality of the algorithm .
[ lemma : mono ]
@xmath171 monotonically increases in @xmath6 ; @xmath172 monotonically increases in @xmath6 also .
the monotonicity of both functions can be verified by taking derivatives , @xmath173 and @xmath174 where the inequality follows since @xmath175 .
therefore , @xmath171 is a strictly concave function , and its first derivative monotonically decreases when @xmath6 increases .
since when @xmath176 , when @xmath177 , we have @xmath178 , therefore , the monotonicity follows . similarly , we have @xmath179 and @xmath180 again , the concavity implies that the first derivative is positive when @xmath177 , and the monotonicity follows .
the algorithm is feasible and optimal .
we first prove the optimality . in order to prove that the algorithm is optimal
, we need to prove that @xmath14 is optimal .
once we prove the optimality of @xmath14 , the optimality of @xmath181 , @xmath182 , @xmath183 follows .
since the solution obtained using our algorithm always has the optimal structure described in lemma [ lemma : cutoff ] , the optimality of the power allocation also implies the optimality of the rate selection , thus , the optimality of the algorithm follows .
therefore , in the following , we prove that @xmath14 is optimal .
first , we note that @xmath14 is the minimal slope up to @xmath148 .
we need to prove that @xmath14 is also the minimal slope up to the final transmission completion time , @xmath6 .
let us define @xmath184 as follows @xmath185 assume that with @xmath147 , we allocate @xmath186 to the first user , and finish @xmath13 using constant rates
. then , we allocate @xmath187 to the first user , and the rest to the second user .
based on lemma [ lemma : mono ] , we have @xmath188 therefore , @xmath184 is an upper bound for the optimal transmission completion time . since @xmath14 is the minimal slope up to @xmath184 ,
we conclude that @xmath14 is optimal throughout the transmission . following similar arguments
, we can prove the optimality of the rest of the power allocations .
this completes the proof of optimality . in order to prove that the allocation is feasible , we need to show that the power allocation for the first user is always feasible in each step . therefore , in the following
, we first prove that @xmath14 is feasible when we assume that @xmath151 .
the feasibility of @xmath14 also implies the feasibility of the rest of the power allocation . with the assumption that @xmath151 , the final transmission time for the first user is @xmath189 based on ( [ eqn : b1 ] ) and ( [ eqn : b2 ] ) , we know that @xmath190 .
since @xmath14 is feasible up to @xmath184 , therefore , @xmath14 is feasible when we assume that @xmath151 .
the feasibility of the rest of the power allocations follows in a similar way .
this completes the feasibility part of the proof .
we consider a band - limited awgn broadcast channel , with bandwidth @xmath191 mhz and the noise power spectral density @xmath192 w / hz . we assume that the path loss between the transmitter and the first receiver is about @xmath193 db , and the path loss between the transmitter and the second user is about @xmath194 db .
then , we have @xmath195 therefore , @xmath196 for the energy harvesting process , we assume that at times @xmath197 $ ] s , we have energy harvested with amounts @xmath198 $ ] mj . we find the maximum departure region @xmath31 for @xmath199 s , and plot them in fig . [ fig : bc_region ] .
we observe that the maximum departure region is convex for each value of @xmath6 , and as @xmath6 increases , the maximum departure region monotonically expands .
then , we aim to minimize the transmission completion time with @xmath200 mbits . following our algorithm
, we obtain the optimal transmission policy , which is shown in fig . [
fig : bc_example1 ] .
we note that the powers change only potentially at instances when energy arrives ( lemma [ lemma : const ] ) ; power sequence is monotonically increasing and `` majorized '' over the whole transmission duration ( lemma [ lemma4 ] ) .
we also note that , for this case , the first user transmits at a constant rate , and the rate for the second user monotonically increases .
the transmitter finishes its transmissions to both users by time @xmath201 s , and the last energy harvest at time @xmath202 s is not used .
next , we consider the example when @xmath203 mbits , we have the optimal transmission policy , as shown in fig .
[ fig : bc_example2 ] . in this example , the cut - off power is greater than @xmath14 , and therefore , @xmath14 is allocated to the first user only over @xmath204 s , and after @xmath205 s , the first user keeps transmitting at a constant rate until all bits are transmitted . in this case , the transmission rates for both users monotonically increase .
the transmitter finishes its transmissions by time @xmath206 s , and the last energy harvest is not used .
we investigated the transmission completion time minimization problem in an energy harvesting broadcast channel .
we first analyzed the structural properties of the optimal transmission policy , and proved that the optimal total transmit power has the same structure as in the single - user channel .
we also proved that there exists a _ cut - off _ power for the stronger user .
if the optimal total transmit power is lower than this cut - off level , all power is allocated to the stronger user , and when the optimal total transmit power is greater than this cut - off level , all power above this level is allocated to the weaker user .
based on these structural properties of the optimal policy , we developed an iterative algorithm to obtain the globally optimal off - line transmission policy .
a. el gamal , c. nair , b. prabhakar , e. uysal - biyikoglu , and s. zahedi , `` energy - efficient scheduling of packet transmissions over wireless networks , '' _ ieee infocom _ , vol . 3 , pp .
17731782 , november 2002 . | in this paper , we investigate the _ transmission completion time _ minimization problem in a two - user additive white gaussian noise ( awgn ) broadcast channel , where the transmitter is able to harvest energy from the nature , using a rechargeable battery .
the harvested energy is modeled to arrive at the transmitter randomly during the course of transmissions .
the transmitter has a fixed number of packets to be delivered to each receiver .
our goal is to minimize the time by which all of the packets for both users are delivered to their respective destinations . to this end , we optimize the transmit powers and transmission rates intended for both users . we first analyze the structural properties of the optimal transmission policy .
we prove that the optimal _ total _ transmit power has the same structure as the optimal single - user transmit power @xcite .
we also prove that there exists a _ cut - off _ power level for the stronger user .
if the optimal total transmit power is lower than this cut - off level , all transmit power is allocated to the stronger user , and when the optimal total transmit power is larger than this cut - off level , all transmit power above this level is allocated to the weaker user . based on these structural properties of the optimal policy , we propose an algorithm that yields the globally optimal off - line scheduling policy . our algorithm is based on the idea of reducing the two - user broadcast channel problem into a single - user problem as much as possible .
energy harvesting , rechargeable wireless networks , broadcast channels , transmission completion time minimization , throughput maximization . |
the performance of @xmath2 superconducting materials@xcite can be greatly enhanced by the addition of small amounts of carbon that will raise the upper critical magnetic field , @xmath0,@xcite and the critical current density , @xmath7.@xcite point defects like carbon that substitute for boron in the host lattice have been shown to raise @xmath16 from @xmath17 in pure @xmath2 to @xmath18 for a carbon level of @xmath19.@xcite at higher carbon contents , @xmath20 rises more slowly and eventually drops to about @xmath21 at @xmath22 carbon.@xcite several different precipitate phases have been used to form small pinning sites in the @xmath2 lattice . for example , @xmath23,@xcite @xmath24,@xcite , @xmath6@xcite and @xmath25@xcite have been added and show a rise of @xmath7 into the range of @xmath26 at @xmath27 and self field with about @xmath28 precipitate additions . in many of these experiments , either a powder - in - tube @xmath29 process or a pressed and sintered pellet method have been used to form the precipitates for the synthesis . in a different approach to sample preparation , a chemical vapor deposition , @xmath30 , method@xcite can be used to co - deposit @xmath8 together with a doping element to form long lengths of carbon doped boron fiber.@xcite subsequent heat treatment in @xmath11 vapor transforms the doped boron into doped @xmath2.@xcite both the @xmath30 and powder methods have advantages .
the advantage of powder methods is that the diffusion lengths are comparable to the powder size giving relatively low reaction temperatures and short reaction times .
the advantage of co - depositing the impurity with the @xmath8 in a @xmath30 process is that the impurity is mixed with the @xmath8 on an atomic scale .
both @xmath3@xcite and @xmath4@xcite separately have been successfully doped into @xmath2 using these @xmath30 methods .
the purpose of the work reported here is to study the combined @xmath4 and @xmath3 doping of @xmath2 to determine whether samples can be prepared with the combined benefits of both dopants , raising @xmath31 to the range of @xmath32 and raising @xmath7 values to the range of @xmath33 at @xmath34 and @xmath35 .
two specific questions need to be addressed .
does the addition of @xmath3 to the @xmath4-doped @xmath2 samples reported earlier,@xcite reduce the amount of @xmath4 in the @xmath2 lattice , possibly by forming @xmath36 precipitates ? in the presence of @xmath4 , can the @xmath3 precipitate size be maintained at @xmath37 for high @xmath7 performance ?
long lengths of doped boron fiber are prepared in a @xmath30 apparatus similar to that used for commercial boron filament production.@xcite all of the materials reported here were deposited on a @xmath38 fiber with an initial diameter of about @xmath39 .
the @xmath38 enters the reaction chamber moving at a few @xmath40 through a @xmath41 seal into a long glass tube containing flowing @xmath42 , @xmath43 , @xmath44 , and @xmath45 . for @xmath3 doping
, @xmath42 is bubbled through liquid @xmath45 to provide @xmath45 molecules in the gas stream .
typically the @xmath43 flow rate was @xmath46 with the @xmath44 and @xmath45 flow rates adjusted for the desired doping levels .
the diameter of the exiting doped @xmath8 fibers were in the @xmath47 to @xmath48 range .
the doped @xmath8 fibers were cut to lengths of @xmath49 and placed in a @xmath50 tube with a @xmath11 to @xmath8 ratio of @xmath51 .
the @xmath50 was welded shut , sealed in a quartz tube , and heat treated in a box furnace for appropriate periods . upon removal from the furnace
, ampoules were quenched in water .
resistivity measurements are made by a four contact method using silver epoxy to make the electrical contacts .
a quantum design @xmath52 system was used to make resistivity vs temperature , @xmath53 , measurements up to @xmath54 . at higher fields ,
@xmath55 measurements are made in a @xmath56 copper coil magnet at the national high magnetic field laboratory at florida state university .
magnetization measurements were made in a quantum design squid magnetometer with a magnetic field range of @xmath57 and @xmath7 was determined from the magnetization hysteresis using the bean model for cylindrical shell samples.@xcite a philips cm30 transmission electron microscope ( @xmath58 ) was employed for the microstructure characterization .
samples for the @xmath58 were made using a crush - flow technique .
in an earlier publication,@xcite we investigated a series of @xmath3-doped @xmath2 samples that were deposited on a commercial carbon coated @xmath23 fiber substrate with a diameter of approximately @xmath59 instead of the @xmath60 diameter @xmath38 substrate used here .
the doped-@xmath2 layers ranged from about @xmath61 to @xmath62 thick .
flux pinning was excellent and gave @xmath7 over @xmath63 at @xmath64 and @xmath65 for a sample with an average doping of about @xmath66.@xcite unfortunately , there were problems with this approach to sample preparation .
first , the diffusion of @xmath11 into the @xmath8 caused swelling and the @xmath2 pulled away from the substrate .
second , the presence of a carbon coated substrate made the amount of @xmath4 in the specimen uncertain .
in addition , these samples had a rather inhomogeneous @xmath3 distribution .
there was a high @xmath3 level near the @xmath23 core and near the outer surface of the fiber.@xcite transmission electron microscope @xmath67 observations for one of these samples that had been reacted for @xmath68 at @xmath69 are shown in fig .
results reveal a random distribution of precipitates scattered through the grain , ranging in size from @xmath70 to @xmath71 with a spacing about 5 times as large .
energy dispersive spectroscopy for a large portion of the grain in fig .
1a , revealed that the @xmath72 ratio was @xmath73 .
selected area diffraction taken along the c - axis showed the prominent hexagonal pattern of @xmath2 with additional powder pattern rings arising from the titanium boride precipitates .
the beam was then tipped off axis to show the rings more prominently .
indexing the rings revealed that the precipitates were @xmath5 rather than @xmath6 .
these results differ from those found by zhao and coworkers.@xcite they found that samples prepared by mixing powders of @xmath11 , @xmath3 , and @xmath8 and reacting gave @xmath6 precipitates on the @xmath2 grain boundaries .
this illustrates that different sample preparation methods can yield different types and locations of the precipitates .
both the precipitation of @xmath5 throughout the grains@xcite and precipitation of @xmath6 on the grain boundaries@xcite seem to give enhanced flux pinning .
samples with @xmath60 diameter tungsten cores were prepared using three different @xmath45 flow rates in the @xmath30 reactor , @xmath74 , @xmath75 and @xmath76 .
after the @xmath30 deposition , the fibers were reacted in @xmath11 vapor to form @xmath3-doped @xmath2 .
energy dispersive spectra ( eds ) in a scanning electron microscope ( sem ) was used to probe the uniformity of @xmath3 distribution across the superconducting fiber and to measure the @xmath3 to @xmath11 ratio , @xmath77/[ti+mg]$ ] .
line scans in various regions of the sample show that the @xmath3 level was uniform to about @xmath78 of the average value .
multiple point scans and area scans in the eds give average values of the @xmath77/[ti+mg]$ ] ratio , shown in fig .
2 indicating that the percent @xmath3 is roughly linearly related to the flow rate in the reaction chamber .
point scans , line scans , and large area raster scans were used for analysis . the large area raster scan shown by the open squares on fig .
2 comprises the most data and are probably the most accurate measure of the @xmath3 content .
these three flow rates of @xmath74 , @xmath75 , and @xmath76 , give @xmath79 , @xmath80 and @xmath81 respectively for the @xmath77/[ti+mg]$ ] ratio . because these @xmath3 doped samples were deposited on a @xmath60 diameter @xmath38 wire and had a @xmath82 outer diameter , these fibers require much longer times or higher temperatures to fully form the @xmath2 phase than the @xmath83 to @xmath62 @xmath8 layers of the earlier work.@xcite therefore , temperatures of @xmath84 to @xmath85 were often used .
a sample with @xmath86 reacted for @xmath87 at @xmath88 were found to be @xmath89 reacted using polarized light in an optical microscope .
a @xmath58 micrograph of this sample , fig .
1b , shows precipitates that are much larger than in fig .
in this figure , the sweeping shaded areas arise from the underlying holey carbon support . in this micrograph
, the beam has been tilted to emphasize areas of dense dislocations as shown , for example , in the top center of the micrograph .
many of the precipitates were @xmath90 to @xmath91 in diameter .
selected area diffraction indicated the precipitates were @xmath6 and had c - axes coaxial with the @xmath2 grains in which they were imbedded . for a sample with @xmath86 plus @xmath92 ,
the sample was fully reacted after @xmath93 at @xmath85 and shows the @xmath94 to @xmath95 diameter precipitates of fig .
the large precipitates here are @xmath6 with the c - axis parallel to the c - axis of the @xmath2 grain in which it is imbedded .
another series of samples having @xmath79 , @xmath80 and @xmath96 were all reacted at @xmath97 for @xmath93 and studied in the @xmath58 . for @xmath98 ,
the precipitate size is less than @xmath99 and widely spaced . for @xmath86 ,
the precipitate size ranges from @xmath94 to @xmath100 . for @xmath101 ,
the precipitate size ranges from @xmath94 to @xmath95 and the density of precipitates is correspondingly higher .
all of these samples show @xmath7 values higher than pure @xmath8 , but lower than shown by the sample of fig .
magnetization data , shown in fig .
3 , indicate that the suppression of @xmath102 with @xmath3 doping depends on the temperature at which the @xmath2 forms .
samples reacted at @xmath103 for @xmath87 have magnetization curves similar to pure @xmath2 except that they are shifted to lower temperature by about @xmath104 .
the @xmath105 and @xmath106 are nearly identical whereas the @xmath107 sample is a bit lower .
the depression is not monotonic in @xmath3 content , and the results on fig .
3 probably represent true scatter in the data .
the cause of the suppression in @xmath102 is not understood .
if the @xmath3 all precipitates as @xmath5 or @xmath6 and no @xmath3 is incorporated in the @xmath2 lattice , then it might be expected that the suppression of @xmath102 might be rather small and arise from lattice strains induced by the precipitates or other defects in the @xmath2 .
it is also possible that some @xmath3 atoms replace @xmath11 in the @xmath2 lattice and reduce the superconducting interaction in that way .
further study is going to be needed to determine which variables contribute to this suppression in @xmath102 .
at @xmath97 for @xmath93 , the suppression of @xmath102 behavior in fig .
3b is similar to fig .
3a except that the downward shift of @xmath102 is somewhat larger . at @xmath108 for @xmath109 , shown in fig .
3c , the suppression of @xmath102 is very large and increases monotonically with @xmath3 content .
because the @xmath110 curves are not always monotonic in the @xmath3 content , a new series of samples were made at @xmath97 to check reproducibility .
the data for the two sets of samples with the same @xmath3 content differed by as much as @xmath111 . with these large @xmath6
precipitates , there are some as yet uncontrolled parameters .
an x - ray study of the shift in the @xmath2 a - axis lattice constant was undertaken to look for a connection between the lattice constant an the amount of impurity .
for the case of carbon,@xcite the a - axis lattice constant contracts linearly with increasing @xmath4 , as shown by fig .
4 , while the c - axis remains essentially unchanged as reported by several authors.@xcite for the case of @xmath3 , the picture is more complicated . as shown in fig .
4 , the a - axis contracts with increasing @xmath3 , but there is considerable scatter in the data and the amount of change in the a - axis may depend on the temperature at which the @xmath2 is formed . an additional sample with @xmath112 flow rate of @xmath45 ( @xmath113 ) and reacted at @xmath85 for @xmath114 is shown by the open square of fig .
the apparent flat region of the a - axis lattice parameter between @xmath115 and @xmath116 as shown by the open squares would appear to indicate that a solubility limit has been reached for @xmath3 in @xmath2 at about @xmath117 .
for the case of @xmath3 additions only , samples with an @xmath118 for @xmath93 reaction temperature are used to illustrate the changes in superconducting properties .
the values of @xmath0 measured up to @xmath119 for the three different @xmath3 concentrations are slightly lower than for pure @xmath2 as shown on fig .
5 . as was found for values of @xmath110 curves in fig .
3b , the @xmath117 sample has the largest suppression of @xmath102 and the largest suppression of @xmath0 . as shown by fig .
6 , the @xmath7 values at @xmath34 for these three @xmath3 concentrations are enhanced by about a factor of @xmath120 at fields , up to about @xmath121 .
there is very little difference in @xmath7 as the @xmath3 level is raised from @xmath105 to about @xmath122 .
values of @xmath7 drop through @xmath123 at about @xmath121 even though @xmath20 is above @xmath124 at @xmath34 for these three samples as shown in fig . 5 .
all the samples reported here for the combined doping were reacted at @xmath85 to ensure full conversion to the @xmath2 phase .
the @xmath4 content is determined from the @xmath44 flow rate in the @xmath30 chamber and previous results.@xcite for a sample with @xmath86 plus @xmath92 shown in fig .
1c , the precipitates are @xmath94 to @xmath95 in diameter and selected area diffraction shows the precipitates to be @xmath6 with the c - axis parallel to the c - axis of the @xmath2 grain in which they are imbedded . in this micrograph , the beam is tilted so that the dislocations are not so apparent , but they are there .
these precipitates of fig .
1c are similar to those in fig . 1b , but in contrast to the @xmath5 precipitates of fig .
an x - ray analysis of the combined @xmath3 and @xmath4 doping is illustrated in fig .
the c - axis lattice parameter from the ( 002 ) peak essentially does not change with @xmath3 and @xmath4 addition .
the a - axis lattice parameter contracts in a regular way as shown by the ( 110 ) peak of fig .
the @xmath117-only peak moves to higher angle than the pure @xmath2 peak by about @xmath125 degree as shown by the heavy dotted line .
the @xmath126-only peak shifts out from the pure @xmath2 peak by about @xmath127 degree as shown by the light dotted line . and , the @xmath117+@xmath126 peak shifts out from the pure @xmath2 peak by about @xmath128 degree as shown by the heavy solid line .
roughly speaking , the decrease in a - axis lattice parameter is additive for @xmath3 and @xmath4 doping at this level .
as was reported earlier,@xcite the @xmath110 curves and the resulting @xmath102 values are depressed monotonically with increasing @xmath4 content as shown by the open squares and solid circles of fig .
if an additional @xmath117 is added to each of these @xmath4 concentrations , the combined suppression of @xmath102 is roughly additive . as shown in fig .
8 , the addition of @xmath129 to @xmath2 suppressed @xmath102 by about @xmath130 and the addition of @xmath129 plus @xmath107 suppressed @xmath102 by about @xmath131 .
values of @xmath20 in samples with the combined doping shown in fig .
9 by the open circles are very similar to values for @xmath4 doping only reported previously@xcite as shown by the solid symbols .
the solid triangles were taken with @xmath132 measurements and the solid circles were taken at the national high magnetic field laboratory as @xmath133 measurements .
for the combined @xmath3 and @xmath4 doped samples there was considerable rounding at the high field end of the @xmath133 transitions .
the two open circles represent two different definitions of @xmath31 , the lower being a linear extrapolation of the long linear region of @xmath133 up to the normal state , and the upper open circle being the field where the resistivity reaches the normal state value within the noise .
the addition of a few percent @xmath4 to the @xmath3 doped samples substantially raises @xmath31 ( open circles ) to values comparable to values for carbon only ( solid symbols ) .
values of @xmath7 for a series of samples reacted at @xmath134 are shown in fig . 10 .
for the combined @xmath126 + @xmath117 sample at @xmath14 , the @xmath7 curve crosses @xmath123 at about @xmath135 shown by the solid circles .
the @xmath136 + @xmath137 sample at @xmath14 crosses @xmath123 at @xmath138 , as shown by the open circles .
the @xmath86 only sample has the highest low field @xmath7 values and crosses @xmath123 at @xmath139 . a pure @xmath2 sample at measured at @xmath14 is shown by the solid squares .
it is not shown here , but the the combined @xmath126 + @xmath117 sample at @xmath34 , the @xmath7 curve crosses @xmath123 at about @xmath121 . to summarize , the addition of @xmath4 to raise @xmath0 and the addition of @xmath3 to form precipitate pinning centers are roughly independent of one another for the samples reported here .
carbon doped @xmath2 shows a rapid rise in @xmath16 from @xmath17 for pure @xmath2 to @xmath21 for @xmath140 added carbon . with both @xmath141 @xmath4 and @xmath142 the sample retains a @xmath143 of @xmath144 and no evidence for the formation of @xmath36 was seen .
the addition of @xmath3 enhances @xmath7 substantially in the magnetic field range of @xmath145 to @xmath146 , as shown in fig .
much work needs to be done to optimize the the @xmath4 and @xmath3 levels and the processing to raise @xmath7 at high fields . for low reaction temperatures and short times to form the @xmath2 phase ,
@xmath58 data show that the precipitates are intra - granular and randomly oriented particles ranging in size from @xmath147 to @xmath71 . for higher reaction temperatures in the range from @xmath84 to @xmath134 ,
the precipitates are much larger .
the precipitates are usually intra - granular @xmath6 particles coplanar with the @xmath2 host that range from @xmath99 to @xmath91 in size .
coarsening of the @xmath6 precipitates at high temperatures and long times is clearly a problem . in magnetic fields from zero to @xmath148
, the addition of carbon to @xmath3-doped @xmath2 gives relatively little change in @xmath7 , but in the @xmath149 to @xmath150 range , carbon additions clearly enhance @xmath7 .
a very practical problem is that the addition of either @xmath4 or @xmath3 slows down the rate at which the @xmath8 fibers transform to the @xmath2 phase . for high @xmath7 values
, it is helpful to react at low temperatures to give small @xmath2 grains and to prevent the @xmath3 precipitate coarsening .
some method is needed to overcome these slow reaction rates , probably the use of fine powders to keep the reaction time short and the reaction temperatures low .
it would be desirable to make doped boron powders so that the diffusion lengths can be much smaller and the reaction temperatures lower .
ames laboratory is operated for the us department of energy by iowa state university under contract no .
this work was supported by the director for energy research , office of basic energy sciences .
a portion of this work was performed at the national high magnetic field laboratory , which is supported by nsf cooperative agreement no .
dmr-0084173 and by the state of florida .
a. gurevich , s. patnaik , v. braccini , k. h. kim , c. mielke , x. song , l. d. cooley , s. dl .
bu , d. m. kim , j. h. choi , l. j. belenky , j. giencke , m. k. lee , w. tian , x. q. pan , a. siri , e. e. helstrom , c. b. eom , d. c. larbalestier , supercon .
* 17 * ( 2004 ) 278 .
. 1 @xmath58 micrograph of @xmath2 with @xmath151 reacted @xmath69-@xmath68 , @xmath4 coated @xmath23 substrate .
b ) @xmath2 with @xmath107 reacted @xmath103-@xmath87 , @xmath38 substrate .
c ) @xmath2 with @xmath152 reacted @xmath85-@xmath93 , @xmath38 substrate . | measurements of the superconducting upper critical field , @xmath0 , and critical current density , @xmath1 , have been carried out for @xmath2 doped with @xmath3 and/or @xmath4 in order to explore the problems encountered if these dopants are used to enhance the superconducting performance .
carbon replaces boron on the @xmath2 lattice and apparently shortens the electronic mean free path of @xmath2 and raising @xmath0 .
titanium forms precipitates of either @xmath5 or @xmath6 that enhance the flux pinning and raise @xmath7 .
most of these precipitates are intra - granular in the @xmath2 phase . for samples containing both @xmath4 and @xmath3 doping
, the @xmath4 appears to still replace @xmath8 in the @xmath2 lattice and the @xmath3 precipitates out as a boride . if approximately @xmath9 and approximately @xmath10 are co - deposited with b to form doped boron fibers and these fibers are in turn reacted in @xmath11 vapor to form doped @xmath2 , the resulting superconductor has @xmath12 and @xmath13 at @xmath14 and @xmath15 . |
the minimal supersymmetric model ( mssm ) is a popular extension of the standard model ( sm ) . in this model
the invariance of @xmath0-parity , defined by @xmath5 for a field with spin @xmath6 , baryon - number @xmath7 and lepton - number @xmath8 , is often imposed on the lagrangian in order to maintain the separate conservation of baryon - number and lepton - number .
although @xmath0-parity plays a crucial role in the phenomenology of the mssm ( e.g. , forbid proton decay and ensure a perfect candidate for cosmic dark matter ) , it is , however , not dictated by any fundamental principle such as gauge invariance and there is no compelling theoretical motivation for it .
the most general superpotential of the mssm consistent with the sm gauge symmetry and supersymmetry contains @xmath0-violating interactions which are given by @xcite @xmath9 where @xmath10 are generation indices , @xmath11 denotes charge conjugation , @xmath12 , @xmath13 and @xmath14 are the color indices with @xmath15 being the total antisymmetric tensor , @xmath16 is the higgs - doublet chiral superfield , and @xmath17 and @xmath18 are the left - handed lepton ( quark ) doublet and right - handed lepton ( quark ) singlet chiral superfields .
the dimensionless coefficients @xmath19 ( antisymmetric in @xmath20 and @xmath21 ) and @xmath22 in the superpotential are @xmath8-violating couplings , while @xmath23 ( antisymmetric in @xmath21 and @xmath24 ) are @xmath7-violating couplings .
so far both theorists and experimentalists have intensively studied the phenomenology of @xmath0-parity breaking supersymmetry in various processes @xcite and obtained some bounds @xcite .
the lepton flavor - changing ( lfc ) processes , which have been searched in various experiments @xcite , are a sensitive probe for new physics because they are extremely suppressed in the sm but can be greatly enhanced in new physics models like supersymmetry @xcite . in @xmath0-parity
breaking supersymmetry , these rare processes may receive exceedingly large enhancement since both @xmath25 and @xmath26 couplings can make contributions .
such enhancement was considered in the decays @xmath27 @xcite and @xmath28 @xcite , the @xmath29 conversion in nuclei @xcite , and the di - lepton productions @xmath30 @xcite and @xmath31 @xcite .
since the gigaz and photon - photon collision options of the ilc can precisely measure the lfc processes @xmath32 and @xmath33 ( @xmath34 and @xmath35 ) , we in this work study these processes in @xmath0-violating mssm . noting that the experimental upper bounds on the lfc @xmath36-decays became more stringent recently @xcite , we will first check the constraints on the relevant @xmath0-violating couplings from the latest measurement of @xmath37 . then , with the updated bounds on the relevant @xmath0-violating couplings , we calculate @xmath32 and @xmath38 to figure out if they can reach the sensitivity of the gigaz and photon - photon collision options of the ilc .
the paper is organized as follows . in sec .
ii we describe the calculations for @xmath37 , @xmath32 and @xmath33 . in sec .
iii we present some numerical results and discussions .
finally , a conclusion is drawn in sec .
in terms of the four - component dirac notation , the lagrangian of the @xmath8-violating interaction is given by ( in our calculations we take the presence of @xmath39 as an example ) @xmath40+h.c.\end{aligned}\ ] ] the lfc interactions @xmath41
( @xmath42 ) are induced at loop level by exchanging a squark @xmath43 or @xmath44 , which is shown in fig.1 . for the decays @xmath37
we take @xmath45 as an example to show the analytic results .
the gauge invariant amplitude of @xmath45 is given by @xmath46 where @xmath47 is given by ( assuming the degeneracy for squark masses ) @xmath48\end{aligned}\ ] ] with @xmath49 , \\
f_2(x)&= & \frac{1}{8(x-1)^3}\left[\frac{1}{3}(2x^2 + 5x-1-\frac{6x^2 \ln x}{x-1 } ) -\frac{2}{3}(x^2 - 5x-2+\frac{6x \ln x}{x-1})\right ] .\end{aligned}\ ] ] the decay branching ratio reads @xmath50 for the decays @xmath51 we calculate the decay rates numerically by using the effective vertex presented in appendix a. note that according to the effective vertex method @xcite , the external legs of the effective vertex can be on - shell or off - shell and thus the vertex can be used in any relevant process . the expression in eqs.(3 - 6 ) can be obtained from the effective vertex in appendix a by putting both leptons on shell . for the process @xmath52 , besides fig.2 ( a , b ) induced by the effective vertex given in appendix a , more diagrams shown in fig.2 ( c - i ) also come into play . the analytic expressions of the amplitudes of these diagrams are given in appendix b. these amplitudes contain the passarino - veltman one - loop functions , which are calculated by using looptools @xcite .
we checked that the amplitudes have gauge invariance and the ultraviolet divergence cancelled .
since the photon beams in @xmath53 collision are generated by the backward compton scattering of the incident electron- and the laser - beam , the events number is obtained by convoluting the cross section of @xmath53 collision with the photon beam luminosity distribution : @xmath54 where @xmath55/@xmath56 is the photon - beam luminosity distribution and @xmath57 ( @xmath58 is the squared center - of - mass energy of @xmath59 collision ) is defined as the effective cross section of @xmath60 . in optimum case , it can be written as @xcite @xmath61 where @xmath62 denotes the energy spectrum of the back - scattered photon for the unpolarized initial electron and laser photon beams given by @xmath63\end{aligned}\ ] ] with @xmath64 here @xmath65 ( @xmath66 is the incident electron energy and @xmath67 is the initial laser photon energy ) and @xmath68 with @xmath69 being the energy of the scattered photon moving along the initial electron direction .
in our calculations we take the sm parameters as @xcite @xmath70 the top quark mass is taken as the new cdf value @xmath71 @xcite .
the relevant susy parameters in our calculations are the masses of squarks as well as the @xmath0-parity violating couplings listed in table i. the strongest bound on squark mass is from the tevatron experiment .
for example , from the search of the inclusive production of squark and gluino in @xmath0-conserving minimal supergravity model with @xmath72 , @xmath73 and @xmath74 , the cdf gives a bound of 392 gev at the 95 @xmath75 c.l . for degenerate gluinos and squarks @xcite .
however , this bound may be not applicable to the @xmath0-violating scenario because the susy signal in case of @xmath0-violation is very different from the @xmath0-conserving case .
the most robust bounds on sparticle masses come from the lep results , which give a bound of about 100 gev on squark or slepton mass @xcite . in our numerical calculations
, we assume the presence of the minimal number of @xmath0-violating couplings , i.e. , for each process only the two relevant couplings ( not summed over the family indices ) are assumed to be present . for @xmath37 , the latest experimental data is @xcite @xmath76 we use these data to update the bounds on the relevant @xmath8-violating couplings .
the new bounds are compared with the old ones in table i for @xmath77 gev ( here we take squark mass of 100 gev for illustration and for heavier squarks the bounds on the @xmath8-violating couplings will become weak , as will be shown later ) .
we can see that the new bounds are much stronger than the old ones .
since the bounds on @xmath78 ( @xmath34 ) are weakest , we only consider the contribution of @xmath78 in our following numerical calculations .
.our new upper bounds on the @xmath8-violating couplings for @xmath77 gev from @xmath37 data @xcite , in comparison with the old ones @xcite .
[ cols="<,<,<",options="header " , ] [ dlambda ] note that the neutrino masses could also constrain the @xmath79 couplings , especially @xmath80 @xcite .
but these constraints depend on more parameters in addition to the squark mass .
for example , the one - loop @xmath79 contributions to the neutrino masses are sensitive to the left - right squark mixings and the two - loop contributions further involve the slepton mass . for small squark mixings with appropriate sign , there may exist a strong cancellation between one - loop and two - loop effects , and in this case , the constraints from the neutrino masses can be avoided .
since the aim of our study is the sensitivity of the lfc processes to @xmath79 couplings and the @xmath79 contributions to these lfc processes are irrelevant to the additional parameters involved in the contributions to the neutrino masses , in our analysis we did not consider such constraints from the neutrino masses . for @xmath32 ,
the upper limits from lep are @xcite @xmath81 the bounds from these lep data are compared with the bounds from @xmath37 in fig.3 .
one can see that the upper bounds on the couplings from the lep @xmath2-decay data @xcite are weaker than the ones from @xmath37 data @xcite .
note that the bounds from the lep @xmath2-decay data were also studied in @xcite and our results are consistent with theirs except that in @xcite the sum over index @xmath24 is implied for @xmath82 with @xmath83 gev .
the possible sensitivity of gigaz to the lfc decays of @xmath2-boson could reach @xcite @xmath84 with the factor @xmath85 ranging from 0.2 to 1.0 . in fig .
3 we take @xmath86 to show the sensitivity .
in contrast to the @xmath0-conserving case in which only @xmath87 is accessible at the gigaz @xcite , the @xmath0-violating couplings under the bound from @xmath27 can still enhance all the channels @xmath3 to the sensitivity of the gigaz .
this implies that the gigaz can further strengthen the bounds on @xmath88 in case of un - observation .
these bounds , unlike the constraints from neutrino masses which involve more parameters , are only dependent on the squark mass . for the @xmath89 collision results shown in fig .
3 , we fixed the parameters as @xmath90 , @xmath91 and @xmath92 @xcite . since the @xmath8-violating couplings relevant to the process @xmath93 is stringently constrained by @xmath45 , we in fig .
3 only show the results for the channels with a tau lepton in the final states , i.e. , @xmath94 .
the background for @xmath95 comes from @xmath96 , @xmath97 and @xmath98 , and we make kinematical cuts @xcite : @xmath99 and @xmath100 ( @xmath101 ) , to enhance the ratio of signal to background . with these cuts , the background cross sections from @xmath102 , @xmath103 and @xmath98 at @xmath104 gev
are suppressed respectively to @xmath105 fb , @xmath106 fb and @xmath107 fb ( see table i of @xcite ) .
to get the @xmath108 observing sensitivity with @xmath109 fb@xmath110 integrated luminosity @xcite , the production rates of @xmath111 after the cuts must be larger than @xmath112 fb @xcite .
we see from fig .
3 that under the current bounds from @xmath27 , the @xmath8-violating couplings can still be large enough to enhance the productions @xmath113 to the @xmath108 sensitivity . we also show the cross sections of @xmath114 as a function of center - of - mass energy @xmath115 of the ilc in fig.4 .
we see that with the increasing of the center - of - mass energy , the cross sections of these processes become smaller .
such a behavior is similar to the results in the r - conserving mssm shown in @xcite .
finally , we point out that the lfc processes can also put bounds on the products @xmath116 and @xmath117 , and our numerical results indicate that such bounds are quite similar to those in fig.3 .
we note that these bounds on @xmath118 and @xmath119 from @xmath120 at gigaz are generally stronger than those from the neutrino masses @xcite .
we evaluated the lepton flavor - changing processes in @xmath0-parity violating mssm .
first , we used the latest data on the rare decays @xmath1 to update the constraints on the relevant @xmath0-violating couplings .
then we calculated the processes @xmath32 and @xmath4 .
we found that with the updated constraints the @xmath0-violating couplings can still enhance the rates of these processes to the sensitivity of gigaz and photon - photon collision options of the ilc .
so , the gigaz and photon - photon collision of the ilc can either observe these @xmath79-induced lfc processes or further strengthen the bounds on the @xmath79 couplings in case of un - observation .
this work was supported in part by the national natural science foundation of china ( nnsfc ) under grant nos .
10505007 , 10821504 , 10725526 and 10635030 , and by hastit under grant no . 2009hastit004 .
here we list the expression for the @xmath8-violating contributions to the effective vertex @xmath122 . other effective vertices @xmath123 and @xmath124 are similar to @xmath122 and can be obtained by replacing the corresponding momentum and mass . the the effective vertex @xmath122 is given by @xmath125 where the two terms on the right side denote the @xmath8-violating loop contributions by exchanging respectively the squarks @xmath126 and @xmath127 , given by @xmath128[l]{$p$}/}}_{\mu}-{{\makebox[.2ex][l]{$p$}/}}_{e})\gamma_{\lambda } \gamma^\alpha ] p_l + \frac{2}{3}[2c^2_{\alpha\lambda}\gamma^\alpha -c^2_{\alpha}(p_{\mu}+p_{e})_\lambda\gamma^\alpha ] p_l\nonumber\\ & & -\frac{1}{m_{\mu}^2}\gamma^\lambda { { \makebox[.2ex][l]{$p$}/}}_{\mu}\gamma^\alpha b^1_{\alpha}p_l + \frac{1}{m_{\mu}^2}[\gamma^\alpha { { \makebox[.2ex][l]{$p$}/}}_{e } \gamma_\lambda p_l + m_{\mu}\gamma^\alpha \gamma_\lambda p_r)]b^2_{\alpha } -\frac{1}{3 } m^{2}_{d_{k}}c^1_{0}\gamma_\lambda p_l \biggl\}\\ \gamma_{\lambda}^{\gamma e \mu}(p_{\mu},p_{e})|_{\tilde { d}_k^r}&= & ae\biggl\{-\frac{2}{3 } [ c^3_{\alpha\beta}\gamma^\alpha\gamma_\lambda\gamma^\beta -c^3_\alpha ( { { \makebox[.2ex][l]{$p$}/}}_{\mu } -{{\makebox[.2ex][l]{$p$}/}}_{e } ) \gamma_{\lambda } \gamma^\alpha ] p_l + \frac{1}{3}[2c^4_{\alpha\lambda}\gamma^\alpha -c^4_{\alpha}(p_{\mu}+p_{e})_\lambda\gamma^\alpha ] p_l\nonumber\\ & & -\frac{1}{m_\mu^2}\gamma^\lambda { { \makebox[.2ex][l]{$p$}/}}_{\mu}\gamma^\alpha b^3_{\alpha}p_l + \frac{1}{m_{\mu}^2}[\gamma^\alpha { { \makebox[.2ex][l]{$p$}/}}_{e } \gamma_\lambda p_l + m_{\mu}\gamma^\alpha \gamma_\lambda p_r)]b^4_{\alpha } -\frac{2}{3}m^{2}_{u_{j}}c_{0}^{3}\gamma_{\lambda } p_{l } \biggl\ } \\ \gamma_{\lambda}^{z e \mu}(p_{\mu},p_{e})|_{\tilde { u}_j^l}&= & be\biggl\{\frac{2s_w^2}{3 } [ c^5_{\alpha\beta}\gamma^\alpha\gamma_\lambda\gamma^\beta -c^5_\alpha ( { { \makebox[.2ex][l]{$p$}/}}_{\mu } -{{\makebox[.2ex][l]{$p$}/}}_{e } ) \gamma_{\lambda } \gamma^\alpha ] p_l-(1-\frac{2s_w^2}{3 } ) m_{d_k}^2c^5_0\gamma_\lambda p_l\nonumber\\ & & + ( 1-\frac{4s_w^2}{3})[2c^6_{\alpha\lambda}\gamma^\alpha -c^6_{\alpha}(p_{\mu}+p_{e})_\lambda\gamma^\alpha ] p_l -\frac{1 - 2s_w^2}{m_{\mu}^2}\gamma^\lambda { { \makebox[.2ex][l]{$p$}/}}_{\mu } \gamma^\alpha b^5_{\alpha}p_l\nonumber\\ & & + \frac{1}{m_{\mu}^2}[(1 - 2s_w^2)\gamma^\alpha { { \makebox[.2ex][l]{$p$}/}}_{e}\gamma_\lambda p_l - 2s_w^2m_{\mu}\gamma^\alpha \gamma_\lambda p_r)]b^6_{\alpha}\biggl\}\\ \gamma_{\lambda}^{z e \mu}(p_{\mu},p_{e})|_{\tilde { d}_k^r}&= & be\biggl\{-(1-\frac{4s_w^2}{3 } ) [ c^7_{\alpha\beta}\gamma^\alpha\gamma_\lambda\gamma^\beta -c^7_\alpha ( { { \makebox[.2ex][l]{$p$}/}}_{\mu } -{{\makebox[.2ex][l]{$p$}/}}_{e } ) \gamma_{\lambda } \gamma^\alpha ] p_l+\frac{4s_w^2}{3}m_{u_j}^2c^7_0 \gamma_\lambda p_l\nonumber\\ & & -\frac{2s_w^2}{3}[2c^8_{\alpha\lambda}\gamma^\alpha -c^8_{\alpha}(p_{\mu}+p_{e})_\lambda\gamma^\alpha ] p_l -\frac{1 - 2s_w^2}{m_{\mu}^2}\gamma^\lambda { { \makebox[.2ex][l]{$p$}/}}_{\mu } \gamma^\alpha b^7_{\alpha}p_l\nonumber\\ & & + \frac{1}{m_{\mu}^2}[(1 - 2s_w^2)\gamma^\alpha { { \makebox[.2ex][l]{$p$}/}}_{e } \gamma_\lambda p_l - 2s_w^2m_{\mu}\gamma^\alpha \gamma_\lambda p_r)]b^8_{\alpha } \biggl\}\end{aligned}\ ] ] with @xmath129 , @xmath130 and @xmath131 and @xmath132 denoting respectively the momenta of the electron and muon . in the above expressions ,
the functions @xmath133 and @xmath134 are the passarino - veltman functions . for these loop functions
, we adopt the definition in @xcite and use looptools @xcite in the calculations .
the functional dependence of these loop functions is given by @xmath135
the amplitudes of the diagrams in fig.2(a - i ) are given by @xmath137[l]{$p$}/}}_{2}-{{\makebox[.2ex][l]{$p$}/}}_{\mu})}\gamma_{\rho}^{\gamma e \mu}(-p_{\mu},p_{2}-p_{\mu})|_{\tilde { u}_j^l,{\tilde{d}_k^r}}v(\mu)\epsilon^{\lambda}_{1}\epsilon^{\rho}_{2}\\ \label{b2 } m_{(b)}|_{\tilde { u}_j^l,{\tilde{d}_k^r}}&= & \overline{u}(e)\gamma_{\lambda}^{\gamma e \mu}(p_{2}-p_{\mu},p_e)|_{\tilde { u}_j^l,{\tilde{d}_k^r}}\frac{i}{({{\makebox[.2ex][l]{$p$}/}}_{2}-{{\makebox[.2ex][l]{$p$}/}}_{\mu})}(ie\gamma_{\rho})v(\mu)\epsilon^{\lambda}_{1}\epsilon^{\rho}_{2}\\ m_{(c)}|_{\tilde { u}_j^l}&= & -\frac{i}{16\pi^{2}}(\frac{8}{9}e^{2})\lambda^{'}_{1jk}\lambda^{'}_{3jk}\overline{u}(e)c^{9}_{\alpha}\gamma^{\alpha}p_{l}v(\mu)\epsilon_{1}\cdotp\epsilon_{2}\\ m_{(c)}|_{\tilde{d}_k^r}&= & \frac{i}{16\pi^{2}}(\frac{2}{9}e^{2})\lambda^{'}_{1jk}\lambda^{'}_{3jk}\overline{u}(e)c^{10}_{\alpha}\gamma^{\alpha}p_{l}v(\mu)\epsilon_{1}\cdotp\epsilon_{2}\\ m_{(d)}|_{\tilde { u}_j^l}&=&\frac{i}{16\pi^{2}}(\frac{1}{9}e^{2})\lambda^{'}_{1jk}\lambda^{'}_{3jk}\overline{u}(e)\biggl\ { d^{1}_{\alpha\beta\delta}\gamma^{\alpha}\gamma_{\rho}\gamma^{\beta}\gamma_{\lambda}\gamma^{\delta}+d^{1}_{\alpha\beta}\gamma^{\alpha}\gamma_{\rho}{{\makebox[.2ex][l]{$p$}/}}_{2}\gamma_{\lambda}\gamma^{\beta}\nonumber\\&&+d^{1}_{\alpha\beta}({{\makebox[.2ex][l]{$p$}/}}_{1}+{{\makebox[.2ex][l]{$p$}/}}_{2})\gamma_{\rho}\gamma^{\alpha}\gamma_{\lambda}\gamma^{\beta}+d^{1}_{\alpha}({{\makebox[.2ex][l]{$p$}/}}_{1}+{{\makebox[.2ex][l]{$p$}/}}_{2})\gamma_{\rho}{{\makebox[.2ex][l]{$p$}/}}_{2}\gamma_{\lambda}\gamma^{\alpha}+m^{2}_{d_k}d^{1}_{\alpha}(\gamma^{\alpha}\gamma_{\rho}\gamma_{\lambda } \nonumber\\ & & + \gamma_{\rho}\gamma_{\lambda}\gamma^{\alpha}+\gamma_{\rho}\gamma^{\alpha}\gamma_{\lambda})+m^{2}_{d_k}d^{1}_{0}[({{\makebox[.2ex][l]{$p$}/}}_{1}+{{\makebox[.2ex][l]{$p$}/}}_{2})\gamma_{\rho}\gamma_{\lambda}+\gamma_{\rho}{{\makebox[.2ex][l]{$p$}/}}_{2}\gamma_{\lambda } ] \biggl\}p_{l}v(\mu)\epsilon^{\lambda}_{1}\epsilon^{\rho}_{2 } , \\ m_{(e)}|_{\tilde{u}_j^l}&=&\frac{i}{16\pi^{2}}(\frac{4}{9}e^{2})\lambda^{'}_{1jk}\lambda^{'}_{3jk}\overline{u}(e)\biggl\{4d^{2}_{\rho\lambda\alpha}-2d^{2}_{\rho\alpha}(2p_e-2p_1-p_2)_{\lambda}\nonumber\\&&-2d^{2}_{\lambda\alpha}(2p_e - p_1)_{\rho}+d^{2}_{\alpha}(2p_e - p_1)_{\rho}(2p_e-2p_1-p_2 ) \biggl\}\gamma^{\alpha}p_{l}v(\mu)\epsilon^{\lambda}_{1}\epsilon^{\rho}_{2 } \\ m_{(f)}|_{\tilde{u}_j^l}&=&\frac{i}{16\pi^{2}}(-\frac{2}{9}e^{2})\lambda^{'}_{1jk}\lambda^{'}_{3jk}\overline{u}(e)\biggl\{[2d^{3}_{\alpha\beta\lambda}-d^{3}_{\alpha\beta}(2p_e - p_2)_{\lambda}]\gamma^{\alpha}\gamma_{\rho}\gamma^{\beta}-[2d^{3}_{\alpha\lambda}\nonumber\\&&-d^{3}_{\alpha}(2p_e - p_2)_{\lambda}]\gamma^{\alpha}\gamma_{\rho}{{\makebox[.2ex][l]{$p$}/}}_{1}+m^{2}_{d_k}[2d^{3}_{\lambda}-d^{3}_0(2p_e - p_2)_{\lambda}]\gamma_{\rho } \biggl\}p_{l}v(\mu)\epsilon^{\lambda}_{1}\epsilon^{\rho}_{2}\end{aligned}\ ] ] @xmath138[l]{$p$}/}}_{2}\gamma_{\lambda}\gamma^{\beta}\nonumber\\&&+d^{4}_{\alpha\beta}({{\makebox[.2ex][l]{$p$}/}}_{1}+{{\makebox[.2ex][l]{$p$}/}}_{2})\gamma_{\rho}\gamma^{\alpha}\gamma_{\lambda}\gamma^{\beta}+d^{4}_{\alpha}({{\makebox[.2ex][l]{$p$}/}}_{1}+{{\makebox[.2ex][l]{$p$}/}}_{2})\gamma_{\rho}{{\makebox[.2ex][l]{$p$}/}}_{2}\gamma_{\lambda}\gamma^{\alpha}+m^{2}_{d_k}d^{4}_{\alpha}(\gamma^{\alpha}\gamma_{\rho}\gamma_{\lambda}\nonumber\\&&+\gamma_{\rho}\gamma_{\lambda}\gamma^{\alpha}+\gamma_{\rho}\gamma^{\alpha}\gamma_{\lambda})+m^{2}_{d_k}d^{4}_{0}[({{\makebox[.2ex][l]{$p$}/}}_{1}+{{\makebox[.2ex][l]{$p$}/}}_{2})\gamma_{\rho}\gamma_{\lambda}+\gamma_{\rho}{{\makebox[.2ex][l]{$p$}/}}_{2}\gamma_{\lambda } ] \biggl\}p_{l}v(\mu)\epsilon^{\lambda}_{1}\epsilon^{\rho}_{2}\\ m_{(h)}|_{\tilde{d}_k^r}&=&\frac{i}{16\pi^{2}}(-\frac{1}{9}e^{2})\lambda^{'}_{1jk}\lambda^{'}_{3jk}\overline{u}(e)\biggl\{4d^{5}_{\rho\lambda\alpha}-2d^{5}_{\rho\alpha}(2p_e-2p_1-p_2)_{\lambda}\nonumber\\&&-2d^{5}_{\lambda\alpha}(2p_e - p_1)_{\rho}+d^{5}_{\alpha}(2p_e - p_1)_{\rho}(2p_e-2p_1-p_2 ) \biggl\}\gamma^{\alpha}p_{l}v(\mu)\epsilon^{\lambda}_{1}\epsilon^{\rho}_{2}\\ m_{(i)}|_{\tilde{d}_k^r}&=&\frac{i}{16\pi^{2}}(-\frac{2}{9}e^{2})\lambda^{'}_{1jk}\lambda^{'}_{3jk}\overline{u}(e)\biggl\{[2d^{6}_{\alpha\beta\lambda}-d^{6}_{\alpha\beta}(2p_e - p_2)_{\lambda}]\gamma^{\alpha}\gamma_{\rho}\gamma^{\beta}-[2d^{6}_{\alpha\lambda}\nonumber\\ & & -d^{6}_{\alpha}(2p_e - p_2)_{\lambda}]\gamma^{\alpha}\gamma_{\rho}{{\makebox[.2ex][l]{$p$}/}}_{1}+m^{2}_{d_k}[2d^{6}_{\lambda}-d^{6}_0(2p_e - p_2)_{\lambda}]\gamma_{\rho } \biggl\}p_{l}v(\mu)\epsilon^{\lambda}_{1}\epsilon^{\rho}_{2 } .\end{aligned}\ ] ] here the effective vertices appearing in eqs.([b1 ] ) and ( [ b2 ] ) are defined in appendix a. the amplitudes for the diagrams with the two photons exchanged are not presented here , which can be obtained from the above corresponding amplitudes with replacement @xmath139 and @xmath140 .
the functional dependence of the passarino - veltman loop functions @xmath141 and @xmath142 is given by @xmath143 99 for some early works on @xmath0-violating supersymmetry , see , e.g. , l. hall and m. suzuki , 231 , 419 ( 1984 ) ; j. ellis _
et al . _ , 150 , 142 ( 1985 ) ; g. ross and j. valle , 151 , 375 ( 1985 ) ; s. dawson , 261 , 297 ( 1985 ) ; r. barbieri and a. masiero , 267 , 679 ( 1986 ) ; h. dreiner and g.g .
ross , 365 , 597 ( 1991 ) ; j. butterworth and h. dreiner , 397 , 3 ( 1993 ) . for phenomenology of @xmath0-violation in some decays , see , e.g. , v. barger , g. f. giudice and t. han , 40 , 2978 ( 1989 ) ; k. agashe , m. graesser , 54 , 4445 ( 1996 ) ; f. zwirner , 132 , 103 ( 1983 ) ; r. n. mohapatra , 34 , 3457 ( 1986 ) ; m. hirsch , h. kleingrothaus , s. g. kovalenko , 75 , 17 ( 1995 ) ; k. s. babu , r. n. mohapatra , 75 , 2276 ( 1995 ) ; g. bhattacharyya , d. choudhury , mod .
a**10 , 1699 ( 1995 ) ; d. e. kaplan , hep - ph/9703347 ; j. jang , j. k. kim , j. s. lee , 55 , 7296 ( 1997 ) ; g. bhattacharyya , a. raychaudhuri,57 , 3837 ( 1998 ) ; j. m. yang , b .-
young , x. zhang , 58 , 055001 ( 1998 ) ; c. h. chang , t. f. feng , l. y. shan , commun .
33 , 421 ( 2000 ) ; s. bar - shalom , g. eilam , a. soni , hep - ph/9812518 ; j. m. yang , 20 , 553 ( 2001 ) ; d. atwood , _ et al .
_ , 66 , 093005 ( 2002 ) ; g. eilam , _ et al .
_ , 510 , 227 ( 2001 ) ; g. bhattacharyya , j. ellis , k. sridhar , 10 , 1583 ( 1995 ) ; g. bhattacharyya , d. choudhury , k. sridhar , 355 , 193 ( 1995 ) ; z. heng _ et al . _ , 79 , 094029 ( 2009 )
. for collider phenomenology of r - violation , see , e.g. , j. erler , j. l. feng , n. polonsky , 78 , 3063 ( 1997 ) ; a. datta , _ et al .
_ , 56 , 3107 ( 1997 ) ; r. j. oakes _ et al . _ , 57 , 534 ( 1998 ) ; m. chemtob and g. moreau , 59 , 116012 ( 1999 ) ; p. chiappetta _
et al . _ , 61 , 115008 ( 2000 ) ; j. l. feng , j. f. gunion , t. han , 58 , 071701 ( 1998 ) ; d. k. ghosh , s. raychaudhuri , k. sridhar , 396 , 177 ( 1997 ) ; s. bar - shalom , g. eilam , a. soni,80 , 4629 ( 1998 ) ; 59 , 055012 ( 1999 ) ; s. bar - shalom , g. eilam , j. wudka , a. soni , 59 , 035010 ( 1999 ) ; b.c .
_ , 420 , 307 ( 1998 ) ; e. perez , y. sirois , h. dreiner , hep - ph/9703444 ; k. hikasa , j. m. yang , b .-
young , 60 , 114041 ( 1999 ) ; p. li _
_ , 51 , 163 ( 2007 ) ; k. j. abraham _ et al .
_ , 63 , 034011 ( 2001 ) ; 514 , 72 ( 2001 ) .
a. belyaev _
et al . _ , jhep 0409 , 012 ( 2004 ) ; j. cao _ et al .
_ , arxiv:0812.1698 [ hep - ph ] . for a review of current bounds ,
see , e.g. , m. chemtob , prog . part .
. phys . * 54 * , 71 ( 2005 ) ; r. barbier _
et al . _ , phys .
420 , 1 ( 2005 ) .
g. bonvicini _ et al . _
[ cleo collaboration ] , phys .
* 79 * , 1221 ( 1997 ) y. enari _ et al . _
[ belle collaboration ] , phys . lett .
b * 622 * , 218 ( 2005 ) y. enari _ et al .
_ [ belle collaboration ] , phys .
lett . * 93 * , 081803 ( 2004 ) y. yusa _ et al . _ [ belle collaboration ] , phys . lett .
b * 589 * , 103 ( 2004 ) m. l. brooks , _ et al .
_ , mega collaboration , phys .
lett . * 83 * , 1521 ( 1999 ) ; b. aubert , _ et al .
_ , babar collaboration , phys .
* 95 * , 041802 ( 2005 ) ; k. hayasaka , _ et al .
_ , belle collaboration , phys .
b * 666 * , 16 ( 2008 ) . m. ahmed _ et al .
_ [ mega collaboration ] , phys . rev .
d * 65 * , 112002 ( 2002 ) ; k. abe _ et al .
_ [ belle collaboration ] , phys .
* 92 * , 171802 ( 2004 )
. for lepton flavor violation in susy , see , e.g. , j. a. casas , a. ibarra , 618 , 171 ( 2001 ) ; j. hisano , t. moroi , k.tobe , m. yamaguchi , 357 , 579 ( 1995 ) ; 53 , 2442 ( 1996 ) ; 391 , 341 ( 1997 ) ; j. hisano , d. nomura , 59 , 116005 ( 1999 ) ; j. hisano , _ et al . _ , 58 , 116010 ( 1998 ) ; j. hisano , d. nomura , t. yanagida 437 , 351 ( 1998 ) ; j. j. cao , _ et al .
_ , phys .
d * 59 * , 095001 ( 1999 ) ; j. ellis , _ et al .
_ , 14 , 319 ( 2000 ) ; j. l. feng , y. nir , y. shadmi , 61 , 113005 ( 2000 ) ; g.k .
leontaros , n.d .
ttracas,431 , 90 ( 1998 ) ; w. buchmuller , d. delepine , l.t .
handoko , 576 , 445 ( 2000 ) ; w. buchmuller , d. delepine , f. vissani , 459 , 171 ( 1999 ) ; j. sato , k. tobe , 63 , 116010;1 j. sato , k. tobe , t. yanagita,498 , 189 ( 2001 ) ; d. f. carvalho , m.e .
gomez , s. khalil , 0107 , 001 ( 2001 ) ; s. f. king , m. oliveira , 60 , 035003 ( 1999 ) ; m. c. chen , k. t. mahanthappa 70 , 113013 ( 2004 ) ; a. gemintern , _ et al .
_ , 67 , 115012 ( 2003 ) ; j. cao , z. xiong , j. m. yang , 32 , 245 ( 2004 ) ; e. arganda , m. j. herrero , 73 , 055003(2006 ) d. f. carvalho , m. e. gomez , j. c. romao , 65 , 093013 ( 2002 ) ; d. atwood , _ et al .
_ , 66 , 093005 ( 2002 ) ; j. i. illana , t. riemann , 63 , 053004 ( 2001 ) ; j. i. illana , m. masip , 67 , 035004 ( 2003 ) . m. chaichian and k. huitu , phys .
b * 384 * , 157 ( 1996 ) .
m. a. mughal , m. sadiq and k. ahmed , phys . lett .
b * 417 * , 87 ( 1998 ) ; w. shao - ming _ et al .
_ , phys .
d * 74 * , 057902 ( 2006);arxiv:0706.3079 [ hep - ph ] .
y. b. sun _ et al .
_ , jhep * 0409 * , 043 ( 2004 ) . j. j. cao _ et al .
_ , phys .
d * 75 * , 075021 ( 2007 ) .
i. f. ginzburg _
et al . _ ,
instrum 219 , 5 ( 1984 ) ; v. i. telnov , nucl .
294 , 72 ( 1990 ) .
t. aaltonen _
[ cdf collaboration ] , arxiv : 0811.2512 .
achard _ et al .
_ [ l3 collaboration ] , 580 , 37 ( 2004 ) .
m. drees , s. pakvasa , x. tata , t. veldhuis , 57,5335 ( 1998 ) ; f. borzumati , j. s. lee , 66 , 115012 ( 2002 ) ; p. dey , a. kundu , b. mukhopadhyaya , s. nandi , 0812 , 100 ( 2008 ) .
r. akers _ et al .
_ [ opal collaboration ] , z. phys .
c * 67 * , 555 ( 1995 ) .
p. abreu _ et al .
_ [ delphi collaboration ] , z. phys .
c * 73 * , 243 ( 1997 ) .
g. wilson , talks at desy - ecfa lc workshops in frascati , 1998 and oxford , 1999 | we examine the lepton flavor - changing processes in @xmath0-parity violating mssm .
first , we update the constraints on the relevant @xmath0-violating couplings by using the latest data on the rare decays @xmath1 .
we find that the updated constraints are much stronger than the old ones from rare @xmath2-decays at lep .
then we calculate the processes @xmath3 and @xmath4 .
we find that with the updated constraints the @xmath0-violating couplings can still enhance the rates of these processes to the sensitivity of gigaz and photon - photon collision options of the ilc . |
authors : : f. d. witherden , a. m. farrington , p. e. vincent program title : : pyfr v0.1.0 licensing provisions : : new style bsd license programming language : : python , cuda and c computer : : variable , up to and including gpu clusters operating system : : recent version of linux / unix ram : : variable , from hundreds of megabytes to gigabytes number of processors used : : variable , code is multi - gpu and multi - cpu aware through a combination of mpi and openmp external routines / libraries : : python 2.7 , numpy , pycuda , mpi4py , sympy , mako nature of problem : : compressible euler and navier - stokes equations of fluid dynamics ; potential for any advection - diffusion type problem . solution method
: : high - order flux reconstruction approach suitable for curved , mixed , unstructured grids .
unusual features : : code makes extensive use of symbolic manipulation and run - time code generation through a domain specific language . running time : : many small problems can be solved on a recent workstation in minutes to hours .
throughout we adopt a convention in which dummy indices on the right hand side of an expression are summed . for example @xmath0 where the limits are implied from the surrounding context .
all indices are assumed to be zero - based
. 2 = + @xmath1 kronecker delta + @xmath2 matrix determinant + @xmath3 matrix dimensions + + * indices . *
+ @xmath4 element type + @xmath5 element number + @xmath6 field variable number + @xmath7 summation indices + @xmath8 summation indices + + * domains .
* + @xmath9 solution domain + @xmath10 all elements in @xmath9 of type @xmath4 + @xmath11 a _ standard _ element of type @xmath4 + @xmath12 boundary of @xmath11 + @xmath13 @@totalleftmargin@-curtab@ element @xmath5 of type @xmath4 in @xmath9 + @xmath14 number of elements of type @xmath4 + + * expansions . *
+ @xmath15 polynomial order + @xmath16 number of spatial dimensions + @xmath17 number of field variables + @xmath18 @@totalleftmargin@-curtab@ nodal basis polynomial @xmath19 for element type @xmath4 + @xmath20 physical coordinates + @xmath21 transformed coordinates + @xmath22 transformed to physical mapping + + * adornments and suffixes .
* + @xmath23 @@totalleftmargin@-curtab@ a quantity in transformed space + @xmath24 @@totalleftmargin@-curtab@ a vector quantity of unit magnitude + @xmath25 transpose + @xmath26 a quantity at a solution point + @xmath27 a quantity at a flux point + @xmath28 @@totalleftmargin@-curtab@ a normal quantity at a flux point + + * operators . * + @xmath29 @@totalleftmargin@-curtab@ common solution at an interface + @xmath30 @@totalleftmargin@-curtab@ common normal flux at an interface + +
there is an increasing desire amongst industrial practitioners of computational fluid dynamics ( cfd ) to undertake high - fidelity scale - resolving simulations of transient compressible flows within the vicinity of complex geometries .
for example , to improve the design of next generation unmanned aerial vehicles ( uavs ) , there exists a need to perform simulations at reynolds numbers @xmath31@xmath32 and mach numbers @xmath33@xmath34of highly separated flow over deployed spoilers / air - brakes ; separated flow within serpentine intake ducts ; acoustic loading in weapons bays ; and flow over entire uav configurations at off - design conditions .
unfortunately , current generation industry - standard cfd software based on first- or second - order accurate reynolds averaged navier - stokes ( rans ) approaches is not well suited to performing such simulations .
henceforth , there has been significant interest in the potential of high - order accurate methods for unstructured mixed grids , and whether they can offer an efficient route to performing scale - resolving simulations within the vicinity of complex geometries .
popular examples of high - order schemes for unstructured mixed grids include the discontinuous galerkin ( dg ) method , first introduced by reed and hill @xcite , and the spectral difference ( sd ) methods originally proposed under the moniker ` staggered - gird chebyshev multidomain methods ' by kopriva and kolias in 1996 @xcite and later popularised by sun et al .
@xcite . in 2007
huynh proposed the flux reconstruction ( fr ) approach @xcite ; a unifying framework for high - order schemes for unstructured grids that incorporates both the nodal dg schemes of @xcite and , at least for a linear flux function , any sd scheme .
in addition to offering high - order accuracy on unstructured mixed grids , fr schemes are also compact in space , and thus when combined with explicit time marching offer a significant degree of element locality . as such , explicit high - order fr schemes are characterised by a large degree of structured computation . over the past two decades improvements in the arithmetic capabilities of processors
have significantly outpaced advances in random access memory .
algorithms which have traditionally been compute bound such as dense matrix - vector products are now limited instead by the bandwidth to / from memory .
this is epitomised in . whereas the processors of two decades ago had flops - per - byte of @xmath35 more recent chips have ratios upwards of @xmath36 .
this disparity is not limited to just conventional cpus .
massively parallel accelerators and co - processors such as the nvidia k20x and intel xeon phi 5110p have ratios of @xmath37 and @xmath38 , respectively .
trends in the peak floating point performance ( double precision ) and memory bandwidth of sever - class intel processors from 19942013 .
the quotient of these two measures yields the flops - per - byte of a processor .
data courtesy of jan treibig . ]
a concomitant of this disparity is that modern hardware architectures are highly dependent on a combination of high speed caches and/or shared memory to maintain throughput .
however , for an algorithm to utilise these efficiently its memory access pattern must exhibit a degree of either spatial or temporal locality . to a first - order approximation
the spatial locality of a method is inversely proportional to the amount of memory indirection . on an unstructured grid indirection
arises whenever there is coupling between elements .
this is potentially a problem for discretisations whose stencil is not compact .
coupling also arises in the context of implicit time stepping schemes .
implementations are therefore very often bound by memory bandwidth . as a secondary trend
we note that the manner in which flops are realised has also changed . in the early 1990s commodity cpus
were predominantly scalar with a single core of execution . however in 2013 processors with eight or more cores are not uncommon .
moreover , the cores on modern processors almost always contain vector processing units .
vector lengths up to 256-bits , which permit up to four double precision values to be operated on at once , are not uncommon .
it is therefore imperative that compute - bound algorithms are amenable to both multithreading and vectorisation .
a versatile means of accomplishing this is by breaking the computation down into multiple , necessarily independent , streams . by virtue of their independence
these streams can be readily divided up between cores and vector lanes .
this leads directly to the concept of _ stream processing_. we will refer to architectures amenable to this form of parallelisation as streaming architectures . a corollary of the above discussion is that compute intensive discretisations which can be formulated within the stream processing paradigm are well suited to acceleration on current and likely future hardware platforms .
the fr approach combined with explicit time stepping is an archetypical of this .
our objective in this paper is to present pyfr , an open - source python based framework for solving advection - diffusion type problems on streaming architectures using the fr approach .
the framework is designed to solve a range of governing systems on mixed unstructured grids containing various element types .
it is also designed to target a range of hardware platforms via use of an in - built domain specific language derived from the mako templating engine .
the current release of pyfr is able to solve the compressible euler and navier - stokes equations on unstructured grids of quadrilateral and triangular elements in two - dimensions , and unstructured grids of hexahedral elements in three - dimensions , targeting clusters of cpus , and nvidia gpus .
the paper is structured as follows . in
we provide a overview of the fr approach for advection - diffusion type problems on mixed unstructured grids . in
we proceed to describe our implementation strategy , and in we present the euler and navier - stokes equations , which are solved by the current release of pyfr .
the framework is then validated in , single - node performance is discussed in , and scalability of the code is demonstrated on up to 104 nvidia m2090 gpus in . finally , conclusions are drawn in .
a brief overview of the fr approach for solving advection - diffusion type problems is given below .
extended presentations can be found elsewhere @xcite .
consider the following advection - diffusion problem inside an arbitrary domain @xmath9 in @xmath16 dimensions @xmath39 where @xmath40 is the _ field variable _ index , @xmath41 is a conserved quantity , @xmath42 is the flux of this conserved quantity and @xmath43 . in defining the flux we have taken @xmath44 in its unscripted form to refer to all of the @xmath17 field variables and @xmath45 to be an object of length @xmath46 consisting of the gradient of each field variable .
we start by rewriting as a first order system @xmath47 where @xmath48 is an auxiliary variable . here , as with @xmath45 ,
we have taken @xmath48 in its unsubscripted form to refer to the gradients of all of the field variables .
take @xmath49 to be the set of available element types in @xmath50 .
examples include quadrilaterals and triangles in two dimensions and hexahedra , prisms , pyramids and tetrahedra in three dimensions .
consider using these various elements types to construct a conformal mesh of the domain such that @xmath51 where @xmath10 refers to all of the elements of type @xmath4 inside of the domain , @xmath14 is the number of elements of this type in the decomposition , and @xmath5 is an index running over these elements with @xmath52 . inside each element
@xmath13 we require that @xmath53 it is convenient , for reasons of both mathematical simplicity and computational efficiency , to work in a transformed space . we accomplish this by introducing , for each element type , a standard element @xmath54 which exists in a transformed space , @xmath55 .
next , assume the existence of a mapping function for each element such that @xmath56 along with the relevant jacobian matrices @xmath57 these definitions provide us with a means of transforming quantities to and from standard element space .
taking the transformed solution , flux , and gradients inside each element to be @xmath58 and letting @xmath59 , it can be readily verified that @xmath60 as required .
we note here the decision to multiply the first equation through by a factor of @xmath61 .
doing so has the effect of taking @xmath62 which allows us to work in terms of the physical solution .
this is more convenient from a computational standpoint .
we next proceed to associate a set of solution points with each standard element . for each type @xmath63 take @xmath64 to be the chosen set of points where @xmath65 .
these points can then be used to construct a nodal basis set @xmath66 with the property that @xmath67 . to obtain such a set we first take @xmath68 to be any basis which spans a selected order @xmath15 polynomial space defined inside @xmath69 .
next we compute the elements of the generalised vandermonde matrix @xmath70 . with these
a nodal basis set can be constructed as @xmath71 . along with the solution points inside of each element
we also define a set of flux points on @xmath72 .
we denote the flux points for a particular element type as @xmath73 where @xmath74 .
let the set of corresponding normalised outward - pointing normal vectors be given by @xmath75 .
it is critical that each flux point pair along an interface share the same coordinates in physical space . for a pair of flux points @xmath76 and @xmath77 at a non - periodic interface
this can be formalised as @xmath78 .
a pictorial illustration of this can be seen in .
solution points ( blue circles ) and flux points ( orange squares ) for a triangle and quadrangle in physical space . for the top edge of the quadrangle
the normal vectors have been plotted .
observe how the flux points at the interface between the two elements are co - located . ]
the first step in the fr approach is to go from the discontinuous solution at the solution points to the discontinuous solution at the flux points @xmath79 where @xmath80 is an approximate solution of field variable @xmath6 inside of the @xmath5th element of type @xmath4 at solution point @xmath81 .
this can then be used to compute a _ common solution _
@xmath82 where @xmath83 is a scalar function that given two values at a point returns a common value . here
we have taken @xmath84 to be the element type , flux point number and element number of the adjoining point at the interface .
since grids in fr are permitted to be unstructured the relationship between @xmath76 and @xmath84 is indirect .
this necessitates the use of a lookup table . as the common solution function
is permitted to perform upwinding or downwinding of the solution it is in general the case that @xmath85 .
hence , it is important that each flux point pair only be visited _ once _ with the same common solution value assigned to both @xmath86 and @xmath87 .
further , associated with each flux point is a vector correction function @xmath88 constrained such that @xmath89 with a divergence that sits in the same polynomial space as the solution . using these fields we can express the solution to as @xmath90_{\tilde{{\mathbf{x } } }
= \tilde{{\mathbf{x}}}^{(u)}_{e\sigma}},\ ] ] where the term inside the curly brackets is the ` jump ' at the interface and the final term is an order @xmath91 approximation of the gradient obtained by differentiating the discontinuous solution polynomial . following the approaches of kopriva @xcite and sun et al .
@xcite we can now compute physical gradients as @xmath92 where @xmath93 .
having solved the auxiliary equation we are now able to evaluate the transformed flux @xmath94 where @xmath95 .
this can be seen to be a collocation projection of the flux . with this
it is possible to compute the normal transformed flux at each of the flux points @xmath96 considering the physical normals at the flux points we see that @xmath97 which is the outward facing normal vector in physical space where @xmath98 is defined as the magnitude . as the interfaces between two elements conform we must have @xmath99 . with these definitions we are now in a position to specify an expression for the _ common normal flux _ at a flux point pair as @xmath100 the relationship @xmath101 arises from the desire for the resulting numerical scheme to be conservative ; a net outward flux from one element must be balanced by a corresponding inward flux on the adjoining element .
it follows that that @xmath102 .
the common normal fluxes in can now be taken into transformed space via @xmath103 where @xmath104 .
it is now possible to compute an approximation for the divergence of the _ continuous _ flux .
the procedure is directly analogous to the one used to calculate the transformed gradient in @xmath105_{\tilde{{\mathbf{x } } } = \tilde{{\mathbf{x}}}^{(u)}_{e\rho}},\ ] ] which can then be used to obtain a semi - discretised form of the governing system @xmath106 where @xmath107 .
this semi - discretised form is simply a system of ordinary differential equations in @xmath108 and can be solved using one of a number of schemes , e.g. a classical fourth order runge - kutta ( rk4 ) scheme .
pyfr is a python based implementation of the fr approach described in section .
it is designed to be compact , efficient , and platform portable .
key functionality is summarised in table .
.[tab : pyfr - func]key functionality of pyfr . [ cols= "
> , < " , ]
in this paper we have described pyfr , an open source python based framework for solving advection - diffusion type problems on streaming architectures .
the structure and ethos of pyfr has been explained including our methodology for targeting multiple hardware platforms .
we have shown that pyfr exhibits spatial super accuracy when solving the 2d euler equations and the expected order of accuracy when solving couette flow problem on a range of grids in 2d and 3d .
qualitative results for unsteady 3d viscous flow problems on curved grids have also been presented .
performance of pyfr has been validated on an nvidia m2090 gpu in three dimensions .
it has been shown that the compute bound kernels are able to obtain between @xmath109 and @xmath110 of realisable peak flop / s whereas the bandwidth bound point - wise kernels are , across the board , able to obtain in excess of @xmath111 realisable peak bandwidth .
the scalability of pyfr has been demonstrated in the strong sense up to 32 nvidia m2090s and in the weak sense up to 104 nvidia m2090s when solving the 3d navier - stokes equations .
the authors would like to thank the engineering and physical sciences research council for their support via two doctoral training grants and an early career fellowship ( ep / k027379/1 ) .
the authors would also like to thank the e - infrastructure south centre for innovation for granting access to the emerald supercomputer , and nvidia for donation of three k20c gpus .
it is possible to cast the majority of operations in an fr step as matrix - matrix multiplications of the form @xmath112 where @xmath113 are constants , @xmath114 is a constant operator matrix , and @xmath115 and @xmath116 are state matrices . to accomplish this we start by introducing the following constant operator matrix @xmath117 and the following state matrices @xmath118 in specifying the state matrices there is a degree of freedom associated with how the @xmath17 field variables for each element are packed along a row of the matrix , with the possible packing choices being discussed in . using these matrices we are able to reformulate as @xmath119 in order to apply a similar procedure to we let @xmath120_{\tilde{{\mathbf{x } } } = \tilde{{\mathbf{x}}}^{(u)}_{e\sigma } } , & \dim { \bm{\mathsf{m}}}^{4}_{e } , & = n_{d}n_{e}^{(u ) } \times n_{e}^{(u)},\\ \big({\bm{\mathsf{m}}}^{6}_{e}\big)_{\rho\sigma } & = \big[\hat{\tilde{{\mathbf{n}}}}^{(f)}_{e\rho}\cdot \tilde{{\bm{\nabla } } } \cdot { \mathbf{g}}^{(f)}_{e\rho}(\tilde{{\mathbf{x}}})\big]_{\tilde{{\mathbf{x } } } = \tilde{{\mathbf{x}}}^{(f)}_{e\sigma } } , & \dim { \bm{\mathsf{m}}}^{6}_{e } , & = n_{d}n_{e}^{(u ) } \times n_{e}^{f},\\ \big({\bm{\mathsf{c}}}^{(f)}_{e}\big)_{\rho(n\alpha ) } & = \mathfrak{c}_{\alpha}u^{(f)}_{e\rho n\alpha } , & \dim { \bm{\mathsf{c}}}^{(f)}_{e } & = n^{(f)}_{e } \times n_v{\left\lvert{\mathbf{\omega}}_e\right\rvert},\\ \big(\tilde{{\bm{\mathsf{q}}}}^{(u)}_{e}\big)_{\sigma(n\alpha ) } & = \tilde{{\mathbf{q}}}^{(u)}_{e\sigma n\alpha } , & \dim \tilde{{\bm{\mathsf{q}}}}^{(u)}_{e } & = n_dn^{(u)}_{e } \times n_v{\left\lvert{\mathbf{\omega}}_e\right\rvert},\end{aligned}\ ] ] here it is important to qualify assignments of the form @xmath121 where @xmath122 is a @xmath16 component vector . as above
there is a degree of freedom associated with the packing . with the benefit of foresight
we take the stride between subsequent elements of @xmath122 in a matrix column to be either @xmath123 or @xmath124 depending on the context . with these matrices
reduces to @xmath125 applying the procedure to we take @xmath126 hence @xmath127 where we note the block diagonal structure of @xmath128 .
this is a direct consequence of the above choices for @xmath129 .
finally , to rewrite we write @xmath130^t_{\tilde{{\mathbf{x } } } = \tilde{{\mathbf{x}}}^{(u)}_{e\sigma } } , & \dim { \bm{\mathsf{m}}}^{1}_{e } & = n^{(u)}_{e } \times n_dn^{(u)}_{e},\\ \big({\bm{\mathsf{m}}}^{2}_{e}\big)_{\rho\sigma } & = \big[\ell^{(u)}_{e\rho}(\tilde{{\mathbf{x}}}^{(f)}_{e\sigma } ) \hat{\tilde{{\mathbf{n}}}}^{(f)}_{e\sigma}\big]^t , & \dim { \bm{\mathsf{m}}}^{2}_{e } & = n^{(f)}_{e } \times n_dn^{(u)}_{e},\\ \big({\bm{\mathsf{m}}}^{3}_{e}\big)_{\rho\sigma } & = \big[\tilde{{\bm{\nabla } } } \cdot { \mathbf{g}}^{(f)}_{e\sigma}(\tilde{{\mathbf{x}}})\big]_{\tilde{{\mathbf{x } } } = \tilde{{\mathbf{x}}}^{(u)}_{e\rho } } , & \dim { \bm{\mathsf{m}}}^{3}_{e } & = n^{(u)}_{e } \times n^{(f)}_{e},\\ \big(\tilde{{\bm{\mathsf{d}}}}^{(f)}_{e}\big)_{\sigma(n\alpha ) } & = \mathfrak{f}^{\vphantom{(f_\perp)}}_{\alpha}\tilde{f}^{(f_\perp)}_{e\sigma n\alpha } , & \dim \tilde{{\bm{\mathsf{d}}}}^{(f)}_{e } & = n^{(f)}_{e } \times n_v{\left\lvert{\mathbf{\omega}}_e\right\rvert},\\ \big(\tilde{{\bm{\mathsf{f}}}}^{(u)}_{e}\big)_{\rho(n\alpha ) } & = \tilde{{\mathbf{f}}}^{(u)}_{e\rho n\alpha } , & \dim \tilde{{\bm{\mathsf{f}}}}^{(u)}_{e } & = n_dn^{(u)}_{e } \times n_v{\left\lvert{\mathbf{\omega}}_e\right\rvert},\\ \big(\tilde{{\bm{\mathsf{r}}}}^{(u)}_{e}\big)_{\rho(n\alpha ) } & = ( \tilde{{\bm{\nabla } } } \cdot \tilde{{\mathbf{f}}})^{(u)}_{e\rho n\alpha } , & \dim \tilde{{\bm{\mathsf{r}}}}^{(u)}_{e } & = n_e^{(u ) } \times n_v{\left\lvert{\mathbf{\omega}}_e\right\rvert},\end{aligned}\ ] ] and after substitution of for @xmath131 obtain @xmath132
in the following section we take @xmath133 and @xmath134 to be the two discontinuous solution states at an interface and @xmath135 to be the normal vector associated with the first state . for convenience
we take @xmath136 , and @xmath137 with inviscid fluxes being prescribed by .
also known as the local lax - friedrichs method a rusanov type riemann solver imposes inviscid numerical interface fluxes according to @xmath138 where @xmath139 is an estimate of the maximum wave speed @xmath140
to incorporate boundary conditions into the fr approach we introduce a set of boundary interface types @xmath141 . at a boundary interface
there is only a single flux point : that which belongs to the element whose edge / face is on the boundary .
associated with each boundary type are a pair of functions @xmath142 and @xmath143 where @xmath133 , @xmath144 , and @xmath135 are the solution , solution gradient and unit normals at the relevant flux point .
these functions prescribe the common solutions and normal fluxes , respectively . instead of directly imposing solutions and normal fluxes
it is oftentimes more convenient for a boundary to instead provide ghost states . in its simplest formulation @xmath145 and @xmath146 where @xmath147 is
the ghost solution state and @xmath148 is the ghost solution gradient .
it is straightforward to extend this prescription to allow for the provisioning of different ghost solution states for @xmath149 and @xmath150 and to permit @xmath148 to be a function of @xmath133 in addition to @xmath144 .
patrice castonguay , pe vincent , and antony jameson .
application of high - order energy stable flux reconstruction schemes to the euler equations . in _
49th aiaa aerospace sciences meeting _ , volume 686 , 2011 . | high - order numerical methods for unstructured grids combine the superior accuracy of high - order spectral or finite difference methods with the geometric flexibility of low - order finite volume or finite element schemes .
the flux reconstruction ( fr ) approach unifies various high - order schemes for unstructured grids within a single framework . additionally , the fr approach exhibits a significant degree of element locality , and is thus able to run efficiently on modern streaming architectures , such as graphical processing units ( gpus ) .
the aforementioned properties of fr mean it offers a promising route to performing affordable , and hence industrially relevant , scale - resolving simulations of hitherto intractable unsteady flows within the vicinity of real - world engineering geometries . in this paper
we present pyfr , an open - source python based framework for solving advection - diffusion type problems on streaming architectures using the fr approach .
the framework is designed to solve a range of governing systems on mixed unstructured grids containing various element types .
it is also designed to target a range of hardware platforms via use of an in - built domain specific language based on the mako templating engine .
the current release of pyfr is able to solve the compressible euler and navier - stokes equations on grids of quadrilateral and triangular elements in two dimensions , and hexahedral elements in three dimensions , targeting clusters of cpus , and nvidia gpus .
results are presented for various benchmark flow problems , single - node performance is discussed , and scalability of the code is demonstrated on up to 104 nvidia m2090 gpus .
the software is freely available under a 3-clause new style bsd license ( see www.pyfr.org ) . _
keywords : _ high - order ; flux reconstruction ; parallel algorithms ; heterogeneous computing |
= 2@xmath1 and @xmath2 = @xmath1 + @xmath3@xmath4 , and the bonds with heisenberg exchange couplings @xmath5 ( ` expanded bonds ' ) and @xmath6 ( ` triangle bonds ' ) . bottom : phase diagram of the antiferromagnetic heisenberg model on the star lattice .
for @xmath7 , the ground state is a valence bond solid ( vbs ) phase ( @xmath8-dimer vbs ) in which every dimer sits on the expanded links connecting neighboring triangles . for @xmath9 ,
the ground state is a vbs with an 18-site unit cell ( the columnar 18-site vbs ) .
, title="fig:",width=6 ] = 2@xmath1 and @xmath2 = @xmath1 + @xmath3@xmath4 , and the bonds with heisenberg exchange couplings @xmath5 ( ` expanded bonds ' ) and @xmath6 ( ` triangle bonds ' ) . bottom : phase diagram of the antiferromagnetic heisenberg model on the star lattice . for @xmath7 ,
the ground state is a valence bond solid ( vbs ) phase ( @xmath8-dimer vbs ) in which every dimer sits on the expanded links connecting neighboring triangles . for @xmath9 ,
the ground state is a vbs with an 18-site unit cell ( the columnar 18-site vbs ) .
, title="fig:",width=8 ] in recent years , several geometrically frustrated spin-1/2 magnets have been synthesized which appear to not order magnetically even at temperatures well below the characteristic exchange couplings . among these
are the quasi - two - dimensional ( 2d ) triangular organic material @xmath10-bedt(cn)@xmath11 , @xcite the kagome lattice herbertsmithite , @xcite and distorted kagome lattice volborthite , @xcite and the three - dimensional hyperkagome lattice magnet @xcite na@xmath12ir@xmath11o@xmath13 . a large class of these magnets appear to exhibit gapless spin liquid behavior down to very low temperatures , leading to the exciting possibility that they may possess exotic ground states with fractionalized excitations.@xcite others among these have been proposed to weakly order into singlet valence bond solid ( vbs ) states which break lattice symmetries.@xcite as yet , there is no clear picture of what combination of geometric effects and spin interactions will lead to spin liquid ground states or vbs ground states ; this necessitates a theoretical and experimental exploration of various new lattice geometries as well as possible ring - exchange interactions beyond the simplest heisenberg spin exchange interaction . in this paper
, we focus on understanding several competing ground states of the nearest neighbor @xmath14 heisenberg model @xmath15 on the 2d star lattice , shown in fig . 1
, as a function of @xmath16 where @xmath6 and @xmath5 are the exchange couplings on the ` triangle bonds ' and ` expanded bonds ' .
our motivation for this study is two - fold .
first , the recent synthesis of a ` star lattice ' organic iron acetate quantum magnet @xcite raises the possibility that a @xmath14 variant may possibly be synthesized in the near future and our results should be applicable to such systems .
second , this lattice has a sufficiently different geometry from more commonly studied quantum magnets it may be viewed either as a variant of the kagome lattice or as a decorated honeycomb lattice which allows us to explore the effect of this new lattice geometry on possible spin liquid physics and valence bond solid phases in quasi-2d systems . motivated by interpolating between this spin-1/2 su(2 ) model and an su(n ) generalization at large n which permits a mean field solution with fermionic spinon excitations,@xcite we examine a large number of interesting u(1 ) spin liquids as candidate ground states of the nearest neighbor @xmath14 heisenberg model on the 2d star lattice . guided by earlier work on the kagome lattice,@xcite we focus on spin liquid states denoted by sl@xmath17 $ ] where @xmath18,@xmath19 , and @xmath20 denote , respectively , the fictitious " fluxes seen by the fermionic spinons as they move around an elementary plaquette of the lattice : an up triangular plaquette @xmath21 , a down triangular plaquette @xmath22 , or the @xmath23-site dodecagon plaquette . in terms of the original spin variable , the fluxes on the triangular plaquettes correspond to scalar spin chiralities of the form @xmath24 , while @xmath20 is related to an operator defined by the twelve spins around the dodecagon loop . depending on the flux values , these spin liquids represent gapped chiral spin liquids which break time - reversal symmetry,@xcite or states with gapless fermi surfaces of spinons , or gapped spin liquids with no broken symmetries . from a study of energetics of various flux values using gutzwiller projected wave function numerics for the physical case of @xmath25 ,
we show that a particular gapped spin liquid , which we denote as sl@xmath26 $ ] , which does not break lattice or time - reversal symmetries emerges as a favorable candidate over a wide range of @xmath16 .
this is in striking contrast to earlier work on the kagome lattice from two perspectives .
first , as we show , the effect of projection is far more dramatic on the star lattice when compared with the kagome lattice ; a numerical gutzwiller projection of the mean field states leads to a complete reordering of the energies of the candidate spin liquids .
second , unlike the kagome lattice case where the lowest energy variational state of this form is a spin liquid with massless dirac fermion excitations,@xcite the sl@xmath26 $ ] is a gapped u(1 ) spin liquid we therefore know that it is ultimately unstable towards spinon confinement at low energies@xcite unlike the dirac fermion states whose stability depends on the number of fermion flavors @xmath27.@xcite we find that the sl@xmath26 $ ] state naturally forms strong dimers on the ` expanded bonds ' for small values of @xmath16 thus leading to a confined state , a @xmath5-dimer vbs , which respects all symmetries of the hamiltonian.@xcite for large @xmath16 , numerical exact diagonalization ( ed ) studies of this model carried out in a restricted nearest - neighbor dimer basis showed signatures of @xmath0 ordering.@xcite we argue that another motivation to study possible dimer orders that break lattice symmetries is that such ordering often appear quite naturally in the large - n fermionic su(n ) theory as recognized in the early work of affleck and marston@xcite and shown in various other models studied recently.@xcite inspired by these results , we consider various candidate vbs phases from different perspectives a large - n route , a bond operator formalism , and gutzwiller projected wave function numerics .
all of these point to a transition to a @xmath0 ordered vbs phase for large enough @xmath28 , leading us to the phase diagram shown in fig . 1 .
our result is in broad agreement with the ed study although the transition point estimated from our work is somewhat larger than the exact diagonalization ( ed ) result which yields @xmath29 ; the ed result may , however , may suffer from significant finite size effects .
we then discuss possible routes by which the sl@xmath26 $ ] state might be unstable towards such @xmath0 vbs order instead of the @xmath5-dimer vbs .
we present results for the triplon dispersion in both vbs states which can be tested in inelastic neutron scattering experiments on candidate materials .
finally , although the various other interesting spin liquids we study do not appear to be energetically viable ground states for the nearest neighbor heisenberg model on the star lattice , they have energies which are close to the ground state .
they might thus be stabilized as ground states by small changes in the hamiltonian , such as further neighbor exchange or spin - phonon coupling , or they might be relevant to understanding the intermediate energy scale properties or finite temperature physics of materials which realize this model .
we therefore elucidate some of the properties of these spin liquid states .
this paper is organized as follows .
we begin , in section [ sec : meanfield ] by formulating the mean field theory of the heisenberg model on this lattice in a slave particle description using fermionic spinons which we relate to a large - n su(n ) approach . based on this , we classify and study the physical properties of a number of candidate spin liquid ansatzes .
we next turn , in section [ sec : dimer ] , to a study of dimerized states on this lattice , and present a group theoretic classification of @xmath0 orders as well as a large - n justification of specific candidate vbs phases .
section [ sec : energetics ] contains a discussion of the energetics of various spin liquid states using mean field theory as well as a gutzwiller projected wave function study and bond operator approaches of candidate vbs phases .
section v discusses the various ways in which the sl@xmath26 $ ] state , which is the lowest energy spin liquid state , might be unstable towards vbs ordering as a result of spinon interactions and from a gutzwiller wave function approach .
the phase transition between vbs phases is described in section [ sec : phasetransition ] .
section [ sec : triplon ] contains a discussion of the triplon dispersion in both vbs states which we think are realized in this model .
we conclude with a discussion about experimental implications in section [ sec : discussion ] .
details of various calculations are contained in appendices a - c .
we investigate the ground state of the s=1/2 heisenberg antiferromagnet on the star lattice , which can be described by the following hamiltonian , @xmath15 where @xmath30 and @xmath31 indicate the positions of nearest neighbor spin pairs .
natural description of this lattice system requires the consideration of two inequivalent links , that is , one link lying on a triangle ( a triangular link ) and the other link connecting two neighboring triangles ( an expanded link ) .
we assign two different exchange couplings @xmath8 and @xmath32 on expanded and triangular links , respectively .
we can also label a site @xmath30 by pairs ( @xmath33 , @xmath34 ) where @xmath33 denotes the location of a unit cell and @xmath34 labels the six sites inside a single unit cell .
( see fig .
[ fig : latticestructure ] . ) to construct spin liquid states we introduce the fermionic spinon operators , @xmath35 ( @xmath36 ) to represent the spin operator ; @xmath37 since this representation alone contains unphysical local configurations enlarging the hilbert space , we have to impose the following local constraint , @xmath38 = 1 , to recover the physical hilbert space . using the fermioninc spinon representation of the spin operator ,
the heisenberg spin hamiltonian can be rewritten as follows .
@xmath39 here we have dropped unimportant constant terms . to decouple the four fermion interaction term we define spin singlet order parameters , @xmath40 @xmath41 @xmath42@xmath43 @xmath44 . after imposing the single occupancy constraint using the lagrange multipliers @xmath45
, the mean field hamiltonian is given by @xmath46 to describe the phase fluctuation of the mean field ansatz , we express @xmath40 as @xmath40 = @xmath47 @xmath48 , which leads to the following hamiltonian , @xmath49 in the above hamiltonian @xmath50 , the local u(1 ) gauge symmetry of the spin hamiltonian which comes from the local conservation of the fermion number is manifest via the following gauge transformation.@xcite @xmath51 here @xmath52 describing the phase fluctuation of @xmath40 plays the role of the spatial components of the u(1 ) gauge field .
namely , we have reformulated the quantum spin model as the problem of the spinons strongly coupled to the u(1 ) gauge field . a systematic way of studying the coupled spinon and gauge field system is to consider the large - n reformulation of the problem extending the spin su(2 ) symmetry to su(n )
( with n even).@xcite we let the flavor index @xmath53 run from 1 to n and modify the single occupancy constraint as , @xmath54 in addition , we scale the interaction strength @xmath55/2 to be @xmath55/n to make each term of the hamiltonian to be of order n. the resulting large - n hamiltonian is given by @xmath56 to treat the quartic interactions we perform a mean field decoupling by introducing su(n ) singlet valence bond , @xmath40 @xmath41 @xmath57@xmath58 @xmath59 . assuming the valence bond amplitude is a complex number , we obtain the mean field hamiltonian given by @xmath60 since the fluctuations of @xmath40 and the average local density @xmath57@xmath58 @xmath61 scale as 1/@xmath62 , we can safely neglect those fluctuations in the large - n limit justifying the mean field approximation . here
we consider the following mean field ansatz @xmath47 = @xmath63 @xmath64 where @xmath63 = @xmath65 on expanded links and @xmath63 = @xmath66 on triangular links , respectively .
we specify the various flux patterns inside the elementary plaquettes , i.e. , the triangles and the dodecagons .
the flux inside a triangle @xmath67 , for example , is defined in the following way , @xmath68 @xmath41 @xmath69 where @xmath70 indicates the three corners of a triangle traversed along the counterclockwise direction
. the flux inside a dodecagon is also defined in the same manner .
since the flux inside a closed loop is a gauge invariant object , different spin liquid ansatz can be distinguished based on the flux values inside the elementary plaquettes .
in particular , we use the term sl@xmath71 $ ] to represent the ansatz which has the fluxes @xmath67 inside an up - pointing triangle , @xmath72 inside a down - pointing triangle and @xmath20 inside a dodecagon . with a given flux configuration
we determine @xmath63 and @xmath73 self - consistently by solving the following coupled mean field equations , @xmath74 where @xmath75 is the number of lattice sites . in this section
we discuss the characteristic properties of various spin liquid phases and their instabilities .
in particular , we focus on translationally invariant mean field states which have nonzero @xmath63 on every link of the lattice . extensive discussion on possible dimerized phases is given later in sec .
[ sec : dimer ] .
as shown in the previous studies about the spin liquid phases on the square@xcite and kagome@xcite lattices , the inclusion of additional spin interactions can change the relative energetics of different spin liquid phases .
therefore it is useful to understand the nature of various competing spin liquid states which are the potential ground states of spin hamiltonians beyond the nearest neighbor heisenberg model .
+ ( a ) + near the fermi level . among the six bands inside the first brillouin zone
we plot the third and four band which are lying close to the fermi energy along the @xmath76 axis .
the flat band is lying at the fermi level which touches with the dispersive band at the @xmath77 point .
, width=8 ] to describe this state we introduce two real mean field order parameters , @xmath65 and @xmath66 , which lie on the expanded and triangular links , respectively . since the unit cell contains six sites we obtain six different bands inside the first brillouin zone .
among the six bands , the third and fourth bands near the fermi level show an interesting structure displayed in fig .
[ fig : band000 ] .
there is a flat band lying on the fermi energy which is touching another dispersive band at the zone center , @xmath78 .
the flat band comes from the existence of the localized eigenstates , which occur due to the destructive interference of hopping amplitudes between the localized units.@xcite the flat band of the uniform spin liquid on the kagome lattice emerges owing to the same reason .
however , in contrast to the kagome lattice problem , the flat band is lying exactly at the fermi level on the star lattice .
the flatness of the band at the fermi energy is not the generic property of the uniform spin liquid .
there are perturbations which do not break any lattice symmetry but spoil the flatness by generating curvature .
the third nearest neighbor hopping is such an example .
however , the quadratic degeneracy at the zone center is protected by the point group symmetry of the underlying unit cell . to understand the stability of the spin liquid we derive the low energy effective hamiltonian , which describes the states near the zone center , expanding the hamiltonian up to the quadratic order of the momentum @xmath79 . the procedure for deriving the effective hamiltonian
is outlined in the appendix [ sec : eff000 ] .
the resulting hamiltonian is given by , @xmath80 in which @xmath81 where the pauli matrix @xmath82 is acting on the two - component space of the continuum field @xmath83=@xmath84 which describes the two low energy states near the @xmath77 point .
the sl[0,0,0 ] state respects all the space group symmetry of the lattice . in particular , if we choose the gauge in which @xmath40=@xmath66 on every triangular link and @xmath40=@xmath65 on every expanded link with @xmath66 and @xmath65 being real constants , the action of the symmetry generators on the spin operator , @xmath85 , is the same as that on the spinon operator , @xmath86 .
since we consider the low energy excitations near the zone center , we focus on the action of the point group symmetry on the continuum fields .
the @xmath87 point group of the star lattice consists of twelve symmetry generators and is generated by the two elements , @xmath88 and @xmath89 .
here @xmath88 means the @xmath90 rotation with respect to the center of a dodecagon and @xmath89 indicates the reflection about the x - axis .
the details on the elements of the @xmath87 point group are discussed in the sec .
[ sec : grouptheory ] . under the @xmath88 and @xmath89 , the continuum field
@xmath91 transforms in the following way , @xmath92 meaning all the fermion bilinears @xmath93 ( @xmath53 = x , y , z ) are forbidden by the point group symmetry .
note that the @xmath94 breaks time - reversal symmetry as well since it shows sign change under complex conjugation .
next we consider the fermion bilinear terms that contain spatial derivatives .
because the dynamical critical exponent is two , the terms with a single spatial derivative are relevant and those with two spatial derivatives are marginal perturbations . investigating the transformation rule under the @xmath87 point group symmetry , it can be easily checked that @xmath94 transforms as the one dimensional @xmath95 irreducible representation and ( @xmath96,@xmath97 ) forms a basis for the two dimensional @xmath98 irreducible representation.@xcite similarly , the transformation properties of derivative terms can be determined . at first , the linear derivative term , ( @xmath99,@xmath100 ) transforms as a two dimensional @xmath101 irreducible representation . to have terms with linear derivatives in the hamiltonian , the product of the fermion bilinear and the derivative must be invariant under the point group symmetry operations . using the decompositions of @xmath102 and @xmath103
, we see that every product of fermion bilinears and the linear derivative is not invariant under the point group symmetry .
therefore linear derivative terms are not allowed . on the other hand
, we have a second derivative term ( @xmath104,@xmath105 ) making a two dimensional @xmath98 irreducible representation . using @xmath106 and @xmath107
, we see that there is a term following @xmath108 irreducible representation , which is nothing but @xmath109\psi$ ] . therefore in addition to the isotropic @xmath110 term
, @xmath109\psi$ ] is the only term allowed by symmetry . since these terms are already present in the hamiltonian , the low energy properties of the sl[0,0,0 ] are not spoiled by these marginal perturbations .
however , these perturbations add curvature to the flat band .
finally , we discuss the effect of the four fermion interaction terms on the stability of the sl[0,0,0 ] state . though a simple power counting shows that the four fermion interactions are marginal , they are actually marginally relevant .
recently , the effects of the four fermion interaction on the quadratic band crossing are studied using the renormalization group approach.@xcite according to ref . , the leading weak coupling instability leads to the state with nonzero @xmath111 , breaking the time reversal symmetry .
it means that sl[0,0,0 ] state is unstable toward a chiral spin liquid state supporting chiral edge states .
+ ( b ) + direction . @xmath112 and
@xmath113 indicate the reciprocal lattice vector corresponding to the doubled unit cell .
( b ) the flux configuration of the sl@xmath26 $ ] state .
the twelve - site unit cell is surrounded by a dotted box .
we assign -1 ( + 1 ) for the hopping amplitude on the thick ( thin ) bond .
, width=6 ] $ ] state near the fermi level . the lower band ( valence band ) is flat and doubly degenerate .
the upper band ( conduction band ) has small dispersion .
the locations of the conduction band minimum ( maximum ) are described by @xmath114 ( @xmath115 ) .
the dispersion is plotted along the @xmath116=@xmath3@xmath76+@xmath117 line passing the @xmath118 and @xmath119 .
, width=8 ] the sl[0,0,@xmath120 state supports @xmath121 flux piercing the dodecagons .
even though this state does not break the translational symmetry , the mean field description requires doubling of the unit cell .
here we consider the doubling of the unit cell along the @xmath122 direction .
for the lattice vectors 2@xmath122 and @xmath2 , the reciprocal lattice vectors are given by @xmath123 the reduced brillouin zone corresponding to the above reciprocal lattice vectors is depicted in fig .
[ fig : gauge00pi](a ) . for the mean field description of the sl@xmath26 $ ] state
, we have chosen the flux configuration as described in fig .
[ fig : gauge00pi](b ) .
since there are twelve sites inside the unit cell , we have twelve bands within the brillouin zone .
the mean field spinon dispersion of the low energy bands near the fermi level is described in fig .
[ fig : band00pi ] .
this state does not have a spinon fermi surface and shows a gapped spectrum .
the lower flat band ( valence band ) is doubly degenerate and the upper band ( conduction band ) is dispersive . according to the projected wave function study that is discussed in detail later in sec .
[ sec : vmc ] , the sl[0,0,@xmath120 state has the lowest ground state energy among the various spin liquid ansatz over a wide parameter range .
unfortunately , however , the sl[0,0,@xmath120 state is unstable once gauge fluctuation is allowed .
since the spinon spectrum has a finite gap , the low energy excitations are described by the compact u(1 ) gauge theory . in 2 + 1
dimension , the compact u(1 ) gauge theory is confining,@xcite which means that free spinons with unit gauge charge can only make charge neutral bound states . in addition , the interaction between spinons can also induce various kinds of broken symmetry states .
extensive discussion on the instability of the sl[0,0,@xmath120 state and its relation with candidate valence bond solid phases are given in sec .
[ sec : sl00pi ] .
+ ( c ) + $ ] and sl@xmath124 $ ] . hopping along ( against ) the direction of an arrow
induce a phase @xmath121/2 ( -@xmath121/2 ) .
( a ) the flux pattern for sl@xmath125 $ ] . counterclockwise motion along an triangular link results in the flux @xmath126 ( b ) the flux pattern for sl@xmath124 $ ] .
counterclockwise motion along an up - pointing ( down - pointing ) triangular link results in the flux -@xmath126
( @xmath126 ) .
, width=8 ] next we consider flux phases which possess finite flux inside triangles . a triangle that supports @xmath126 flux breaks time - reversal and parity symmetry but preserves the combination of them
. a convenient way to investigate the time - reversal symmetry breaking in spin systems is to consider the expectation value of the scalar spin chirality operator@xcite defined as follows , @xmath127 since @xmath128 is odd under both the time reversal ( @xmath129 ) and parity operations , the ground state breaks both symmeties when the expectation value of the scalar spin chirality operator @xmath130 is nonzero . in other words ,
the scalar spin chirality plays the role of the order parameter measuring time reversal symmetry breaking .
because the unit cell contains two triangles , we can define the following two different scalar spin chirality operators , @xmath131 where @xmath132 and @xmath133 are the uniform and staggered scalar spin chiralities , respectively . to understand the symmetry properties of the @xmath132 and @xmath133 we have to recognize that there are two different reflection symmetries on the star lattice .
the reflection ( @xmath134 ) with respect to the axis connecting the center of a dodecagon with the mid - point of an expanded link ( for example , the @xmath135 axis in fig .
[ fig : twofold](a ) ) interchanges the up - pointing triangles and the down - pointing triangles . on the other hand , the other reflection ( @xmath136 ) about the axis connecting the center of a dodecagon with a vertex of a triangle , ( for example , the @xmath137 axis in fig .
[ fig : twofold](a ) ) , leaves each triangle as it is . under the @xmath136 reflection
both @xmath132 and @xmath133 change their signs .
here we use the term parity to indicate the @xmath134 reflection symmetry .
note that the parity operation is equivalent to the reflection in two dimensional space .
@xmath132 is odd under both the time - reversal ( @xmath129 ) and parity ( @xmath134 ) but even under inversion ( @xmath138 ) . on
the other hand @xmath133 is odd under time - reversal and inversion but even under parity . here
the inversion operation is defined with respect to the mid - point of the expanded link connecting two triangles . in both cases ,
the combination of the time - reversal , parity transformation , and inversion ( @xmath139 ) is equivalent to the identity operation under which both the @xmath132 and @xmath133 are invariant.@xcite $ ] and sl@xmath124 $ ] along the @xmath76 axis when @xmath32=2@xmath8 .
( a ) the first brillouin zone . here
q and -q denote the points in which minimum ( maximum ) of the conduction ( valence ) band of the sl@xmath125 $ ] occurs .
the k and - k indicate the points where the linear band touching between the two bands near the fermi energy of the sl@xmath124 $ ] occurs .
( b ) the dispersion of the valence and conduction band corresponding to the sl@xmath125 $ ] ansatz .
the low energy excitation near the @xmath140q can be described by the massive dirac particles .
( c ) the dispersions of the two bands near the fermi level for sl@xmath124 $ ] .
the low energy excitation near the @xmath140k point can be described by the dirac particles under the staggered chemical potential . , width=8 ] the sl@xmath125 $ ] is characterized by nonzero @xmath132 but with vanishing @xmath133 .
therefore it breaks time - reversal and parity transformation while it is invariant under the combined operation .
it is a chiral spin liquid state which has a finite energy gap . in fig .
[ fig : fluxband](a ) we plot the spinon dispersion near the fermi level corresponding to the valence and conduction bands .
the energy gap is minimum at the momentum @xmath141=(@xmath121/3,0 ) and -@xmath141 .
if we expand the mean field hamiltonian near the dispersion minimum , @xmath141 and -@xmath141 in the limit of large @xmath142 , we can get the following effective low energy hamiltonian , @xmath143-m\tau_{z}\big\}\psi(q),\end{aligned}\ ] ] where the fermi velocity @xmath144 = @xmath145 and the mass @xmath146 = @xmath147 . in the above
we define the eight component dirac fermion field , @xmath148 in which 1 and 2 are the two - component dirac indices , @xmath53 and @xmath149 are indices for the nodes ( @xmath150 ) and spins . the pauli matrix @xmath151 acts on the two - component dirac space . for later convenience
we define two additional pauli matrices , @xmath152 and @xmath153 acting on the nodal and spin spaces , respectively .
since the mass term has the same sign in the two nodal positions , integrating out fermions leads to the chern - simons gauge field action . as a consequence ,
the charge neutral spinon hall conductivity should be finite .
the chern - simons term stabilizes the spin liquid ground state by providing a finite mass to the u(1 ) gauge field .
therefore the u(1 ) gauge field can only mediate a short range interaction between the spinons , which makes the fractionalized particles ( the spinons ) to be the elementary excitations of the spin liquid ground state.@xcite + ( d ) + the sl@xmath124 $ ] is characterized by nonzero @xmath133 but with vanishing @xmath132 .
therefore it breaks both the time - reversal and inversion operation but is invariant under the parity transformation .
because the fluxes of the two triangles within the unit cell have opposite sign , the six - fold rotational symmetry is broken down to the three - fold symmetry .
( see fig . [
fig : fluxconfig](b ) . )
thus it is a nematic spin liquid .
the mean field spinon dispersion corresponding to the two bands near the fermi energy is plotted in fig .
[ fig : fluxband ] ( b ) .
the spin liquid ansatz has a spinon fermi surface which consists of an electron pocket at the @xmath154 = ( @xmath155,0 ) point and a hole pocket at the - @xmath154 point .
expanding the mean field hamiltonian using @xmath156 as an expansion parameter , the following effective low energy hamiltonian can be obtained , @xmath157-m\mu_{z}\big\}\psi(q),\end{aligned}\ ] ] where the fermi velocity @xmath144 = @xmath145 and the staggered " field @xmath158 = @xmath147 . since the effective chemical potentials coming from the staggered " field @xmath158 have the opposite signs at the two nodal points , we have both an electron pocket ( at the @xmath154 point ) and a hole pocket ( at the -@xmath154 point ) on the fermi surface .
in contrast to the sl@xmath125 $ ] state which has a gapped spinon spectrum , the sl@xmath124 $ ] state has gapless low energy excitations . to confirm that the low energy description based the above effective hamiltonian in eq .
( [ eq : effectiveh2 ] ) is valid after including the fluctuation beyond the mean field description , we have to check whether there are relevant perturbations which are allowed by symmetry
. especially , some of the fermion bilinears , which are made of @xmath159 , can potentially generate various mass terms which spoil the low energy description of eq .
( [ eq : effectiveh2 ] ) . to judge the stability of this spin liquid state
, we have to understand how the symmetries of the microscopic hamiltonian are realized in the effective continuum theory . even though the original spin hamiltonian is invariant under the full space group transformations , after the gauge theory formulation of the problem , the symmetry of the mean field hamiltonian
is realized projectively .
that is , under the symmetry transformation @xmath160 with the mapping @xmath161@xmath162@xmath163 , the spinon operator @xmath164 transforms in the following way , @xmath165 where @xmath166 is a phase factor which depends on the symmetry operation @xmath160 , and a local coordinate @xmath161 .
the group of the symmetry operations which make the mean field hamiltonian invariant is called the projective symmetry group ( psg).@xcite $ ] .
@xmath89 maps y to - y while @xmath167 induce the rotation by @xmath168 with respect to the center of the dodecagon . , width=5 ] to perform the psg analysis we have to specify the symmetry group of the spin hamiltonian .
the star lattice has the @xmath87 point group symmetry generated by the six - fold rotation symmetry with respect to the center of a dodecagon and the reflections .
however , due to the finite fluxes inside triangles , the sl@xmath124 $ ] state breaks some @xmath87 point group symmetries .
especially , the six - fold rotational symmetry is broken down to three - fold rotational symmetry .
the point group symmetry of the sl@xmath124 $ ] state is generated by the 2@xmath121/3 rotation ( @xmath169 ) around the center of the dodecagon and the reflection ( @xmath89 ) which maps @xmath170 to - @xmath170 .
the symmetry operations which generate the point group of the sl@xmath124 $ ] state are depicted in fig .
[ fig : plusminussymmetry ] .
the sl@xmath124 $ ] state is also invariant under the translations ( @xmath171 and @xmath172 ) by the lattice vectors @xmath122 and @xmath2 . in combination with the above point group symmetry ,
the translational symmetry defines the space group of the sl@xmath124 $ ] state .
in addition , the sl@xmath124 $ ] state is invariant under the combination ( @xmath173 ) of the time - reversal ( @xmath129 ) and inversion ( @xmath138 ) as well as the spin rotation . finally , it has the charge conjugation symmetry ( @xmath174 ) via the mapping @xmath175
@xmath162 @xmath176 where @xmath177 = 1 for i=1,2,3 and -1 for i=4,5,6 . under these symmetry operations the continuum field
@xmath159 transforms as follows .
@xmath178^{t } , \nonumber \\
t_{a_{1 } } : \psi & \longrightarrow e^{-\imath \frac{2 \pi}{3}\mu_{z } } \psi , \nonumber \\
t_{a_{2 } } : \psi & \longrightarrow e^{\imath \frac{2 \pi}{3}\mu_{z } } \psi , \nonumber \\
r_{y } : \psi & \longrightarrow \mu_{z } \tau_{x } \psi , \nonumber \\ c_{\frac{2 \pi}{3 } } : \psi & \longrightarrow e^{-\imath \frac{2 \pi}{3}\tau_{z } } \psi.\end{aligned}\ ] ] using the above transformation rules we can easily check that @xmath179 is the only fermion bilinear which is allowed by symmetry .
therefore the low energy hamiltonian in eq .
( [ eq : effectiveh2 ] ) is valid even after we include the fluctuations and protected by the projective symmetry group .
some details about how we have determined the transformation rule of the continuum fields are explained in the appendix [ sec : fieldsymmetry ] .
next we discuss about possible instability of the sl@xmath180 $ ] state .
the low energy effective hamiltonian ( eq . ( [ eq : effectiveh2 ] ) ) which is obtained from the perturbative expansion in powers of @xmath156 implies that the electron pocket ( at the @xmath154 point ) and the hole pocket ( at the @xmath181 point ) are nested in the large @xmath32/@xmath8 limit .
therefore the instability in the particle - hole channel with the momentum @xmath182 is expected .
the following two fermion bilinears , @xmath183=@xmath184 and @xmath185=@xmath186 , are especially important in this respect .
addition of the mass term @xmath187 to the effective hamiltonian in eq .
( [ eq : effectiveh2 ] ) leads to the mass gap of @xmath188 . since these mass terms are anticommuting with the effective hamiltonian in eq .
( [ eq : effectiveh2 ] ) , the pair ( @xmath183,@xmath185 ) opens the largest mass gap than any other pairs of possible mass terms .
interestingly , ( @xmath183,@xmath185 ) transforms nontrivially under the space group operations .
its symmetry property is consistent with some ordered state with @xmath189-type translational symmetry breaking . using the terminology defined in sec .
[ sec : grouptheory ] , ( @xmath183,@xmath185 ) transforms as the @xmath190 irreducible representation under the enlarged point group @xmath191 .
the detailed discussion on the group theory for the star lattice is given in sec .
[ sec : grouptheory ] .
+ ( e ) + $ ] and sl@xmath192 $ ] plotted along the @xmath116=@xmath3@xmath76 direction . ( a ) for sl@xmath193 $ ] .
the locations of the minimum ( maximum ) band gap are described by @xmath194 ( @xmath195 ) .
( b ) for sl@xmath192 $ ] . here
@xmath196 and @xmath197 ( @xmath198 and @xmath199 ) indicate the location of the electron ( hole ) pockets .
, width=8 ] we obtain sl@xmath193 $ ] ( sl@xmath192 $ ] ) phase by putting additional @xmath121 flux on every dodecagon on top of sl@xmath200 $ ] ( sl@xmath201 $ ] ) states . due to the introduction of the additional @xmath121 flux , the mean field description requires unit cell doubling
although the actual physical wave function maintains the translational invariance .
the spinon dispersion of the sl@xmath193 $ ] state is described in fig .
[ fig : bandhalfpi](a ) . basically , the structure of the low energy spectrum of sl@xmath193 $ ] is similar to that of sl@xmath200 $ ] , except that the number of the momentum points which support low energy excitations is doubled .
both of them are characterized by finite @xmath132 indicating the time - reversal and parity symmetry breaking .
therefore the sl@xmath193 $ ] state is also a chiral spin liquid state .
the low energy excitations can be described by the effective hamiltonian similar to eq .
( [ eq : eqdirac1 ] ) which can be obtained following the same procedure we used to derive eq .
( [ eq : eqdirac1 ] ) for sl@xmath200 $ ] . in fig .
[ fig : bandhalfpi](b ) we have drawn the low energy spinon excitation spectrum of the sl@xmath192 $ ] state .
there are two electron pockets ( around @xmath118 and @xmath202 ) and two hole pockets ( around @xmath119 and @xmath203 ) .
it is characterized by finite @xmath133 showing broken time reversal and inversion symmetry .
since the fluxes inside up - triangles and down - triangles have opposite signs , the six - fold rotational symmetry is broken down to three - fold rotational symmetry .
therefore it is another nematic spin liquid state .
according to the pioneering work by d.s.rokhsar,@xcite when the lattice system is dimerizable , the best mean field ansatz is one of dimerized states in the large - n limit of the su(n)-generalized heisenberg model . here
we call a lattice to be dimerizable when it is possible to make every site belong to a dimer and a lattice site be paired with one and only one of its neighboring site . in particular , when every dimer is lying on the link which has the maximum exchange coupling @xmath204 , the dimer state belongs to the ground state manifold of the mean field hamiltonian . in terms of the variable @xmath40
, we have finite @xmath47 only on the dimers lying on the link which has the maximum spin coupling @xmath204 .
the star lattice is dimerizable with respect to @xmath8 .
therefore when @xmath8 is larger than @xmath32 , it has a unique dimerized ground state ( we call it the @xmath8-dimer vbs ) in which every dimer is lying on an expanded link connecting neighboring triangles . in fig .
[ fig : jedimer ] we describe the geometric arrangement of singlet dimers of the @xmath8-dimer vbs phase . on the other hand ,
the rokhsar s general theorem can not be applied when @xmath32 is larger than @xmath8 .
this is because the star lattice is not dimerizable with respect to the @xmath32 links and every dimer configuration defined on the star lattice contains a finite number of dimers lying on the @xmath8 links .
therefore it is possible that the translationally invariant mean field ansatz can be the ground state even in the large - n limit .
-dimer vbs state .
this is the ground state in the large - n limit when @xmath8 @xmath205 @xmath32 .
, width=4 ] when @xmath32 @xmath205 @xmath8 , we have to maximize the number of the dimers lying on triangular links to minimize the ground state energy of dimerized states .
since every triangle can support a single dimer at most ( we call the triangle with a dimer lying on it a filled triangle ) , the remaining unpaired lattice point of the filled triangle has to be a part of the dimer lying on an expanded link . in other wards ,
every dimer lying on an expanded link is connecting two filled triangles and this describes a representative local dimer configuration of the lowest energy dimerized states when @xmath32 @xmath205 @xmath8 .
( see fig .
[ fig : jtdimer](a ) . ) using this local dimer configuration as a building block we can construct infinite number of degenerate dimerized ground states .
@xmath205 @xmath8 .
a dimer lying on an expanded link connects two neighboring filled triangles which are supporting dimers on them .
( b ) the 18-site flippable loop which consists of the alternating bright(blue ) and dark(black ) thick lines . , width=6 ] to understand how the degeneracy of these dimerized phases is lifted by fluctuations , we consider the 1/n corrections to the ground state energy . in ref . , read and sachdev investigated the 1/n correction systematically for a similar problem defined on the square lattice .
we first review the main ideas of their work briefly and extend the approach to our star lattice problem . for a given dimer configuration , @xmath206 we include the fluctuations @xmath207 as @xmath40 = @xmath206 + @xmath207 . here
@xmath206 is nonzero only on the link supporting a dimer lying on it . expanding the effective action to the quadratic order in the fluctuations ,
the ground state degeneracy of the dimerized states on the square lattice could be lifted by the following terms , @xmath208 @xmath209 n @xmath206@xmath210@xmath211@xmath212 . here
@xmath213 and @xmath214 indicate the four corners of a square plaquette .
when a pair of links lying in parallel are occupied by two dimers ( @xmath206 and @xmath211 ) and the remaining pair of the links are assigned to the fluctuations ( @xmath210 and @xmath212 ) , the @xmath208 term above can induce additional lowering of the ground state energy .
it means that dimer configurations which support the maximum number of the parallel dimer pairs span the ground state manifold .
the four - fold degenerate columnar valence bond solid is selected as the ground state following these procedures .
the above idea of the 1/n - correction can also be rephrased in the following way . for every square plaquette composed of two parallel dimers ,
we can define a loop which consists of alternating occupied and empty links . here
we call such a loop as a flippable loop@xcite because two degenerate dimer configurations are connected via a loop flip , i.e. , the interchange of the occupied and empty links .
the 1/n - correction captures the energy lowering through the resonance process which can also be described as a loop flip .
the resulting ground state ( a columnar dimer state ) supports the maximum number of the flippable loops .
this idea can be generalized to the higher order corrections and the degeneracy of dimerized states begins to be lifted from the lowest order correction corresponding to the smallest flippable loop .
marston and zeng@xcite discussed the effect of the 1/n - correction on the degeneracy lifting process for the kagome lattice antiferromagnet .
there , the first term that lifts the degeneracy involves the six - site flippable loop , which is the so - called perfect hexagon with three dimers on it .
the valence bond solid ground state of the kagome lattice , which contains 36-site within the unit cell , results from the condition of maximizing the number of the perfect hexagons.@xcite the similar idea was also applied to the square - kagome antiferromagnet.@xcite in the star lattice problem with @xmath32 @xmath205 @xmath8 , the dimerized ground states are constructed by repeating the representative local dimer configuration displayed in fig .
[ fig : jtdimer](a ) . in this ground state manifold
, the smallest flippable loop contains 18 sites with a dodecagon at the center , which is shown in fig .
[ fig : jtdimer](b ) . here when the bright ( blue ) thick link is occupied by a dimer , the neighboring dark ( black ) thick link is empty and vice versa . by interchanging the roles played by the bright ( blue ) links and the dark ( black ) links , two degenerate
dimerized phases can be connected .
therefore the fluctuation corrections pick the patterns that maximize the number of the 18-site dimer units participating in the 18-site flippable loops . in fig .
[ fig : jtdimerconfig](a ) we show the valence bond solid order which has the maximum number of the 18-site flippable loops . among the six neighboring dodecagons around an 18-site dimer unit , three can be the centers of the 18-site dimer units .
this is in contrast to the case of the kagome lattice problem .
there , none of the six neighboring hexagons around a perfect hexagon can be perfect hexagons .
in fact , the valence bond solid states on the star lattice have similarity with those on the square lattice . in the case of the square lattice , among the four neighboring square plaquettes around a central plaquette supporting two parallel dimers , half of them ( two square plaquettes ) can support two parallel dimers . by maximizing the number of the square plaquette composed of two parallel dimers ,
the columnar valence bond solid emerges . based on the similarity with the square lattice problem
, we can call the valence bond solid in fig . [
fig : jtdimerconfig](a ) as a columnar 18-site valence bond solid . in this figure
all the dodecagons except the central one support the 18-site dimer unit .
the three - fold degeneracy of the columnar 18-site vbs comes from the broken translation symmetry .
in addition we also consider another low energy valence bond order which is displayed in fig .
[ fig : jtdimerconfig](b ) . here
the 18 links around the 18-site unit have the same finite value of @xmath206 . in analogy with the square lattice problem
this phase can be called as the 18-site box vbs phase . the translational symmetry breaking results in the three - fold degeneracy in this phase as well . in the recent numerical study by g. misguich et al.@xcite , this phase was suggested as a possible valence bond solid ground state when @xmath32 @xmath205 1.3 @xmath8 .
we expect that the 1/n corrections would select the 18-site columnar vbs as the ground state that can maximize the resonance energy gain from the smallest flippable loop . however , since the length of the flippable loop is quite large compared to those on the square and kagome lattice problem , the energy difference of the two candidate valence bond solids in fig .
[ fig : jtdimerconfig ] could be very small .
the columnar and box 18-site vbs phases discussed in the above section break the lattice translational symmetry and are described by the enlarged @xmath0 unit cell . here
we perform the detailed symmetry analysis on the bond order that are compatible with the @xmath0 enlarged unit cell .
+ and @xmath215 corresponding to the @xmath0 ordered states . , width=8 ] \(a ) + + the star lattice has the @xmath87 point group symmetry .
the twelve elements of the @xmath87 group are as follows .
@xmath216 where @xmath217 means the rotation by 2@xmath121/6 around the center of a dodecagon and @xmath218 is the identity operator .
@xmath219 indicates the reflection with respect to a plane orthogonal to the lattice plane .
the reflection planes ( denoted by @xmath53 ) are described in the fig .
[ fig : twofold](a ) .
the space group , @xmath220 , of the star lattice is generated by the translation group , @xmath221 , and the @xmath87 point group .
an element of the space group can be written using the seitz operator @xmath222 , where @xmath223 is an element of the @xmath87 group and @xmath224=@xmath225+@xmath226 .
( @xmath227 and @xmath228 are integers . ) the action of a seitz operator on a lattice point @xmath229 is defined as @xmath230 .
note that the translation group , @xmath221 , is an invariant subgroup of the space group @xmath220 and the point group @xmath87 is the corresponding factor group , i.e. , @xmath87= @xmath220/@xmath221 . to understand the symmetry of the enlarged unit cell
, we define another translation group , @xmath231 , whose elements can be written as , @xmath232 where @xmath233 = @xmath234 and @xmath215 = @xmath235 are the lattice vectors corresponding to the @xmath0 ordered state.(see fig .
[ fig : twofold](b ) ) since @xmath231 is an invariant subgroup of the space group @xmath220 , the enlarged point group @xmath191 can be defined as the factor group @xmath191= @xmath220/@xmath231 .
therefore the elements of the space group can be written as @xmath236 @xmath237 @xmath238 in which @xmath239 is an element of the enlarged point group , @xmath191 .
( note that similar approach was used by hermele et al .
, @xcite to investigate the symmetry properties of the object invariant under the translations by 2@xmath122 and 2@xmath2 on the kagome lattice . )
the construction of the @xmath191 group is straightforward , whose elements can be written using the seitz operator @xmath240 with @xmath241=0 , @xmath122 , @xmath2 .
the 36 elements of the @xmath240 can be grouped into the nine conjugate classes , @xmath242 .the character table of the enlarged point group @xmath191 [ cols="^,^,^,^,^,^,^,^,^,^",options="header " , ] measured relative to the @xmath5-dimer vbs state shown in fig.[fig : jedimer ] which has an energy @xmath243 per spin .
( b ) energies for @xmath244 measured relative to the @xmath0 vbs state shown in fig.[fig : jtdimerconfig ] which has an energy @xmath245 per spin . note that sl[0,0,@xmath120 ( solid line is a guide to the eyes ) has the lowest energy over the whole parameter space .
sl[0,0,0 ] has the second lowest energy for a wide range for @xmath246 . for @xmath247 , the energy per spin of all six spin liquid ansatzes
are given in table.[table : energy ] .
( these computations were carried out on a system with @xmath248 unit cells , i.e. , with @xmath249 spins , and the statistical error bars on the energy are of the order of the symbol size ) .
, title="fig:",width=8 ] measured relative to the @xmath5-dimer vbs state shown in fig.[fig : jedimer ] which has an energy @xmath243 per spin .
( b ) energies for @xmath244 measured relative to the @xmath0 vbs state shown in fig.[fig : jtdimerconfig ] which has an energy @xmath245 per spin . note that sl[0,0,@xmath120 ( solid line is a guide to the eyes ) has the lowest energy over the whole parameter space .
sl[0,0,0 ] has the second lowest energy for a wide range for @xmath246 . for @xmath247 , the energy per spin of all six spin liquid ansatzes
are given in table.[table : energy ] .
( these computations were carried out on a system with @xmath248 unit cells , i.e. , with @xmath249 spins , and the statistical error bars on the energy are of the order of the symbol size ) .
, title="fig:",width=8 ] in the above mean field calculation , the single occupancy constraint is imposed only on average .
therefore the mean field wave functions contain unphysical states with zero or two fermions at a point . to obtain physical spin wave functions
we therefore perform a numerical gutzwiller projection on the mean field wave functions .
the ground state energies of the projected states are computed numerically using the variational monte carlo ( vmc ) method@xcite .
the resulting energies are displayed in fig.[fig : energy_wfn ] for a range of @xmath16 where we have optimized the state with respect to @xmath250 for each value of @xmath16 .
table [ table : energy ] shows the numerical energy values for @xmath16=2.0 to facilitate a comparison with the mean field numerics .
we find that gutzwiller projection dramatically changes the relative ordering of the various states , and that the state sl@xmath26 $ ] appears , upon projection , to be the lowest energy spin liquid over the entire parameter range . according to the projected wave function study , sl[0,0,@xmath120
state is the ground state over a wide parameter space .
however , the sl[0,0,@xmath120 state is unstable due to the confinement in the 2 + 1 dimensional pure gauge theory .
it is also inconsistent with our expectation for the @xmath251 limit .
when @xmath251 , the @xmath5-dimer vbs phase ( see fig .
[ fig : jedimer ] . ) is the exact ground state .
in addition the recent exact diagonalization study shows that the @xmath5-dimer vbs phase remains as the ground state up to the isotropic limit of @xmath252.@xcite therefore sl[0,0,@xmath120 state should have higher energy than @xmath5-dimer vbs phase at least in some finite range of @xmath253 .
this discrepancy comes from the lack of the interdimer interaction in the decoupled dimer limit . for the description of dimerized phases beyond the decoupled dimer limit
, we undertake the self - consistent bond operator approach.@xcite if the correction coming from the inter - dimer interaction is significant , we also have to check the possibility that the true ground state is a valence bond solid even when @xmath254 . in the bond operator formulation ,
the dimer singlet degrees of freedom are used as natural building blocks and the quantum corrections coming from the triplet fluctuations can systematically be investigated.@xcite here we present a brief explanation of the bond operator formulation .
let us consider the two @xmath255 spins constituting a dimer singlet , @xmath256 and @xmath257 .
the hilbert space is spanned by four states that can be taken as a singlet state , @xmath258 , and three triplet states , @xmath259 , @xmath260 and @xmath261 . then , the singlet and triplet boson operators are introduced such that each of the above states can be created from the vacuum @xmath262 as follows : @xmath263 to eliminate unphysical states from the enlarged hilbert space , the following constraint needs to be imposed on the bond - particle hilbert space : @xmath264 where @xmath265 and @xmath266 , and we adopt the summation convention for the repeated indices hereafter unless mentioned otherwise .
constrained by this equation , the exact expressions for the spin operators can be written in terms of the bond operators : @xmath267 where @xmath268 is the third - rank totally antisymmetric tensor with @xmath269 . utilizing the bond operator representation of spin operators , the heisenberg spin hamiltonian in eq .
( [ eqn : hamiltonian ] ) can be rewritten solely in terms of bond particle operators .
since all dimers of @xmath5-dimer vbs phase are symmetry equivalent , the singlet condensate density @xmath270 and the chemical potential @xmath271 can be set to be @xmath270 = @xmath272 and @xmath271=@xmath273 in our mean field theory .
here @xmath274 denotes the location of dimers .
the hard - core constraint on the bond - particle operators is imposed by adding the following lagrange multiplier term , @xmath275 .
the resuling hamiltonian can be written as follows : @xmath276 where @xmath277 and @xmath278 in the above , @xmath27 is the number of unit cells and @xmath279.\end{aligned}\ ] ] ] to that of @xmath8-dimer vbs for 0@xmath280 . to emphasize the importance of the interdimer interactions we present the energies of @xmath8-dimer vbs obtained in three different ways .
the decoupled indicates the energy of the decoupled dimers .
quartic ( quadratic ) is the energy from the bond operator theory with ( without ) the quartic interaction effect .
, width=7 ] the quartic interactions between triplet particles are decoupled using the mean field order parameters @xmath281 and @xmath282 , where @xmath283 and @xmath284 . here
@xmath281 and @xmath282 denote the diagonal and off - diagonal triplet correlations between neighboring dimers .
these two order parameters @xmath281 and @xmath282 together with @xmath272 and @xmath273 are determined self - consistently by solving the coupled saddle point equations.@xcite the ground state energy of @xmath5-dimer vbs phase obtained from the self - consistent bond operator calculation is displayed and compared to the energy of sl[0,0,@xmath120 state in fig . [
fig : energy_jedimer ] .
here we have obtained the energy of @xmath5-dimer vbs phase in three different ways .
if we neglect the inter - dimer couplings ( the decoupled dimer limit ) completely , the energy is independent of @xmath6/@xmath5 .
the inclusion of the inter - dimer interaction lowers the ground state energy significantly . in the end
, @xmath5-dimer vbs phase has the lower ground energy than the sl[0,0,@xmath120 state over the entire parameter range of 0@xmath285@xmath6/@xmath5@xmath2851 when we include the quartic interactions .
the inter - dimer interactions generate huge correction to the ground state energy of dimerized phases . ] state .
( a ) energetics for @xmath32@xmath205@xmath8 .
the energies of the columnar 18-site vbs are obtained in three different ways as in fig.[fig : energy_jedimer ] .
( b ) the relative energies obtained by subtracting the decoupled dimer energy . note that there is a level crossing around @xmath32/@[email protected] between the energy of sl[0,0,@xmath120 state and that of the columnar 18-site vbs including the quartic interaction .
, width=8 ] now we concentrate on the other limit where @xmath6@xmath205@xmath5 .
in contrast to the @xmath5@xmath287@xmath6 limit , it is nontrivial to identify the ground state even when we restrict our attention to valence bond solid phases . taking into account the information from the exact diagonalization study and 1/n fluctuation from the large - n limit , we suggest the columnar 18-site vbs phase as a promising candidate for the ground state as explained below .
we apply the bond operator approach to the columnar 18-site vbs phase .
the nine dimers within the unit cell can be divided into two groups .
one group is made of the six dimers lying on the triangular links .
note that all these six dimers are lying on a dodecagon .
( see the central dodecagon in fig .
[ fig : jtdimerconfig](a ) . )
we call such a dodecagon surrounded by six dimers as a perfect " dodecagon .
the remaining three dimers lying on the expanded links make the other group .
every dimer belonging to the same group is symmetry equivalent as one can easily notice from the patterns around the central perfect " dodecagon in fig .
[ fig : jtdimerconfig](a ) . to apply the bond operator approach we have to introduce two independent sets of order parameters to distinguish the two different groups of dimers .
we use @xmath288 and @xmath289 ( @xmath290 and @xmath291 ) to indicate the singlet condensate density and the chemical potential corresponding to the expanded ( triangular ) link . to decouple the quartic triplet interactions we introduce two sets of the order parameters , that is , @xmath292@xmath293 , @xmath294@xmath295 and @xmath292@xmath296 , @xmath297@xmath295 .
@xmath293 ( @xmath294 ) describes the diagonal ( off - diagonal ) correlation between the neighboring dimers lying on a perfect " dodecagon . on the other hand @xmath296 ( @xmath297 )
describes the diagonal ( off - diagonal ) correlation between a dimer lying on an expanded link and its neighboring dimer lying on a perfect " dodecagon .
we have determined the eight parameters @xmath298 , @xmath299 ( @xmath53 = t or e ) , and @xmath300 , @xmath301 ( @xmath302 = @xmath303 or @xmath304 ) self - consistently by solving the coupled saddle point equations .
the self - consistent solution shows that @xmath296=@xmath297=0 , @xmath305=1 and @xmath306=-3/4@xmath8 .
since @xmath296 and @xmath297 describe the coupling between the dimers lying on expanded links and the dimers lying on perfect " dodecagons , these two groups of dimers are completely decoupled when @xmath296=@xmath297=0 . in this situation ,
every dimer lying on expanded links is decoupled from the surrounding , leading to @xmath305=1 and @xmath306=-3/4@xmath8 .
the triplet fluctuations are confined inside every isolated perfect dodecagon , which is reflected in the finite @xmath293 and @xmath294 values .
this interesting structure would result in the highly localized triplet excitation spectrum .
the ground state energies of the columnar 18-site vbs and sl[0,0,@xmath120 are compared in fig .
[ fig:00pi_dimer ] .
the energies of the columnar 18-site vbs are obtained in three different ways again .
that is , for the decoupled dimer limit , including the inter - dimer coupling neglecting quartic interactions , and finally including the inter - dimer quartic interactions .
for clarity we also calculated the energy difference relative to the decoupled dimer energy as shown in fig .
[ fig:00pi_dimer](b ) .
interestingly , there is a critical ratio @xmath307@xmath286 2.4 beyond which the columnar 18-site vbs becomes the ground state when we include the quartic triplet interactions .
even though the critical ratio @xmath307 is a bit larger than the suggested phase boundary from the numerical study@xcite , the existence of the critical values of @xmath307 is quite encouraging .
in particular , because the slopes of the lines in fig .
[ fig:00pi_dimer](a ) are almost parallel , small additional energy correction could induce a large shift of the crossing point as shown in fig .
[ fig:00pi_dimer](b ) .
since the simple hartree - fock approximation does not take into account the fluctuations coming from the cooperative interaction between the dimers on the expanded links and those on the perfect dodecagons , we expect that the quantum correction beyond the hartree - fock limit could shift the energy level crossing point down to @[email protected] as suggested by the numerical study.@xcite
summarizing the previous discussions , @xmath5-dimer vbs phase is the ground state for @xmath308 while the columnar 18-site vbs is the ground state in the opposite limit of @xmath309 .
( here @xmath310 . )
although this result is obtained based on the energy comparison with sl[0,0,@xmath120 state , one may still expect that the two valence bond solid states are intimately related to the sl[0,0,@xmath120 state .
in particular , the @xmath5-dimer vbs and the columnar 18-site vbs may arise as a consequence of the confinement in the sl[0,0,@xmath120 spin liquid state . in this section
we describe the possible relation between these two valence bond solid phases and the sl[0,0,@xmath120 state . due to the finite spinon gap , the u(1 ) gauge field is the only low energy excitation in the sl[0,0,@xmath120 state in the long wavelength limit . since the compact u(1 )
gauge theory without matter field is confining in 2 + 1 dimension , we expect that the monopole proliferation would lead the sl@xmath26 $ ] ansatz to some confined phases . to understand the properties of the confined phases resulting from the monopole condensation , we have to determine the symmetry properties of the monopole operators . here
we discuss the possibility that the monopole operators are invariant under all possible symmetry transformations .
the @xmath5-dimer vbs phase , which is the ground state for @xmath311 , is invariant under space group operations .
the bond ordering pattern of the @xmath5-dimer vbs phase belongs to the trivial @xmath108 irreducible representation of the @xmath87 point group .
therefore if we interpret the @xmath8-dimer vbs phase to be induced by the confinement transition , which is reasonable in the limit of @xmath312 , this reflects the fact that monopole operators are invariant under symmetry transformations .
irreducible representation .
the thick solid ( red ) link has @xmath313 = 1 while the thin solid link has @xmath313=0 .
( a ) @xmath5-bond order .
( b ) @xmath6-bond order .
, width=7 ] extending the group theory analysis we performed in sec .
[ sec : grouptheory ] , we investigate all possible bond orders invariant under the space group operations .
these are displayed in fig .
[ fig : a1order ] .
here we have finite singlet correlation ( @xmath314 ) only on the thick solid ( red ) links . the bond order in fig .
[ fig : a1order](a ) is nothing but the @xmath8-dimer vbs phase . on the other hand , the bond order in fig .
[ fig : a1order](b ) has finite @xmath313 only on the triangular links ( we call it as a @xmath32-bond ordered phase ) .
we expect the @xmath32-bond ordered phase is the natural low energy bond ordering pattern when @xmath32@xmath287@xmath8 . since the arbitrary superposition of these two orders follows the same @xmath108 irreducible representation , we expect the actual ground states would have finite @xmath313 values over all the links on the lattice . however , it is natural to expect that the @xmath313 on the expanded ( triangular ) link would be larger than that on the triangular ( expanded ) link when @xmath8 @xmath205 @xmath32 ( @xmath8 @xmath315 @xmath32 )
. therefore the bond ordered phase corresponding to the @xmath108 irreducible representation successfully describes the low energy manifold over the whole parameter range of @xmath32/@xmath8 .
interestingly , the recent work by choy and kim@xcite has suggested that the same bond ordered states are the ground states of the same model hamiltonian in the strong quantum limit based on the bosonic sp(n ) approach . since the spinon bandstructure of the sl[0,0,@xmath120 state does not change qualitatively by varying @xmath32/@xmath8
, we expect that the change of @xmath32/@xmath8 ratio would not affect the trivial monopole quantum number .
therefore if the ordered phase is coming from the confinement transition , it will transform trivially under the symmetry operations .
however , it is also possible that the instability of the sl[0,0,@xmath120 state is caused by the interactions between spinons leading to some broken symmetry phases .
we discuss about this possibility in the following section . here
we investigate the instability of the sl[0,0,@xmath120 state coming from the interactions between spinons and the symmetry properties of the resulting ordered phase .
especially we focus on the instability towards the states with the @xmath0-type translational symmetry breaking . as shown in fig .
[ fig : band00pi ] , the valence band is completely flat without any preferred momentum .
however , the conduction band supports several dispersion minima .
the four minimum points of the conduction band are given by @xmath316 interestingly if we double the vectors connecting neighboring minimum points , they sit on the brillouin zone corners which are nothing but the momentum corresponding to the @xmath0 ordering
. motivated by this observation we study the symmetry properties of the bound states made of low energy fermions near the conduction band minima .
we introduce the fermion fields @xmath317 which describe the low energy excitations near the four conduction band minima @xmath114 ( @xmath30 = 1,2,3 and 4 ) , @xmath318 where @xmath319 is the eigenvector of the mean field hamiltonian at the momentum @xmath114 , and @xmath320 is a slowly varying fermion field near the conduction band minimum , with @xmath34 labelling the twelve sites within the unit cell .
to determine the symmetry of bound states made of the above low energy fermions , we have to understand how the symmetries of the microscopic hamiltonian are realized in the effective continuum fields , @xmath317 . here
we follow the same procedure which we use to determine the transformation properties of the continuum field for the sl@xmath124 $ ] state in sec . [
sec : properties ] . to extract the necessary information on the transformations of the continuum field , we consider a finite system of a 6@xmath32112 unit cell . by solving the mean field hamiltonian on this finite system , we determine the properties of the eigenvectors at the four momentum points , @xmath114 .
the detailed explanation of the procedure as to how to determine the symmetry of the continuum fields is discussed in the appendix [ sec : fieldsymmetry ] . through the projective symmetry group analysis , we determine the following transformation properties of the continuum fields , @xmath322 @xmath323 @xmath324
@xmath325 we first investigate the symmetry of all possible fermion bilinears which can be written as @xmath326 .
for the description of the 4 by 4 unitary matrix @xmath327 we introduce two sets of the pauli matrices @xmath82 and @xmath328 . here
@xmath82 is acting on the space spanned by either ( @xmath329,@xmath330 ) or ( @xmath331,@xmath332 ) . on the other hand @xmath45
is defined in the space spanned by ( @xmath329,@xmath331 ) or ( @xmath330,@xmath332 ) . among the sixteen possible bilinears ,
there are only two terms which can form a basis of the enlarged point group @xmath191
. these are @xmath333 and @xmath334 which transform as the @xmath95 and @xmath108 irreducible representation of the enlarged point group @xmath191 , respectively . however , since these two bilinears have zero total momentum , they can not describe the @xmath0 type symmetry breaking .
the net momentums carried by the other fourteen bilinears are neither zero nor @xmath154=(@xmath335,0 )
. therefore the transformation properties of them are not compatible with the symmetry of the @xmath0 type enlarged unit cell .
next we consider the instability in the particle - particle channels .
defining the pairing amplitude as @xmath336@xmath41@xmath337 , we have sixteen different @xmath336 .
we first omit the indices for the spin degrees of freedom and study how they transform under the space group symmetry operations .
it can easily be checked that the sixteen pairing amplitudes are divided into the four different sets @xmath338 , @xmath339 , @xmath340 and @xmath341 which transform independently .
the six - component vectors @xmath342 and the two - component vectors @xmath343 are given by @xmath344 @xmath345 @xmath346 under a space group symmetry operation s , they transform in the following way , @xmath347 where @xmath348 ( @xmath349 ) is the 6 by 6 ( 2 by 2 ) matrix representing the symmetry operation s , whose detailed expressions are displayed in the appendix [ sec : matrices ] . since the pairing amplitude @xmath336 is not a gauge invariant object , we have to look into the symmetry of the bilinears such as @xmath350@xmath41@xmath351@xmath352 .
even though the number of all possible tensors @xmath350 is very large , we can reduce the complexity of the symmetry analysis by focusing on the objects carrying the momentum compatible with the @xmath0 ordering .
this idea leads us to exclude @xmath350 made of the basis @xmath339 and @xmath341 .
in addition , @xmath338 and @xmath340 have opposite spin parities , that is , @xmath338 is spin singlet while @xmath340 is spin triplet . therefore all we have to consider are the terms like @xmath353@xmath354@xmath338 and @xmath355@xmath356@xmath340 .
first , we define a set of pauli matrices @xmath82 acting on the space spanned by @xmath340 .
using the transformation properties of @xmath340 under the space group , we obtain a pair ( @xmath355@xmath357@xmath340,@xmath355@xmath358@xmath340 ) which has the momentum @xmath154=(@xmath335,0 ) and transforms as the @xmath190 irreducible representation of the enlarged point group , @xmath191 . to understand the symmetry of @xmath353@xmath354@xmath338
, we introduce a set of the gell - mann matrices @xmath359 ( @xmath135=1, .... ,8)@xcite as well as the pauli matrices @xmath82 . for convenience
, we also define the matrix , @xmath360=@xmath361@xmath362 where @xmath362 is the 3@xmath3213 identity matrix .
the gell - mann matrices are acting on the space spanned by either ( @xmath363 , @xmath364 , @xmath365 ) or ( @xmath366 , @xmath367 , @xmath368 ) and @xmath82 connects these two three - component vectors . here
@xmath369 indicates the nth component of @xmath338 .
we have examined the symmetry of all possible bilinears @xmath353@xmath354@xmath338 and found out that there are only two sets of bilinears which have the momentum @xmath154 compatible with the @xmath0 ordering .
these are given by @xmath370 @xmath371 in the above @xmath372 transforms as the two dimensional @xmath190 irreducible representation of the enlarged point group @xmath191 .
on the other hand @xmath373 constitutes a basis of the four dimensional @xmath282 irreducible representation .
therefore the instability given by @xmath372 has the symmetry consistent with the @xmath0 orders which we have discussed in detail in previous sections . since we are considering an instability from a gapped phase , to stabilize the resulting ordered state , the condensation energy should be larger than the excitation gap . however , because the magnitude of the energy gap reduces as @xmath6/@xmath5 increases , the instability can occur beyond the critical value of @xmath6/@xmath5 . finally ,
since the condition of @xmath374 does not constrain the magnitude of @xmath375 , @xmath375 can have both zero and nonzero values .
if @xmath376 with @xmath374 , we have @xmath377 spin liquid supporting fractionalized quasi - particles and breaking the translational symmetry at the same time .
this state is similar to the amperean paired state@xcite , which is recently suggested as a possible ground state of the organic compound @xmath10-(bedt - ttf)@xmath378cu@xmath378(cn)@xmath379 . on the other hand , if @xmath380 while @xmath374 , we have more conventional phase transforming as an @xmath190 irreducible representation . in conclusion ,
the symmetry analysis of the low energy fermions near the conduction band minima shows that the instability of the sl[0,0,@xmath120 state from the particle - particle channel supports valence bond solid phases , which have the @xmath189 unit cell transforming as @xmath190 irreducible representations of the enlarged point group @xmath191 .
remember that the @xmath8-dimer vbs phase is induced via the monopole condensation from the sl[0,0,@xmath120 state for small @xmath381 limit .
therefore we can obtain the valence bond solid ground states both for @xmath7 and @xmath9 limits from the instability of the sl[0,0,@xmath120 state .
we next assess the stability of the sl@xmath26 $ ] spin liquid towards columnar dimer order in the guzwiller projected state . in order to do this ,
we include a parameter @xmath382 in the mean field hamiltonian which corresponds to strengthening the fermion hopping on those bonds which dimerize in the ` classical ' columnar dimer state , shown in fig .
[ fig : jtdimerconfig](a ) , and gutzwiller project the resulting state . clearly ,
if @xmath383 , the resulting wave function will be precisely the ` classical ' columnar dimer pattern .
the energy change of the weakly distorted state as a function of the distortion parameter @xmath382 serves as a measure of the inverse susceptibility of the sl@xmath26 $ ] state towards columnar dimer order . as seen from fig .
[ fig : energy_dim ] , the sl@xmath26 $ ] state is stable , with a positive inverse susceptibility , for @xmath384 , but is unstable , with a negative inverse susceptibility for @xmath385 .
further , the optimal @xmath382 appears to increase continuously for @xmath385 .
this suggests that the sl@xmath26 $ ] state possibly undergoes a continuous transition into a state with @xmath0 columnar dimer order at @xmath386 .
we discuss the phase transition more carefully in the next section .
$ ] spin liquid state upon including a distortion @xmath382 corresponding to increased fermion hopping amplitude on the dimerized bonds of the @xmath0 columnar dimer state shown in fig . [ fig : jtdimerconfig](a ) . for @xmath387 ,
the sl@xmath26 $ ] state is found to be stable against this distortion , while it is seen to be unstable for @xmath388 .
( this calculation was carried out on a system with @xmath389 unit cells , i.e. , with @xmath390 spins .
the statistical errors on the computed energy are of the order of the symbol size . ) , title="fig:",width=4 ] $ ] spin liquid state upon including a distortion @xmath382 corresponding to increased fermion hopping amplitude on the dimerized bonds of the @xmath0 columnar dimer state shown in fig . [ fig : jtdimerconfig](a ) . for @xmath387 ,
the sl@xmath26 $ ] state is found to be stable against this distortion , while it is seen to be unstable for @xmath388 .
( this calculation was carried out on a system with @xmath389 unit cells , i.e. , with @xmath390 spins .
the statistical errors on the computed energy are of the order of the symbol size . ) , title="fig:",width=4 ] $ ] spin liquid state upon including a distortion @xmath382 corresponding to increased fermion hopping amplitude on the dimerized bonds of the @xmath0 columnar dimer state shown in fig . [ fig : jtdimerconfig](a ) .
for @xmath387 , the sl@xmath26 $ ] state is found to be stable against this distortion , while it is seen to be unstable for @xmath388 .
( this calculation was carried out on a system with @xmath389 unit cells , i.e. , with @xmath390 spins .
the statistical errors on the computed energy are of the order of the symbol size . ) , title="fig:",width=4 ] $ ] spin liquid state upon including a distortion @xmath382 corresponding to increased fermion hopping amplitude on the dimerized bonds of the @xmath0 columnar dimer state shown in fig . [
fig : jtdimerconfig](a ) . for @xmath387 ,
the sl@xmath26 $ ] state is found to be stable against this distortion , while it is seen to be unstable for @xmath388 .
( this calculation was carried out on a system with @xmath389 unit cells , i.e. , with @xmath390 spins .
the statistical errors on the computed energy are of the order of the symbol size . ) , title="fig:",width=4 ]
the results in previous sections show that there is a phase transition between the @xmath8-dimer vbs phase and the columnar 18-site vbs phase with increasing @xmath16 . to describe the phase transition between these two vbs phases , we construct a landau - ginzburg free energy introducing a two - component vector ( @xmath391,@xmath392 ) , which transforms as an @xmath190 irreducible representation of the enlarged point group @xmath191 .
since the vbs phases are time - reversal invariant , we can use two real numbers , @xmath391 and @xmath392 .
the landau - ginzburg free energy can be written using all possible invariants made of @xmath391 and @xmath392 .
in particular , it is important to note that there is a third order invariant , which belongs to the @xmath108 irreducible representation of the following decomposition , @xmath393 in terms of @xmath391 and @xmath392 , the third order invariant is given by @xmath394 .
straightforward extension of the same group theoretical analysis to quartic order shows that there is only one quartic invariant of @xmath395 .
collecting all invariants up to quartic order , the landau - ginzburg free energy density is written as @xmath396 for convenience we define a complex variable @xmath397 as follows , @xmath398 then the free energy density is given by @xmath399 given @xmath400 and @xmath401 , the above mean field free energy predicts two different phases separated by a first order transition point at @xmath402 . for @xmath403
, we have a disordered phase with @xmath404 .
on the other hand , ordered phases with @xmath405 and @xmath406 ( n=0,1,2 ) appear when @xmath407 .
the three ordered states describe the three - fold degenerate vbs phases with @xmath0 pattern . considering slow spatial variation of @xmath397 ,
the euclidean landau - ginzburg effective action is given by @xmath408 the above effective action is nothing but the action for the three dimensional @xmath409-clock model .
previous monte carlo simulations suggest that the phase transition associated with the @xmath409 symmetry breaking is weakly first order.@xcite the wave function numerics in the previous section , however , are not inconsistent with this scenario considering the small system size .
in addition , the nature of the transition may not have been completely settled.@xcite
-dimer vbs phase for @xmath410 .
( b ) triplet dispersion of the columnar 18-site vbs phase for @xmath411 . ,
width=8 ] in fig .
[ fig : tripletdispersion ] , we plot the triplet dispersions of the two vbs phases , which are obtained from the bond operator mean field theory .
the triplet dispersion of the @xmath8-dimer vbs state is shown in fig .
[ fig : tripletdispersion ] ( a ) . since the unit cell of the @xmath8-dimer vbs state is composed of three dimers , we have three triplet modes in the spectrum . here
we neglect the fact that each triplet particle has three components ( x , y , and z ) when we count the number of bands .
interestingly , the lowest band is flat and touches another dispersive band at the brillouin zone center .
the emergence of this band touching has a topological origin.@xcite the flat band reflects the existence of localized eigenstates .
in fact , there are two different types of localized eigenstates .
the first set is given by states which are confined within dodecagons .
each dodecagon supports a single localized eigenstate .
in addition , there are non - contractible " loop states constituting the second group of localized eigenstates .
in contrast to the states confined within dodecagons , these loop states extend over the whole lattice system one - dimensionally .
the topological characteristics of these loop states can be easily understood using the periodic boundary condition , under which a two - dimensional system has a torus geometry . in this situation , there are two independent loop states winding the torus once . since these states can not be shrunk to points , they are non - contractible " . counting the number of independent localized states carefully , we see that the number of degenerate localized eigenstates is larger than the number of dimers on the lattice.@xcite it means that a single flat band is not enough to support all independent localized eigenstates
. therefore additional degrees of freedom must be provided by another band , leading to the band touching .
the triplet dispersion of the columnar 18-site vbs state is displayed in fig .
[ fig : tripletdispersion ] ( b ) .
we have nine triplet bands which are all flat within our mean field approach .
the degeneracies of the flat bands are given by 3 , 1 , 2 , 2 , and 1 counting from the bottom to the top bands .
in contrast to the case of the @xmath5-dimer vbs phase , the flat structure emerges simply because the dimers on expanded links are completely decoupled from those on neighboring perfect " dodecagons .
the lowest flat band is triply degenerate , which comes from the three dimers on expanded links .
the remaining six dimers of the unit cell lying on perfect " dodecagons constitute the other six bands with higher energies .
therefore we expect that the nature of the valence bond solid ground states can be demonstrated via neutron scattering experiments measuring triplet dispersion spectra .
in summary , we have shown that the ground state of the nearest neighbor heisenberg model on the star lattice undergoes a phase transition from the @xmath8-dimer vbs phase " which respects all lattice symmetries to the columnar 18-site vbs phase " which exhibits @xmath0 order with increasing @xmath16 . from the landau - ginzburg analysis
, this appears to be a conventional quantum phase transition which is described as the thermal transition of the @xmath412 dimensional @xmath409-clock model .
if @xmath413 variants of the organic iron - acetate magnet can be synthesized , they would be particularly good candidates to study the phase diagram discussed in this paper since it may be possible to pressure tune the ratio @xmath16 significantly in such systems .
both vbs states obtained here would exhibit a spin gap in uniform susceptibility measurements .
we expect a direct signature of the 18-site vbs order to appear in x - ray scattering or neutron diffraction studies which would see a change in the crystal periodicity .
ignoring coupling to phonons , the 18-site vbs state should exhibit a thermal transition in the universality class of the @xmath409-clock model in @xmath414 dimensions .
in addition , the two vbs phases exhibit quite distinct behaviors in their triplet excitation spectra as discussed above which could be tested using inelastic neutron scattering experiments .
we thank daniel podolsky , tingpong choy , fa wang , and jason alicea for helpful discussions .
this work was supported by the nserc of canada , the canada research chair program and the canadian institute for advanced research .
ap acknowledges support from the sloan foundation , the connaught foundation and an ontario era .
here we present the details of how we have derived the low energy effective hamiltonians of the spin liquid states discussed in sec .
[ sec : properties ] . for the gauge choice depicted in fig .
[ fig : fluxconfig](a ) , the mean field hamiltonian corresponding to the sl@xmath125 $ ] can be written in momentum space as , @xmath415 in which @xmath416 where @xmath417=@xmath418 and @xmath419=@xmath420 .
@xmath146 , @xmath34 are indices for the six sites inside a unit cell and @xmath421=@xmath422/@xmath423 . the indices for the spin quantum number are dropped for simplicity . here
we define the fourier transformation via @xmath424 = @xmath425@xmath426 @xmath427@xmath428 . as described in fig .
[ fig : fluxband](b ) , the conduction ( valence ) band shows the dispersion minimum ( maximum ) at the momentum @xmath150 .
the energy eigenvalues of @xmath429 at @xmath79=@xmath150 are given by @xmath430 where @xmath140 indicates the two momentum position @xmath150 and 1 and 2 represent the conduction ( 1 ) and valence ( 2 ) bands , respectively .
to make the analytic treatment of the problem possible we focus on the small @xmath421 limit .
note that the overall spinon band structure does not change upon varying @xmath421 .
expanding the energy eigenvalues in powers of @xmath421 , we get @xmath431 the corresponding eigenvectors are @xmath432 where the superscript @xmath129 means taking transposition .
the above eigenvalues and eigenvectors satisfy @xmath433@xmath434 = @xmath435@xmath434 ( @xmath53=1 and 2 ) correctly up to the third order in @xmath421 .
now we want to construct the effective hamiltonian describing the states which have small momentum deviation from @xmath150 , i.e. , states with @xmath79=@xmath150+@xmath436 .
we first define @xmath437 @xmath41 @xmath438-@xmath433 . keeping terms which are first order in @xmath436 and projecting them into the low energy space spanned by the eigenvectors @xmath434 ( @xmath53 = 1 , 2 )
, we obtain the following effective hamiltonian , @xmath439 finally , defining the continuum fermion fields using the spinon variables as @xmath440 we arrive at the low energy effective hamiltonian written in eq .
( [ eq : eqdirac1 ] ) which is nothing but the hamiltonian for the massive dirac particles .
the low energy hamiltonian corresponding to the sl@xmath124 $ ] can be obtained following the similar steps used to construct the massive dirac hamiltonian of the sl@xmath125 $ ] . adopting the flux configuration depicted in fig .
[ fig : fluxconfig](b ) , we have an electron pocket centered at the momentum @xmath154 and an hole pocket centered at the momentum -@xmath154 . at each momentum
@xmath441 there is a linear band touching as is shown in fig .
[ fig : fluxconfig](b ) .
the energy eigenvalues of the degenerate states at @xmath79=@xmath441 are given by @xmath442 where @xmath443 ( @xmath444 ) represents the degenerate energy eigenvalue at the momentum @xmath154 ( -@xmath154 ) .
we choose the corresponding eigenvectors in the following way , @xmath445 satisfying @xmath446@xmath434 = @xmath447@xmath434 ( @xmath53=1 and 2 ) correctly up to the third order in @xmath421 .
the first order perturbation theory combined with the projection into the low energy space spanned by @xmath434 ( @xmath53=1 , 2 ) leads to the following hamiltonian , @xmath448 using the continuum fermion field defined in eq .
( [ eq : continuumfield ] ) we obtain the low energy effective hamilonian displayed in eq .
( [ eq : effectiveh2 ] ) the mean field band structure of the sl@xmath449 $ ] has a flat band lying at the fermi energy , which is touching a dispersive band at the zone center , i.e. , at the momentum @xmath450 = ( 0,0 ) .
to describe the low energy states near the @xmath450 point , we use the degenerate perturbation theory again .
however , since the two bands touch quadratically , we keep the perturbation expansion up to the quadratic order in momentum . for the sl@xmath449 $ ] ,
the mean field hamiltonian is given by @xmath415 in which @xmath451 we divide @xmath429 into two pieces such that @xmath429 = @xmath452 + @xmath453 in which @xmath452 @xmath41 @xmath454 .
diagonalization of @xmath452 gives the eigenvalues @xmath455=@xmath292 - 1-@xmath421,-1-@xmath421,-1+@xmath421,-1+@xmath421,2-@xmath421,2+@xmath421@xmath295 .
note that there are two pairs of doubly degenerate eigenvalues .
here we focus on one of the degenerate eigenvalues @xmath456=-1+@xmath421 which is lying at the fermi level .
we choose the following two degenerate eigenvectors corresponding to @xmath456 , @xmath457 now we introduce the projection operator @xmath458 ( @xmath459 ) which projects states into ( out of ) the low energy space spanned by @xmath460 and @xmath461 .
that is , @xmath458=@xmath462 + @xmath463 and @xmath459 = @xmath464 - @xmath458 .
we also define @xmath465 @xmath41 @xmath466 and @xmath467 @xmath41 @xmath453 -@xmath465
. then @xmath429 = @xmath452 + @xmath465 + @xmath467 .
the projected hamiltonian is given by@xcite @xmath468\hat{p}_{0 } + \hat{p}_{0}v_{1}\hat{p}_{1}\frac{1}{w_{0}-h_{0}}\hat{p}_{1}v_{1}\hat{p}_{0},\end{aligned}\ ] ] which is valid up to the quadratic order in @xmath79 .
the resulting hamiltonian can be written as , @xmath469 we define the continuum fields as @xmath470 finally , combining the above results we obtain the following low energy effective hamiltonian , @xmath80 in which @xmath471
in this section we show how the continuum fields of the sl@xmath124 $ ] state transform under the microscopic symmetries of the lattice . here
we follow the procedures suggested by hermele et al . in ref . .
to obtain the necessary information we consider a finite system with periodic boundary conditions in both @xmath122 and @xmath2 directions .
in particular , to determine how the wave functions at the momentum @xmath441 ( we call it as the nodal wave functions ) transform under the space group symmetries of the ansatz , we consider the 3@xmath3213 lattice system , that is , the system is periodic under the translation by 3@xmath122 and 3@xmath2 .
since a unit cell ( indexed by a vector @xmath33 ) contains six sites labeled by n , the finite system consists of 54 sites .
the nine points within the brillouin zone of the finite system contain the the nodal points @xmath441 and respect all the point group symmetries of the ansatz . using the eigenvectors @xmath472 ( a = + , - and @xmath53 = 1 , 2 ) in eq .
( [ eq : eigenvector2 ] ) , the nodal wave function is given by @xmath473 then the continuum field is written as @xmath474 now we consider a symmetry operation @xmath160 under which the spinons transform as @xmath475 in particular we consider the following symmetry operations , \(i ) the @xmath122 translation ( @xmath476 ) ; @xmath477 \(ii ) the @xmath2 translation ( @xmath478 ) ; @xmath479 \(iii ) the @xmath335 rotation ( @xmath480 ) ; @xmath481 \(iv ) the y reflection ( @xmath89 ) ; @xmath482 where @xmath33=(x , y ) and ( n , m)@xmath483@xmath484 .
\(v ) time - reversal and inversion ( @xmath485 ) ; @xmath486 in which ( n+3 ) represents the remainder when it is divided by six .
\(vi ) charge conjugation ( @xmath174 ) ; @xmath487 in which @xmath177 = 1 for i = 1 , 2 , 3 and -1 otherwise . for each of the symmetry operation @xmath160 , the matrix representation of the symmetry
@xmath160 is defined as @xmath488 the action of the symmetry operation @xmath160 on the nodal wave function is given by @xmath489=@xmath490 . now
a and b denote the nodal and the two - component dirac indices collectively . finally , the transformation of the nodal wave function reflects the action of @xmath160 on the continuum field such as @xmath491 the transformation rule of the continuum field under all the above symmetry operations is summarized in eq .
( [ eq : symmetryofcontinuum ] ) .
here we present the expressions of the matrices @xmath348 and @xmath492 ( @xmath160=@xmath493 , @xmath494 , @xmath89 , and @xmath88 ) , which are defined in sec . [
sec : psg ] . at first @xmath495 are given by | we investigate various competing paramagnetic ground states of the heisenberg antiferromagnet on the two dimensional star lattice which exhibits geometric frustration .
using slave particle mean field theory combined with a projective symmetry group analysis , we examine a variety of candidate spin liquid states on this lattice , including chiral spin liquids , spin liquids with fermi surfaces of spinons , and nematic spin liquids which break lattice rotational symmetry .
motivated by connection to large - n su(n ) theory as well as numerical exact diagonalization studies , we also examine various valence bond solid ( vbs ) states on this lattice .
based on a study of energetics using gutzwiller projected states , we find that a fully gapped spin liquid state is the lowest energy spin liquid candidate for this model .
we also find , from a study of energetics using gutzwiller projected wave functions and bond operator approaches , that this spin liquid is unstable towards two different vbs states a vbs state which respects all the hamitonian symmetries and a vbs state which exhibits @xmath0 order depending on the ratio of the heisenberg exchange couplings on the two inequivalent bonds of the lattice .
we compute the triplon dispersion in both vbs states within the bond operator approach and discuss possible implications of our work for future experiments on candidate materials . |
uncertainty quantification of complex systems mandates stochastic computations of a multivariate output function @xmath4 that depends on @xmath5 , a high - dimensional random input with a positive integer @xmath1 . for practical applications , encountering hundreds of variables or more
is not uncommon , where a function of interest , defined algorithmically via numerical solution of algebraic , differential , or integral equations , is all too often expensive to evaluate .
therefore , there is a need to develop low - dimensional approximations of @xmath4 by seeking to exploit the hidden structure potentially lurking underneath a function decomposition .
the dimensional decomposition of @xmath6 can be viewed as a finite , hierarchical expansion in terms of its input variables with increasing dimensions , where @xmath7 is a subset with the complementary set @xmath8 and cardinality @xmath9 , and @xmath10 is a @xmath11-variate component function describing a constant or the cooperative influence of @xmath12 , @xmath13 , a subvector of @xmath14 , on @xmath4 when @xmath15 or @xmath16 .
the summation in ( [ 1 ] ) comprises @xmath17 terms , with each term depending on a group of variables indexed by a particular subset of @xmath18 , including the empty set @xmath19 .
this decomposition , first presented by hoeffding @xcite in relation to his seminal work on @xmath20-statistics , has been studied by many other researchers @xcite : sobol @xcite used it for quadrature and analysis of variance ( anova ) @xcite ; efron and stein @xcite applied it to prove their famous lemma on jackknife variances ; owen @xcite presented a continuous space version of the nested anova ; hickernell @xcite developed a reproducing kernel hilbert space version ; and rabitz and alis @xcite made further refinements , referring to it as high - dimensional model representation ( hdmr ) .
more recently , the author s group formulated this decomposition from the perspective of taylor series expansion , solving a number of stochastic - mechanics problems @xcite . in a practical setting ,
the multivariate function @xmath4 , fortunately , has an effective dimension @xcite much lower than @xmath1 , meaning that @xmath4 can be effectively approximated by a sum of lower - dimensional component functions @xmath10 , @xmath21 .
given an integer @xmath22 , the truncated dimensional decomposition @xmath23 then represents a general @xmath0-variate approximation of @xmath24 , which for @xmath25 includes cooperative effects of at most @xmath0 input variables @xmath26 , @xmath27 , on @xmath4 .
however , for ( [ 2 ] ) to be useful , one must ask the fundamental question : what is the approximation error committed by @xmath28 for a given @xmath29 ?
the answer to this question , however , is neither simple nor unique , because there are multiple ways to construct the component functions @xmath10 , @xmath30 , spawning approximations of distinct qualities .
indeed , there exist two important variants of the decomposition : ( 1 ) referential dimensional decomposition ( rdd ) and ( 2 ) anova dimensional decomposition ( add ) , both representing sums of lower - dimensional component functions of @xmath24 . while add has desirable orthogonal properties , the anova component functions are difficult to compute due to the high - dimensional integrals involved .
in contrast , the rdd lacks orthogonal features with respect to the probability measure of @xmath14 , but its component functions are much easier to obtain . for rdd , an additional question arises regarding the reference point , which , if improperly selected , can mar the approximation .
existing error analysis , limited to the univariate truncation of @xmath24 , reveals that the expected error from the rdd approximation is at least four times larger than the error from the add approximation @xcite .
although useful to some extent , such result alone is not adequate when evaluating multivariate functions requiring higher - variate interactions of input @xcite .
no error estimates exist yet in the current literature even for a bivariate approximation .
therefore , a more general error analysis pertaining to a general @xmath0-variate approximation of a multivariate function should provide much - needed insights into the mathematical underpinnings of dimensional decomposition .
the purpose of this paper is twofold .
firstly , a brief exposition of add and rdd is given in section 2 , including clarifications of parallel developments and synonyms used by various researchers .
the error analysis pertaining to the add approximation is described in section 3 .
secondly , a direct form of the rdd approximation , previously developed by the author s group , is tapped for providing a vital link to subsequent error analysis .
section 4 introduces new formulae for the lower and upper bounds of the expected errors from the bivariate and general @xmath0-variate rdd approximations .
these error bounds , so far available only for the univariate approximation , are used to clarify why add approximations are exceedingly more precise than rdd approximations .
there are seven new results stated or proved in this paper : proposition [ p4 ] , theorems [ t6 ] , [ t7 ] , and corollaries [ c2 ] , [ c3 ] , [ c4a ] , [ c4 ] . proofs of other results can be obtained from the references cited , including a longer version of this paper ( http://www.engineering.uiowa.edu/~rahman/moc_longpaper.pdf ) .
conclusions are drawn in section 5 .
let @xmath31 , @xmath32 , @xmath33 , @xmath34 , and @xmath35 represent the sets of positive integer ( natural ) , non - negative integer , integer , real , and non - negative real numbers , respectively . for @xmath36 , denote by @xmath37 the @xmath38-dimensional euclidean space and by @xmath39 the @xmath38-dimensional multi - index space .
these standard notations will be used throughout the paper .
let @xmath40 be a complete probability space , where @xmath41 is a sample space , @xmath42 is a @xmath43-field on @xmath41 , and @xmath44 $ ] is a probability measure . with @xmath45 representing the borel @xmath43-field on @xmath46 , consider an @xmath46-valued independent random vector @xmath47 , which describes the statistical uncertainties in all system and input parameters of a given stochastic problem .
the probability law of @xmath14 is completely defined by its joint probability density function @xmath48 .
assuming independent coordinates of @xmath14 , its joint probability density @xmath49 can be expressed by a product of marginal probability density functions @xmath50 of @xmath51 , @xmath52 , defined on the probability triple @xmath53 with a bounded or an unbounded support on @xmath34 . consider a non - negative , multiplicative , otherwise general , weight function @xmath54 , satisfying @xmath55 with the marginal weight functions @xmath56 , @xmath52 . for @xmath7 ,
let @xmath57 define the joint weight function associated with @xmath58 . without loss of generality , assume that the weight functions have been normalized to integrate to @xmath59 let @xmath60 ) , a real - valued , measurable transformation on @xmath61 , define a stochastic response of interest and @xmath62 represent a hilbert space of square - integrable functions @xmath4 with respect to the induced generic measure @xmath63 supported on @xmath46 . the representation in ( [ 1 ] ) is called dimensional decomposition if the component functions @xmath10 , @xmath7 , are uniquely determined from the requirements @xmath64 indeed , integrating ( [ 1 ] ) with respect to the measure @xmath65 , that is , over all variables except @xmath66 , and using ( [ 3 ] ) yields a recursive form @xmath67 of the decomposition with @xmath68 denoting an @xmath1-dimensional vector whose @xmath69th component is @xmath51 if @xmath70 and @xmath71 if @xmath72 when @xmath73 , the sum in the last line of ( [ 3b ] ) vanishes , resulting in the expression of the constant function @xmath74 in the second line . when @xmath75 , the integration in the last line of ( [ 3b ] ) is on the empty set , reproducing ( [ 1 ] ) and hence finding the last function @xmath76 .
indeed , all component functions of @xmath4 can be obtained by interpreting literally the last line of ( [ 3b ] ) . on inversion , ( [ 3b ] )
results in @xmath77 providing an explicit form of the same decomposition .
it is important to emphasize that the measure involved in expressing the dimensional decomposition in ( [ 3b ] ) or ( [ 3c ] ) may or may not represent the probability measure of @xmath14 . indeed
, different measures will create distinct yet exact representations of @xmath4 , all exhibiting the same structure of ( [ 1 ] ) .
there exist two important variants of dimensional decomposition , described as follows .
the add is generated by selecting the probability measure of @xmath14 as the generic measure , that is , @xmath63=@xmath78 in ( [ 3b ] ) or ( [ 3c ] ) , yielding the recursive form @xmath79 [ 4 ] that is commonly found in the anova literature @xcite , although , for the uniform probability measure @xmath80 .
the explicit version takes the form @xmath81 where @xmath82 is the marginal probability density function of @xmath83
. equations ( [ 4 ] ) and ( [ 5 ] ) can also be derived from other perspectives , including commuting projections on a linear space of real - valued functions invoked by kuo _
@xcite for the uniform probability measure . if @xmath84 is the expectation operator with respect to the measure @xmath78 , then two important properties of the add component functions , inherited from ( [ 3 ] ) , are as follows .
+ the add component functions @xmath85 , @xmath86 , have zero means , i.e. , @xmath87=\int_{\mathbb{r}^{|u|}}y_{u , a}(\mathbf{x}_{u})f_{u}(\mathbf{x}_{u})d\mathbf{x}_{u}=0.\label{6}\ ] ] [ p1 ] two distinct add component functions @xmath85 and @xmath88 , where @xmath86 , @xmath89 , and @xmath90 , are orthogonal , i.e. , they satisfy the property , @xmath91=\int_{\mathbb{r}^{|u\cup v|}}y_{u , a}(\mathbf{x}_{u})y_{v , a}(\mathbf{x}_{v})f_{u\cup
v}(\mathbf{x}_{u\cup v})d\mathbf{x}_{u\cup v}=0.\label{7}\ ] ] [ p2 ] traditionally , ( [ 4 ] ) or ( [ 5 ] ) with @xmath92 , @xmath93 , following independent , standard uniform distributions , has been identified as the anova decomposition @xcite ; however , the author s recent works @xcite reveal no fundamental requirement for a specific probability measure of @xmath14 , provided that the resultant integrals in ( [ 4 ] ) or ( [ 5 ] ) exist and are finite . in this work , the add should be interpreted with respect to an arbitrary but product type probability measure for which it is always endowed with desirable orthogonal properties .
however , the add component functions are difficult to ascertain , because they require calculation of high - dimensional integrals .
consider a reference point @xmath94 and the associated dirac measure @xmath95 .
the rdd is created when @xmath95 is chosen as the generic measure in ( [ 3b ] ) , leading to the recursive form @xmath96 [ 8 ] presented as cut - hdmr @xcite , anchored decomposition @xcite , and anchored - anova decomposition @xcite , with the latter two referring to the reference point as the anchor .
xu and rahman introduced ( [ 8 ] ) with the aid of taylor series expansion , calling it dimension - reduction @xcite and decomposition @xcite methods for calculating statistical moments and reliability of mechanical system responses , respectively . again , these various synonyms of the same decomposition exist due to diverse perspectives employed by researchers in disparate fields .
analogous to add , the rdd can also be described explicitly , for instance @xcite , @xmath97 where @xmath98 denotes an @xmath1-dimensional vector whose @xmath69th component is @xmath51 if @xmath99 and @xmath100 if @xmath101 the second argument `` @xmath102 '' introduced in @xmath103 and @xmath104 is a reminder that the rdd component functions depend on the reference point , although @xmath4 does not , as ( [ 9 ] ) is exact .
an important property of the rdd component functions , also inherited from ( [ 3 ] ) , is as follows @xcite .
the rdd component functions @xmath103 , @xmath86 , vanish when any of its own variables @xmath51 with @xmath70 takes on the value of @xmath100 , i.e. , @xmath105 [ p3 ] clearly , the rdd component functions lack orthogonal features with respect to the probability measure of @xmath14 , but are relatively easy to obtain as they only involve function evaluations at a chosen reference point . however , the rdd component functions can be orthogonal with respect to other inner products @xcite . when a dimensional decomposition , whether add or rdd or others , of a multivariate function is truncated by retaining only lower - dimensional terms , the result is an approximation .
however , due to the special structure of the decomposition , the approximation is endowed with an error - minimizing property , described in theorem [ t1 ] .
[ @xcite ] for a multivariate function @xmath106 and @xmath29 , if @xmath107 represents an @xmath0-variate approximation , obtained by truncating at @xmath30 the dimensional decomposition of @xmath24 with respect to the generic measure @xmath63 , then its component functions @xmath10 , @xmath30 , are uniquely determined from @xmath108^{2}w(\mathbf{x})d\mathbf{x},\\ \mathrm{subject\ : to } & \int_{\mathbb{r}}y_{u}(\mathbf{x}_{u})w_{i}(x_{i})dx_{i}=0\;\mathrm{for}\ ; i\in u.\end{split}\label{10c}\ ] ] [ t1 ] rabitz and alis @xcite proved this theorem in pages 202 - 207 of their paper .
the @xmath109 error in ( [ 10c ] ) committed by a truncated dimensional decomposition is minimized , but only for a specific measure . given a measure ,
no other choices of the component functions of @xmath110 will produce approximations that are better than the one derived from ( [ 10b ] ) .
however , different measures will create different truncated decompositions , resulting in distinct approximation errors .
therefore , selecting a measure is vitally important for determining the approximation quality of a dimensional decomposition .
the @xmath0-variate add approximation @xmath111 , say , of @xmath24 , where @xmath29 , is obtained by truncating the right side of ( [ 4a ] ) at @xmath30 , yielding @xmath112 applying the expectation operator on @xmath24 and @xmath111 from ( [ 4a ] ) and ( [ 11 ] ) , respectively , and noting proposition [ p1 ] , the mean @xmath113=y_{\emptyset , a}$ ] of the @xmath0-variate add approximation matches the exact mean @xmath114:=\int_{\mathbb{r}^{n}}y(\mathbf{x})f_{\mathbf{x}}(\mathbf{x})d\mathbf{x}=y_{\emptyset , a}$ ] , regardless of @xmath0 . applying the expectation operator again , this time on @xmath115 , and recognizing proposition [ p2 ] results in splitting the variance @xmath116=\sum_{{\textstyle { \emptyset\ne u\subseteq\{1,\cdots , n\}\atop 1\le|u|\le s}}}\sigma_{u}^{2}=\sum_{s=1}^{s}\:\sum_{{\textstyle { \emptyset\ne u\subseteq\{1,\cdots , n\}\atop |u|=s}}}\sigma_{u}^{2}\label{12}\ ] ] of the @xmath0-variate add approximation , where @xmath117 $ ] represents the variance of the _ _
zero-__mean add component function @xmath85 , @xmath118 .
clearly , the approximate variance in ( [ 12 ] ) approaches the exact variance @xmath119=\sum_{\emptyset\ne u\subseteq\{1,\cdots , n\}}\sigma_{u}^{2}=\sum_{s=1}^{n}\:\sum_{{\textstyle { \emptyset\ne u\subseteq\{1,\cdots , n\}\atop |u|=s}}}\sigma_{u}^{2},\label{13}\ ] ] the sum of all variance terms , when @xmath120 .
a normalized version @xmath121 is often called the global sensitivity index of @xmath4 for @xmath122 @xcite .
define a mean - squared error @xmath123:=\int_{\mathbb{r}^{n}}\left[y(\mathbf{x})-\hat{y}_{s , a}(\mathbf{x})\right]^{2}f_{\mathbf{x}}(\mathbf{x})d\mathbf{x}\label{14}\ ] ] committed by the @xmath0-variate add approximation @xmath111 of @xmath24 . replacing @xmath4 and @xmath124 in ( [ 14 ] ) with the right sides of ( [ 4a ] ) and ( [ 11 ] ) , respectively , and then recognizing propositions [ p1 ] and [ p2 ] yields @xmath125 which completely eliminates the variance terms of @xmath126 that are associated with @xmath0- and all lower - variate contributions , an attractive property of add . by setting @xmath127 ,
the error can be expressed for any truncation of add . among all possible measures
, the probability measure endows the add approximation with an error - minimizing property , explained as follows . for a given @xmath0 , the @xmath0-variate add approximation is optimal in the mean - square sense .
[ p4 ] consider a generic @xmath0-variate approximation @xmath28 of @xmath24 other than the add approximation @xmath111 . since @xmath128 contains only higher than @xmath0-variate terms and @xmath129 contains at most @xmath0-variate terms ,
@xmath130 and @xmath131 are orthogonal , satisfying @xmath132=0 $ ] .
consequently , the second - moment error from an @xmath0-variate approximation @xmath133\\ & = \mathbb{e}\left[\left(y(\mathbf{x})-\hat{y}_{s , a}(\mathbf{x})\right)^{2}\right]+\mathbb{e}\left[\left(\hat{y}_{s , a}(\mathbf{x})-\hat{y}_{s}(\mathbf{x})\right)^{2}\right]\\ & = e_{s , a}+\mathbb{e}\left[\left(\hat{y}_{s , a}(\mathbf{x})-\hat{y}_{s}(\mathbf{x})\right)^{2}\right]\ge e_{s , a},\label{16}\end{aligned}\ ] ] proving the mean - square optimality of the @xmath0-variate add approximation .
therefore , given a truncation , an rdd approximation , regardless of how the reference point is selected , can not be better than an add approximation for calculating variance .
further details of rdd approximation errors are described in the next section .
the @xmath0-variate rdd approximation @xmath134 , say , of @xmath24 , where @xmath29 , is obtained by truncating the right side of ( [ 8a ] ) at @xmath30 , yielding @xmath135 which depends on the reference point , needing the second argument `` @xmath102 '' in @xmath136 . for error analysis
, however , a suitable direct form of ( [ 17 ] ) is desirable .
theorem [ t2 ] supplies such a form , which was originally obtained by xu and rahman @xcite using the taylor series expansion .
the same form was reported later by kuo _
et al_. @xcite .
let @xmath137 , @xmath138 , be a @xmath38-dimensional multi - index with each component representing a non - negative integer .
the multi - index , used in theorem [ t2 ] , obeys the following standard notations : ( 1 ) @xmath139 ; ( 2 ) @xmath140 ; ( 3 ) @xmath141 ; ( 4 ) @xmath142 , @xmath143 .
[ multivariate function theorem @xcite ] for a differentiable multivariate function @xmath106 and @xmath29 , if @xmath144 represents an @xmath0-variate rdd approximation of @xmath24 , then @xmath134 consists of all terms of the taylor series expansion of @xmath24 at @xmath102 that have less than or equal to @xmath0 variables , i.e. , @xmath145 where @xmath146 [ t2 ] xu and rahman @xcite proved this theorem in pages 1996 - 2000 of their paper when @xmath147 without loss of generality .
the stochastic method associated with the rdd approximation was simply called `` decomposition method '' @xcite .
theorem [ t2 ] implies that the rdd approximation @xmath134 in ( [ 18 ] ) , when compared with the taylor series expansion of @xmath24 , yields residual error that includes only terms of dimensions @xmath148 and higher .
all higher - order @xmath0- and lower - variate terms of @xmath24 are included in ( [ 18 ] ) , which should therefore generally provide a higher - order approximation of a multivariate function than the equation derived from an @xmath0-order taylor expansion .
equations ( [ 17 ] ) and ( [ 18 ] ) both follow the same structure of ( [ 2 ] ) . however , due to the distinct perspectives involved , it is not obvious if these equations represent the same function @xmath136 . a lemma and a theorem recently proved by the author
demonstrate that , indeed , they do @xcite . when @xmath149 , @xmath150 , and @xmath151 , ( [ 18 ] ) degenerates to the zero - variate rdd approximation @xmath152 the univariate rdd approximation @xmath153 and the bivariate rdd approximation @xmath154 respectively .
[ c0 ] since the right side of ( [ 26 ] ) comprises only univariate functions , the interpolation or integration of @xmath155 is essentially univariate .
similarly , the right side of ( [ 27 ] ) , which contains at most bivariate functions , requires at most bivariate interpolation or integration of @xmath156 .
therefore , appellation of the terms `` univariate approximation '' and `` bivariate approximation '' for @xmath155 in ( [ 26 ] ) and @xmath156 in ( [ 27 ] ) , respectively , is more appropriate than referring to them as first - order and second - order approximations .
following similar consideration , define another mean - squared error @xmath157 : = \int_{\mathbb{r}^{n}}\left[y(\mathbf{x})-\hat{y}_{s , r}(\mathbf{x};\mathbf{c})\right]^{2}f_{\mathbf{x}}(\mathbf{x})d\mathbf{x } \end{split}\label{28}\ ] ] associated with the @xmath0-variate rdd approximation @xmath134 of @xmath24 , which depends on the reference point @xmath102 .
wang @xcite suggested choosing a random reference point uniformly distributed over @xmath158^{n}$ ] and then calculating the error on average .
but , @xmath14 defined here may follow an arbitrary probability law with density @xmath159 ; therefore , selecting the reference point characterized by the probability density @xmath160 is more appropriate , which leads to @xmath161 & : = \int_{\mathbb{r}^{n}}e_{s , r}(\mathbf{c})f_{\mathbf{x}}(\mathbf{c})d\mathbf{c } \\ & = \int_{\mathbb{r}^{2n}}\left[y(\mathbf{x})-\hat{y}_{s , r}(\mathbf{x};\mathbf{c})\right]^{2}f_{\mathbf{x}}(\mathbf{x})f_{\mathbf{x}}(\mathbf{c})d\mathbf{x}d\mathbf{c } , \end{split}\label{29}\ ] ] as the expected value of the rdd error . simplifying ( [ 29 ] ) in terms of the variance components of @xmath4 , as done for the add error in ( [ 15 ] ) , for arbitrary @xmath0 and @xmath1 may appear formidable . here , the _ zero_-variate ( @xmath149 ) , univariate ( @xmath162 ) , and bivariate ( @xmath163 ) approximation errors for arbitrary @xmath1 will be derived first , followed by error analysis for a general @xmath0-variate approximation . in all cases ,
the derivations require using ( 1 ) the relationships , @xmath164 [ 30 ] that exist between add and rdd component functions and approximations and ( 2 ) sobol s formula @xcite , @xmath165 for select choices of @xmath166 described in the following subsection .
equations ( [ 30a ] ) and ( [ 30b ] ) follow from propositions [ p1 ] , [ p2 ] , and [ p3 ] and definitions of respective component functions in ( [ 4 ] ) and ( [ 8 ] ) , eventually leading to ( [ 30c ] ) .
the term @xmath167 in sobol s formula represents a sum of variance terms contributed by the add component functions that belong to @xmath86 .
theorems [ t4 ] , [ t5 ] , and [ t6 ] show how the expected errors from the _ zero_-variate , univariate , and bivariate rdd approximations , respectively , depend on the variance components of @xmath4 .
let @xmath94 be a random vector with the joint probability density function of the form @xmath168 , where @xmath169 is the marginal probability density function of its @xmath170th coordinate .
then the expected error committed by the zero - variate rdd approximation for @xmath171 is @xmath172=2\sigma^{2},\label{31b}\ ] ] where @xmath173=\mathbb{e}\left[y^2(\mathbf{x})\right]-y_{\emptyset , a}^2 $ ] is the variance of @xmath4 .
[ t4 ] setting @xmath149 in ( [ 29 ] ) and using the expression of @xmath174 from ( [ 25b ] ) , the expected error from the _ zero_-variate rdd approximation becomes @xmath175 & = \int_{\mathbb{r}^{2n}}\left[y(\mathbf{x})-y(\mathbf{c})\right]^{2}f_{\mathbf{x}}(\mathbf{x})f_{\mathbf{x}}(\mathbf{c})d\mathbf{x}d\mathbf{c}\\ & = \int_{\mathbb{r}^{n}}y^{2}(\mathbf{x})f_{\mathbf{x}}(\mathbf{x})d\mathbf{x}+\int_{\mathbb{r}^{n}}y^{2}(\mathbf{c})f_{\mathbf{x}}(\mathbf{c})d\mathbf{c}\\ & \;\;\;\ ; -2\int_{\mathbb{r}^{n}}y(\mathbf{x})f_{\mathbf{x}}(\mathbf{x})d\mathbf{x}\int_{\mathbb{r}^{n}}y(\mathbf{c})f_{\mathbf{x}}(\mathbf{c})d\mathbf{c}\\ & = \sigma^{2}+y_{\emptyset , a}^{2}+\sigma^{2}+y_{\emptyset , a}^{2}-2y_{\emptyset , a}^{2}\\ & = 2\sigma^{2 } , \end{split}\label{31c}\ ] ] where the third equality exploits the add - rdd relationship in ( [ 30a ] ) .
hence , the theorem is proven .
let @xmath94 be a random vector with the joint probability density function of the form @xmath168 , where @xmath169 is the marginal probability density function of its @xmath170th coordinate .
then the expected error committed by the univariate rdd approximation for @xmath176 is @xmath177={\displaystyle \sum_{s=2}^{n}\left(s^{2}-s+2\right)\sum_{{\textstyle { \emptyset\neq u\subseteq\{1,\cdots , n\}\atop |u|=s}}}}\sigma_{u}^{2},\label{32}\ ] ] where @xmath178 $ ] is the variance of the _
zero_-mean add component function @xmath85 , @xmath86 .
[ t5 ] setting @xmath162 in ( [ 29 ] ) , the expected error from the univariate rdd approximation on expansion is a sum @xmath177=i_{1,1}+i_{1,2}+i_{1,3}\label{33}\ ] ] of three integrals @xmath179 on @xmath180 , where their first indices represent the univariate approximation .
the first integral @xmath181=y_{\emptyset , a}^{2}+\sigma^{2}=y_{\emptyset , a}^{2}+\sum_{s=1}^{n}\:\sum_{{\textstyle { \emptyset\ne u\subseteq\{1,\cdots , n\}\atop |u|=s}}}\sigma_{u}^{2},\label{35}\ ] ] expressed in terms of the variance components , is independent of @xmath0 . however
, the second integral depends on @xmath0 , yielding @xmath182 \\ & = -2\left(y_{\emptyset , a}^{2}+\hat{\sigma}_{1,a}^{2}\right ) \\ & = -2y_{\emptyset , a}^{2}-2{\displaystyle \sum_{{\textstyle { \emptyset\ne u\subseteq\{1,\cdots , n\}\atop |u|=1}}}}\sigma_{u}^{2 } , \end{split}\label{36}\ ] ] where the first , second , and fifth lines are obtained ( 1 ) employing the add - rdd relationships in ( [ 30c ] ) for @xmath162 , ( 2 ) recognizing @xmath183 and @xmath184 to be orthogonal , satisfying @xmath185=0 $ ] , and ( 3 ) applying ( [ 12 ] ) for @xmath162 , respectively . using the expression of @xmath186 from ( [ 26 ] ) and noting independent coordinates of @xmath14 and @xmath102 , the expanded third integral becomes @xmath187^{2}f_{\mathbf{x}}(\mathbf{x})f_{\mathbf{x}}(\mathbf{c})d\mathbf{x}d\mathbf{c}\\ & = \int_{\mathbb{r}^{2n}}\biggl[{\displaystyle \sum_{i=1}^{n}}y^{2}(x_{i},\mathbf{c}_{-\{i\}})+2{\displaystyle \sum_{i=1}^{n-1}\sum_{j = i+1}^{n}}y(x_{i},\mathbf{c}_{-\{i\}})y(x_{j},\mathbf{c}_{-\{j\}})\\ & \;\;\ ; + ( n-1)^{2}y^{2}(\mathbf{c})-2(n-1){\displaystyle \sum_{i=1}^{n}}y(x_{i},\mathbf{c}_{-\{i\}})y(\mathbf{c})\biggr]f_{\mathbf{x}}(\mathbf{x})f_{\mathbf{x}}(\mathbf{c})d\mathbf{x}d\mathbf{c}. \end{split}\label{37}\ ] ] further evaluation of this integral requires exploiting sobol s formula in ( [ 31 ] ) for @xmath188 and @xmath189 , where @xmath190 , @xmath191 , yielding @xmath192 where @xmath193 is the generic @xmath194-variate coefficient for the univariate approximation , obtained by counting the number of @xmath195 or @xmath196 for @xmath197 @xmath198 _ e.g. _ , @xmath199 for @xmath200 , @xmath201 for @xmath202 , and so on @xmath198 due to symmetry .
adding all terms in ( [ 35 ] ) , ( [ 36 ] ) , and ( [ 37b ] ) , with the recognition that @xmath203 , yields ( [ 32 ] ) , completing the proof .
let @xmath94 be a random vector with the joint probability density function of the form @xmath168 , where @xmath169 is the marginal probability density function of its @xmath170th coordinate .
then the expected error committed by the bivariate rdd approximation for @xmath204 is @xmath205={\displaystyle \sum_{s=3}^{n}{\displaystyle \frac{1}{4}}\left(s^{4}-2s^{3}-s^{2}+2s+8\right)\sum_{{\textstyle { \emptyset\neq u\subseteq\{1,\cdots , n\}\atop |u|=s}}}}\sigma_{u}^{2},\label{39}\ ] ] where @xmath178 $ ] is the variance of the _
zero_-mean add component function @xmath85 , @xmath86 .
[ t6 ] setting @xmath163 in ( [ 29 ] ) , the expected error from the bivariate rdd approximation on expansion is another sum @xmath205=i_{2,1}+i_{2,2}+i_{2,3}\label{40}\ ] ] of three @xmath206-dimensional integrals @xmath207 where their first indices represent the bivariate approximation . since the first integral
does not depend on @xmath0 , @xmath208 is the same as @xmath209 .
following a similar reasoning employed for the univariate approximation , the second integral @xmath210 = -2\left(y_{\emptyset , a}^{2}+\hat{\sigma}_{2,a}^{2}\right ) \\ % & = -2\left(y_{\emptyset , a}^{2}+\hat{\sigma}_{2,a}^{2}\right)\\ & = -2y_{\emptyset , a}^{2}-2{\displaystyle \sum_{{\textstyle { \emptyset\ne u\subseteq\{1,\cdots , n\}\atop |u|=1}}}}\sigma_{u}^{2}-2{\displaystyle \sum_{{\textstyle { \emptyset\ne u\subseteq\{1,\cdots , n\}\atop |u|=2}}}}\sigma_{u}^{2 } \end{split}\label{43}\ ] ] contains variance terms associated with at most two variables . using the expression of @xmath211 from ( [ 27 ] ) ,
the expanded third integral becomes @xmath212^{2}f_{\mathbf{x}}(\mathbf{x})f_{\mathbf{x}}(\mathbf{c})d\mathbf{x}d\mathbf{c}\\ & = \int_{\mathbb{r}^{2n}}\biggl\{\biggl[{\displaystyle \sum_{i=1}^{n-1}\sum_{j = i+1}^{n}}y(x_{i},x_{j},\mathbf{c}_{-\{i , j\}})\biggr]^{2}{\displaystyle + ( n-2)^{2}{\displaystyle \biggl[\sum_{i=1}^{n}y(x_{i},\mathbf{c}_{-\{i\}})\biggr]^{2}}}\\ & \;\;\ ; + { \displaystyle \frac{1}{4}(n-1)^{2}(n-2)^{2}}y^{2}(\mathbf{c})\\ & \;\;\ ; -2(n-2)\biggl[{\displaystyle \sum_{i=1}^{n-1}\sum_{j = i+1}^{n}}y(x_{i},x_{j},\mathbf{c}_{-\{i , j\}})\biggr]\biggl[{\displaystyle \sum_{i=1}^{n}}y(x_{i},\mathbf{c}_{-\{i\}})\biggr]\\ & \;\;\ ; -(n-1)(n-2)^{2}\biggl[{\displaystyle \sum_{i=1}^{n}}y(x_{i},\mathbf{c}_{-\{i\}})\biggr]y(\mathbf{c})\\ & \;\;\ ; + ( n-1)(n-2)\biggl[{\displaystyle \sum_{i=1}^{n-1}\sum_{j = i+1}^{n}}y(x_{i},x_{j},\mathbf{c}_{-\{i , j\}})\biggr]y(\mathbf{c})\biggr\ } f_{\mathbf{x}}(\mathbf{x})f_{\mathbf{x}}(\mathbf{c})d\mathbf{x}d\mathbf{c}. \end{split}\label{43b}\ ] ] employing sobol s formula , this time for @xmath188 , @xmath213 , @xmath189 , @xmath214 , @xmath215 , @xmath216 , and @xmath217 , where @xmath218 , @xmath219 , results in @xmath220 producing the generic @xmath194-variate coefficient @xmath221 for the bivariate approximation . adding all terms in ( [ 42 ] ) , ( [ 43 ] ) , and ( [ 44 ] ) , with the understanding that @xmath222 , yields ( [ 39 ] ) , proving the theorem .
the expected error @xmath223 $ ] from the zero - variate rdd approximation , expressed in terms of the error @xmath224 from the zero - variate add approximation , is @xmath225=2e_{0,a},\;1\le n<\infty.\label{45b}\ ] ] [ c1 ] the lower and upper bounds of the expected errors @xmath226 $ ] and @xmath227 $ ] from the univariate and bivariate rdd approximations , respectively , expressed in terms of the errors @xmath228 and @xmath229 from the univariate and bivariate add approximations , are @xmath230 \le \left(n^{2}-n+2\right)e_{1,a},\;2\le n<\infty,\ ] ] and @xmath231 \le { \displaystyle \frac{1}{4}}\left(n^{4}-2n^{3}-n^{2}+2n+8\right)e_{2,a},\;3\le n<\infty,\ ] ] respectively .
[ c2 ] when @xmath14 comprises independent and identically distributed uniform random variables over @xmath158 $ ] , the results of the _ zero_-variate and univariate rdd approximations presented in theorems [ t4 ] and [ t5 ] coincide with those derived by wang @xcite . however , the results of the bivariate rdd approximation @xmath198 that is , theorem [ t6 ] @xmath198 are new .
theorems [ t5 ] and [ t6 ] demonstrate that on average the error from the univariate rdd approximation eliminates the variance terms associated with the univariate contribution .
for the bivariate rdd approximation , the variance portions resulting from the univariate and bivariate terms have been removed as well .
the univariate and bivariate add approximations also satisfy this important property .
however , the coefficients of higher - variate terms in the rdd errors are larger than unity , implying greater errors from rdd approximations than from add approximations . from corollary [ c1 ] , the _
zero_-variate rdd approximation on average commits twice the amount of error as does the _ zero_-variate add approximation . since a _ zero_-variate approximation , whether derived from add or rdd , does not capture the random fluctuations of a stochastic response , the error analysis associated with a _
zero_-variate approximation is useless . nonetheless , the _ zero_-variate results are reported here for completeness .
corollary [ c2 ] shows that the expected error from the univariate rdd approximation is at least four times larger than the error from the univariate add approximation .
in contrast , the expected error from the bivariate rdd approximation can be eight times larger or more than the error from the bivariate add approximation . given a truncation , an add approximation is superior to an rdd approximation .
in addition , rdd approximations may perpetrate very large errors at upper bounds when there exist a large number of variables and appropriate conditions .
for instance , consider a contrived example involving a function of @xmath232 variables with a finite variance @xmath233 and the following distribution of the variance terms : @xmath234 , @xmath235 , and @xmath236 .
then , the errors from the univariate and bivariate add approximations are both equal to @xmath237 , which is negligibly small .
in contrast , the error from the univariate rdd approximation reaches @xmath238 , an unacceptably large magnitude already .
furthermore , the error from the bivariate rdd approximation jumps to an enormously large value of @xmath239 .
more importantly , the results reveal a theoretical possibility for a higher - variate rdd approximation to commit a larger error than a lower - variate rdd approximation @xmath198 an impossible scenario for the add approximation .
however , it is unlikely for this odd behavior to be exhibited for realistic functions , where the variances of higher - variate component functions attenuate rapidly or vanish altogether .
nonetheless , a caution is warranted when employing rdd approximations for stochastic analysis of high - dimensional systems .
[ r1 ] the error analysis presented so far is limited to at most the bivariate approximation . in this subsection ,
the approximation error from a general @xmath0-variate truncation is derived as follows .
let @xmath94 be a random vector with the joint probability density function of the form @xmath168 , where @xmath169 is the marginal probability density function of its @xmath170th coordinate .
then the expected error committed by the @xmath0-variate rdd approximation for @xmath240 @xmath241 is @xmath242={\displaystyle \sum_{s = s+1}^{n}\left[{\displaystyle 1+\sum_{k=0}^{s}}{\displaystyle \binom{s - s+k-1}{k}^{2}}\binom{s}{s - k}\right]\sum_{{\textstyle { \emptyset\neq u\subseteq\{1,\cdots , n\}\atop |u|=s}}}}\sigma_{u}^{2},\label{47}\ ] ] where @xmath178 $ ] is the variance of the _
zero_-mean add component function @xmath85 , @xmath86 .
[ t7 ] expanding the square in ( [ 29 ] ) , the expected error from the @xmath0-variate rdd approximation is @xmath242=i_{s,1}+i_{s,2}+i_{s,3},\label{48}\ ] ] where @xmath243 are three generic @xmath206-dimensional integrals .
the first integral @xmath244 is the same as before .
the second integral @xmath245=-2\left(y_{\emptyset , a}^{2}+\hat{\sigma}_{s , a}^{2}\right)\\ & = -2y_{\emptyset , a}^{2}-2{\displaystyle \sum_{s=1}^{s}\:\sum_{{\textstyle { \emptyset\ne u\subseteq\{1,\cdots , n\}\atop |u|=s}}}}\sigma_{u}^{2 } \end{split}\label{51}\ ] ] comprises variance terms associated with at most @xmath0 variables .
the third integral , using @xmath246 from ( [ 18 ] ) , takes the form @xmath247^{2}f_{\mathbf{x}}(\mathbf{x})f_{\mathbf{x}}(\mathbf{c})d\mathbf{x}d\mathbf{c},\\ & = b_{s}(0)y_{\emptyset , a}^{2}+{\displaystyle \sum_{s=1}^{n}b_{s}(s)\sum_{{\textstyle { \emptyset\ne u\subseteq\{1,\cdots , n\}\atop |u|=s}}}}\sigma_{u}^{2 } \end{split}\label{52}\ ] ] with the generic @xmath194-variate coefficient @xmath248 , @xmath249 for the @xmath0-variate approximation yet to be determined .
of @xmath250 such coefficients , the last one , @xmath251 is easier to determine .
it is obtained from the expansion coefficients of the square in ( [ 52 ] ) that are associated with all variance terms of @xmath126 . to determine other coefficients , @xmath248 , @xmath252 , the procedure used before for the univariate or bivariate approximation is unwieldy .
an alternative scheme proposed here stems from the realization that the expressions of those coefficients consist of terms from two sources : ( 1 ) terms that depend solely on @xmath1 , which can be described by a function @xmath253 , say , of @xmath1 ; and ( 2 ) terms that depend on both @xmath1 and @xmath194 , which can be described by another function @xmath254 , say , of @xmath1 and @xmath194 . following this rationale ,
let a generic coefficient be expressed by @xmath255 for any @xmath256 and @xmath257 switching the variables @xmath1 and @xmath194 , ( [ 54 ] ) produces @xmath258 at @xmath259 , ( [ 54 ] ) and ( [ 55 ] ) result in @xmath260 .
either of these two equations at @xmath259 with @xmath260 in mind yields @xmath261 or @xmath262 , where the function @xmath263 is already described in ( [ 53 ] ) .
therefore , the generic @xmath194-variate coefficient @xmath264 where the binomial coefficients should be interpreted more generally than their classical combinatorial definition , for instance , @xmath265 valid for any real number @xmath266 and any integer @xmath267 .
adding all terms in ( [ 50 ] ) , ( [ 51 ] ) , and ( [ 52 ] ) , with the cognizance that @xmath268 for all @xmath269 , yields ( [ 47 ] ) , proving the theorem .
the lower and upper bounds of the expected error @xmath270 $ ] from the @xmath0-variate rdd approximation , expressed in terms of the error @xmath271 from the @xmath0-variate add approximations , are @xmath272\le\left[{\displaystyle 1+\sum_{k=0}^{s}}\binom{n - s+k-1}{k}^{2}\binom{n}{s - k}\right]e_{s , a},\label{57}\ ] ] @xmath2 , where the coefficients of the lower and upper bounds are obtained from @xmath273 and @xmath274 respectively .
[ c3 ] both theorem [ t7 ] and corollary [ c3 ] are new and provide a general result pertaining to rdd error analysis for an arbitrary truncation .
the specific results of the _ zero_-variate or univariate or bivariate rdd approximation , derived in the preceding subsection , can be recovered by setting @xmath149 or @xmath150 or @xmath151 in ( [ 47 ] ) through ( [ 59 ] ) . from corollary [ c3 ] ,
the expected error from the @xmath0-variate rdd approximation of a multivariate function is at least @xmath3 times larger than the error from the @xmath0-variate add approximation .
in other words , the ratio of rdd to add errors doubles for each increment of the truncation . consequently , add approximations are exceedingly more precise than rdd approximations at higher - variate truncations .
although the relative disadvantage of using rdd over add worsens drastically with the truncation @xmath0 , one hopes that the approximation error is also decreasing with increasing @xmath0 .
for instance , given a rate at which @xmath275 decreases with @xmath11 , what can be inferred on how fast @xmath271 and @xmath270 $ ] decay with respect to @xmath0 ?
corollary [ c4a ] and subsequent discussions provide some insights . if the variance of the _ zero_-mean add component function @xmath85 diminishes according to @xmath276 , where @xmath86 , and @xmath277 and @xmath278 are two real - valued constants , then @xmath279 and @xmath280 \le c{\displaystyle \sum_{s = s+1}^{n}\left[{\displaystyle 1+\sum_{k=0}^{s}}{\displaystyle \binom{s - s+k-1}{k}^{2}}\binom{s}{s - k}\right ] { \displaystyle \binom{n}{s}}p^{-s}}\label{59d}\ ] ] for @xmath2 . [ c4a ]
when the equality holds in ( [ 59c ] ) , @xmath271 decays strictly monotonically with respect to @xmath0 for any rate parameter @xmath281 .
in contrast , @xmath270 $ ] , according to ( [ 59d ] ) , does not follow suit for an arbitrary @xmath281
. however , there exists a minimum threshold , say , @xmath282 , when crossed , @xmath270 $ ] also decays monotonically .
the threshold can be determined from the condition that @xmath223 = \mathbb{e}\left[e_{1,r}\right]$ ] , resulting in the relationship @xmath283 between @xmath1 and @xmath282 .
equation ( [ 59e ] ) supports an exact solution of @xmath284 in terms of @xmath282 , expressed employing the lambert w function @xmath285 and can be inverted easily .
for instance , when @xmath286 , ( [ 59f ] ) yields @xmath287 , the only real - valued solution of interest . depicted in figure [ f1 ] ( left ) ,
@xmath282 derived from ( [ 59f ] ) increases monotonically and strikingly close to linearly with @xmath1 for the ranges of the variables examined . using the equalities in ( [ 59c ] ) and ( [ 59d ] ) , figure [ f1 ] ( right ) presents plots of two normalized errors ,
@xmath270/\sigma^2 $ ] and @xmath288 , against @xmath0 , each obtained for @xmath286 and @xmath289 or @xmath290 , where the variance @xmath291 $ ] .
when the rate parameter is sufficiently low ( _ e.g. _ , @xmath292 ) , the expected rdd error initially rises before falling as @xmath0 becomes larger .
the non - monotonic behavior of the rdd error is undesirable , but it vanishes when the rate parameter is sufficiently high ( _ e.g. _ , @xmath293 ) .
no such anomaly is found in the add error for any @xmath281 . and
@xmath1 ( left ) and normalized rdd and add errors versus @xmath0 for @xmath286 , @xmath289 or @xmath290 ( right ) . ] the expected error @xmath294 $ ] from the best rdd approximation , expressed in terms of the error @xmath295 from the best add approximation , where the best approximations are obtained by setting @xmath296 , is @xmath297=2^{n}e_{n-1,a},\;1\le n<\infty.\label{60}\ ] ] [ c4 ] due to the factor @xmath17 in ( [ 60 ] ) , the expected error from the best rdd approximation as @xmath298 can be significantly large unless the best add approximation commits an error equal to or smaller than @xmath299 . in reference to corollary [ c4a ] ,
suppose that @xmath276 .
then @xmath294 \le c(2/p)^n$ ] .
therefore , @xmath294 \to 0 $ ] as @xmath298 for @xmath300 .
the error analysis presented in this paper pertains to only second - moment characteristics of @xmath24 .
similar analyses or definitions aimed at higher - order moments or probability distribution of @xmath4 can be envisioned , but no closed - form solutions and simple expressions are possible . however ,
if @xmath4 satisfies the requirements of the chebyshev inequality or its descendants @xmath198 a condition fulfilled by many realistic functions @xmath198 then the results and findings from this work can be effectively exploited for stochastic analysis .
see the longer version of the paper for further details .
two variants of dimensional decomposition , namely , rdd and add , of a multivariate function , both representing finite sums of lower - dimensional component functions , were studied .
the approximations resulting from the truncated rdd and add are explicated , including clarifications of parallel developments and synonyms used by various researchers . for the rdd approximation , a direct form , previously developed by the author s group ,
was found to provide a vital link to subsequent error analysis .
new theorems were proven about the expected errors from the bivariate and general rdd approximations , so far available only for the univariate rdd approximation , when the reference point is selected randomly .
they furnish new formulae for the lower and upper bounds of the expected error committed by an arbitrarily truncated rdd , providing a means to grade rdd against add approximations .
the formulae indicate that the expected error from the @xmath0-variate rdd approximation of a function of @xmath1 variables , where @xmath2 , is at least @xmath3 times larger than the error from the @xmath0-variate add approximation .
consequently , add approximations are exceedingly more precise than rdd approximations at higher - variate truncations .
the analysis also finds the rdd approximation to be sub - optimal for an arbitrarily selected reference point , whereas the add approximation always results in minimum error .
therefore , rdd approximations should be used with caveat . | the main theme of this paper is error analysis for approximations derived from two variants of dimensional decomposition of a multivariate function : the referential dimensional decomposition ( rdd ) and analysis - of - variance dimensional decomposition ( add ) .
new formulae are presented for the lower and upper bounds of the expected errors committed by bivariately and arbitrarily truncated rdd approximations when the reference point is selected randomly , thereby facilitating a means for weighing rdd against add approximations .
the formulae reveal that the expected error from the @xmath0-variate rdd approximation of a function of @xmath1 variables , where @xmath2 , is at least @xmath3 times greater than the error from the @xmath0-variate add approximation .
consequently , add approximations are exceedingly more precise than rdd approximations .
the analysis also finds the rdd approximation to be sub - optimal for an arbitrarily selected reference point , whereas the add approximation always results in minimum error .
therefore , the rdd approximation should be used with caution . |
we acknowledge helpful discussions with dr .
e. lstedt , dr . x. wang and prof .
j. h. eberly .
we thank dr .
c. cirelli and prof .
u. keller for kindly providing experimental data .
this work was supported by the national science funds ( no .
60925021 , no .
61275126 and no .
11234004 ) and the 973 program of china ( no .
2011cb808103 ) .
99 see recent reviews , v. s. popov , phys .
uspekhi * 47 * , 855 ( 2004 ) ; w. becker , x. j. liu , p. j. ho , j. h. eberly , rev .
phys . * 84 * , 1011 ( 2012 ) .
m. v. ammosov , n. b. delone and v. p. krainov , sov .
jetp * 64 * , 1191 ( 1986 ) .
a. m. perelomov , v. s. popov , m. v. terentev , sov .
jetp * 23 * , 924 ( 1966 ) .
l. v. keldyshi , sov .
jetp * 20 * , 1307 ( 1965 ) .
x. m. tong , c. d. lin , j. phys .
b * 38 * , 2593 ( 2005 ) .
b. walker , _ et al .
lett . * 73 * , 1227 ( 1994 ) .
a. n. pfeiffer , c. cirelli , m. smolarski , r. dorner , u. keller , nature phys .
* 7 * , 428 ( 2011 ) .
a. n. pfeiffer , _ et al .
. j. phys .
* 13 * , 093008 ( 2011 ) .
a. fleischer , _ et al .
lett . * 107 * , 113003 ( 2011 ) .
p. eckle , _ et al .
_ , nature phys . * 4 * , 565 ( 2008 ) ; p. eckle , _ et al .
_ , science * 322 * , 1525 ( 2008 ) .
y. m. zhou , _
et al . _ ,
lett . * 109 * , 053004 ( 2012 ) ; _ ibid .
_ , phys . rev .
a * 86 * , 043427 ( 2012 ) .
x. wang , j. tian , a. n. pfeiffre , c. cirelli , u. keller , j. h. eberly , arxiv:1208.1516 ; x. wang , j. h. eberly , j. chem .
* 137 * , 22a542 ( 2012 ) ; j. h. eberly , invited talk on the workshop on super intense laser - atom physics , spe .
2012 , suzhou , china .
x. wang , j. h. eberly , phys .
lett . * 103 * , 103007 ( 2009 ) .
f. mauger , c. chandre , t. uzer , phys .
lett . * 105 * , 083002 ( 2010 ) ; _ ibid . , _ * 104 * , 043005 ( 2010 ) .
p. j. ho , _ et al .
lett . * 94 * , 093002 ( 2005 ) ; _ ibid .
_ , * 97 * , 083001 ( 2006 ) ; s. l. haan , _ et al .
97 * , 103008 ( 2006 ) .
w. quan , _ et al .
lett . * 103 * , 093001 ( 2009 ) ; h. liu , _ et al .
lett . * 109 * , 093001 ( 2012 ) .
w. becker , h. rottke , contemp . phys . * 49 * , 199 ( 2008 ) . j. b. watson , _ et al . _ , phys . rev . lett . * 78 * , 1884 ( 1997 ) .
g. d. gillen , m. a. walker , l. d. van woerkom , phys .
a * 64 * , 043413 ( 2001 ) .
note that the `` knee '' structure of mg in circularly polarized field reported in this work can be simulated by using the crapola model .
m. feit , j. fleck , jr . , and a. steiger , j. comput . phys . * 47 , * 412 ( 1982 ) .
x. m. tong , k. hino , and n. toshima , phys . rev .
a * 74 , * 031405(r ) ( 2006 ) p. q. wang , a. max sayler , k. d. garnes , b. d. esry , i. ben - itzhak , opt .
* 30 * , 664 ( 2005 ) see the supplementary material . c. m. maharjan , _ et al .
a * 72 * , 041403(r ) ( 2005 ) . | the experimental observations of sequential double ionization ( sdi ) of ar [ a. n. pfeiffer _
et al . _ , nature phys . * 7 * , 428 ( 2011 ) ] , such as the four - peak momentum distribution and the ionization time of the first and second electrons , are investigated and reproduced with a quantum model by including and excluding the @xmath0-@xmath0 correlation effect . based on the comparison of experiment and simulation , the role of @xmath0-@xmath0 correlation in sdi is discussed .
it is shown that the inclusion of @xmath0-@xmath0 correlation is necessary to reproduce the momentum distribution of electrons .
tunneling ionization of atoms , molecules and semiconductors exposed to strong laser fields is one of the most fundamental quantum processes and has been of significant interest over the past several decades @xcite .
theories @xcite have been well established for single electron ionization based on the keldysh frame @xcite . by numerically solving the time - dependent schrdinger equation ( tdse ) with single - active - electron ( sae ) , tong and
lin also proposed a more accurate formula @xcite , which draws up the well - known ammosov - delone - krainov ionization rate @xcite to the numerical simulation via an empirical correction factor , thereby predicting the ionization rate quite precisely .
nevertheless , for double ( or multielectron ) ionization , the underlying dynamics become more complicated and have not been fully understood so far .
it is generally recognized that two electrons can be ionized either in a non - sequential or sequential process @xcite . in the former mechanism ,
so - called non - sequential double ionization ( nsdi ) , the @xmath0-@xmath0 correlation plays an essential role and the nsdi yield is remarkably enhanced compared with the independent - electron ionization theory @xcite . in the latter mechanism , so - called sequential double ionization ( sdi ) , two electrons are released independently step by step , therefore the sdi yield can be simulated by using ionization theory based on sae .
nevertheless , recent experiments @xcite have doubted the validity of independent - electron approximation in the sdi regime .
specifically , fleischer _ et al . _
@xcite reported an angular correlation between the first and second ionization steps . in ref .
@xcite , pfeiffer _ et al .
_ measured the release time of sdi for ar by using the attoclock technique @xcite .
the measured release time is in good agreement with the prediction of tong - lin s formula for the first electron , but is much earlier than the theoretical prediction for the second electron .
such a finding has attracted significant interests immediately .
zhou _
et al . _
@xcite have reproduced the experimental release time with a classical model by including the @xmath0-@xmath0 correlation .
later , wang _
et al .
_ @xcite argued that the release time can be successfully explained with a similar classical model by properly adopting the soft - core parameter but excluding the @xmath0-@xmath0 correlation , which implies that @xmath0-@xmath0 correlation is not essential in sdi . even though the classical model is robust , which has been demonstrated extensively in previous investigations @xcite , and
also provides a straightforward interpretation with electron trajectory analysis @xcite , the physical picture of sdi is still unclear .
actually , to produce the experimental data , a scaling factor is adopted in @xcite to shift the simulation curve .
strictly speaking , the experiment has not been reproduced .
several curial questions remain : does the @xmath0-@xmath0 correlation play an essential role in sdi ?
what is its influence and how can its effect be observed ?
can we find another indicator to identify the @xmath0-@xmath0 correlation , except for ionization times and angular correlation @xcite ? to clarify these questions , we investigate sdi of ar in an elliptically polarized field by using a quantum model .
we show that the @xmath0-@xmath0 correlation does affect sdi .
it leads to that a small part of the electrons ionize in a correlated way .
this effect influences the observed ionization time only slightly , but it affects the overall shape of the momentum distributions . a complete description of a two - electron system in an elliptically polarized field requires to simulate the motion for each electron in at least two dimensions , which is almost impossible at present @xcite . to overcome the enormous computational challenge , the so - called crapola quantum model @xcite was introduced , which has excellently reproduced the `` knee '' structure of nsdi @xcite . in this model , the total wave function is written , @xmath1 in the high - intensity elliptically polarized field , the electron is quickly removed from the core and the recollision is significantly suppressed
then we can assume that the overlap between the two electrons is sufficiently small for the exchange term and can be negligible in comparison to the coulomb repulsion .
thus only the first term needs to be considered .
the time evolutions of the first ( i.e. , outer electron , @xmath2 ) and second ( i.e. , inner electron , @xmath3 ) electrons are described by @xmath4\psi_n(\mathbf { r}_n , t),\ ] ] where atomic units ( a.u . )
are adopted and @xmath5 denotes two - dimensional coordinates @xmath6 , @xmath7 refer to the first and second electrons , v@xmath8 and @xmath9 $ ] is the driving laser field .
@xmath10 , @xmath11 and @xmath12 refer to the ellipticity , duration and carrier - envelope phase ( cep ) , respectively . in the crapola quantum model , the first electron
was assumed to move in a static effective potential due to the nucleus and inner electrons , i.e. , @xmath13 .
for the second electron , the interaction with the first electron was taken into account .
thus , v@xmath14 includes a static effective potential due to the nucleus plus a time - dependent potential due to the first electron , @xmath15 where @xmath16 , @xmath17 and @xmath18 are the soft - core parameters for the e - ion and @xmath0-@xmath0 interactions .
the second electron is correlated with the first one through the time - dependent term of eq .
[ e3 ] . we could include or exclude the @xmath0-@xmath0 correlation by setting @xmath19 or 0 . in the former case
, we adjust the @xmath0-@xmath0 correlation strength by changing @xmath18 to identify its influence .
note that in the crapola quantum model , the correlation of the first electron on the second is taken into account while its counteractive , i.e. , the time - dependent potential of the second electron on the first electron , is not explicitly included . in order to remedy this seemingly unreasonable treatment ,
we performed other calculations where the time - dependent potential of the second electron on the first electron is also explicitly taken into account .
thus , the potential for the first electron is @xmath20 we call this method revised correlation model in this work .
we employ the split - operator method @xcite to numerically solve eq .
( 2 ) .
like in ref .
@xcite , the two - dimensional space of each electron is partitioned into two regions : the outer region a \{@xmath21 } and the inner region b \{@xmath22 } with @xmath23=100 a.u . in the inner region ,
the wave function is propagated exactly in the presence of combination of the coulombic potential and the laser field . in the outer region ,
the wave function is propagated under the volkov hamiltonian analytically and the final momentum spectra of the first electron @xmath24 and the second @xmath25 are obtained from the wave function in this region @xcite .
the inner and the outer regions are smoothly divided by a splitting technique@xcite .
the ground states of the two electrons are obtained by imaginary - time method . in order to identify
how @xmath0-@xmath0 correlation affects sdi , calculations were performed by treating the time - dependent potential of @xmath0-@xmath0 interaction in different ways as mentioned above , i.e. , including ( revised correlation model ) and excluding ( uncorrelated model ) it for both electrons , and the crapola model .
the data shown below are the results from the uncorrelated and revised correlation model , while the results from the crapola model is shown in the supplemental material for comparison @xcite . , @xmath26 , @xmath27 .
dashed lines show the simulation result with revised correlation model , i.e. , @xmath19 , @xmath28 , @xmath29 , @xmath30 .
, width=377 ] figure [ fig1 ] shows the simulated ionization times of the first and second electrons for the 7-fs and 33-fs pulses .
the central wavelengths of the 7-fs and 33-fs pulses are 740 nm and 780 nm and the ellipticities are 0.78 and 0.77 , respectively . for the 7-fs pulse , four cep values , 0 , @xmath31 and @xmath32 ,
are adopted and the final simulation results are averaged over the cep . for a comparison with experimental data , the intensity profile of the laser focus , the density distribution of the atoms in the gas jet , and the geometrical overlap of the focus and the gas jet need to be considered .
because of uncertainties of these parameters in experiment @xcite , a 3d focal volume averaging ( assumption of constant gas density across the entire 3-dimensional gaussian beam @xcite ) and a 2d focal volume averaging ( assumption of constant gas density only at the waist of the beam and zero gas density otherwise , indicating a gas jet small compared to the rayleigh range of the laser focus ) are tested here .
we first analyze the result for the 3d focal volume averaging .
as shown in fig .
[ fig1](a ) , the simulated ionization time agrees well with the experimental data for both electrons , regardless of the @xmath0-@xmath0 correlation being excluded ( solid lines ) or included ( dashed lines ) . for the 2d focal volume averaging , as shown in fig .
1(b ) , both calculations also agree very well with the experimental data for the first electron . for the second electron ,
the calculations with uncorrelated model agree with the experimental data at the high laser intensity range while deviates slightly at the low laser intensity range .
the calculations with revised correlation model agree better with the experimental data at low intensity .
figure [ fig1](c ) and ( d ) show the simulation results for the 33-fs pulse by assuming a 3d and 2d focal configuration , respectively .
one can see that the @xmath0-@xmath0 correlation indeed influences the ionization time , but the influence is slight .
additionally , the absolute values of the ionization times depend also on the details of the focal geometry , which is unknown in experiment .
the ionization times simulated by uncorrelated model , revised correlation model and crapola model ( see fig .
s2 in @xcite ) all can explain the experiment if accounting for the uncertainties in experiment .
thus , it is difficult to judge the role of @xmath0-@xmath0 correlation in sdi based the ionization time .
these ionization times read in the experiments and our calculations are integral signals , which have erased the details in sdi , obstructing our way on revealing the role of @xmath0-@xmath0 correlation in sdi .
therefore , another indicator which keeps more details of the ionization process and robust against uncertainties in the focal volume averaging , is needed to identify the role of @xmath0-@xmath0 correlation . ] and low [ ( b ) : 1 pw/@xmath33 intensities .
the solid and dashed lines show the projection onto the x and y axis .
( c ) and ( d ) : momentum distribution of ar@xmath34 ion along the x axis as a function of laser intensity for the 33-fs and 7-fs pulses .
electron correlation is taken into account and 3d focal volume averaging is applied.,width=321 ] figure [ fig2](a ) and ( b ) show the two - dimensional momentum distribution of ar@xmath34 for 33-fs pulse at 3 pw/@xmath35 and 1 pw/@xmath35 , respectively . here
the @xmath0-@xmath0 correlation is taken into account by using the revised correlation model and 3d focal volume configure is considered . by projecting onto the major axis of the ellipse ( y axis ) , the momentum distribution is close to gaussian , whereas along the minor axis ( x axis ) , it displays 3 peaks at low intensity and 4 peaks at high intensity .
the outer two peaks correspond to electrons that are emitted into parallel direction and the inner peaks correspond to antiparallel electron emission @xcite .
figure [ fig2](c ) shows the momentum distribution along the minor axis as a function of laser intensity .
as shown in this figure , the inner band gradually expands with increasing the intensity and bifurcates into two branches from 1.5 pw/@xmath35 . while for the 7-fs pulse , as shown in fig .
[ fig2](d ) , the inner band also expands with increasing the intensity but does not bifurcate over the intensity range considered in this work .
all these features agree well with the experiment @xcite .
simulations are also performed by using the uncorrelated model and crapola model .
the ion momentum distributions do not display remarkable difference except that the depth of valley between the inner and outer peaks change slightly ( see figs . s3 and s4 in @xcite ) . and
p@xmath36 for the 7-fs pulse with peak intensity 3 pw/@xmath35 .
uncorrelated model and revised correlation model are used in ( a ) and ( b ) .
( c ) experimental correlation spectrum .
( d ) reduced spectrum of p@xmath37 and p@xmath36 . here
, the 3d focal volume averaging is applied.,width=321 ] to get a deeper insight into the role of @xmath0-@xmath0 correlation , we investigate the correlation spectrum of the radial momentum , which is defined as p@xmath38p_x^2+(\epsilon^2 + 1)p_y^2}$ ] , and p@xmath37 and p@xmath36 are the radial momenta of the first and the second electrons , respectively .
figures [ fig3](a ) and ( b ) show the simulation results by using the uncorrelated model and the revised correlation model for the 7-fs pulse .
figure [ fig3](c ) presents the correlation spectrum observed in the experiment .
the difference between figs .
[ fig3](a ) and ( b ) is remarkable .
the distribution including @xmath0-@xmath0 correlation is fatter than that excluding @xmath0-@xmath0 correlation .
the result including @xmath0-@xmath0 correlation agrees better with the experimental observations .
this feature can be more clearly seen in the reduced spectra , which are respectively obtained from the momentum spectra of the first electron @xmath39 and the second electron @xmath25 . as shown in fig .
[ fig3](d ) , for both electrons , the widths of the spectra excluding @xmath0-@xmath0 correlation are narrower than the experimental data . while for the spectra including @xmath0-@xmath0 correlation , the widths are in agreement with the experimental data .
in order to further confirm that the broadening of the reduced spectra originates from the @xmath0-@xmath0 correlation , we performed other calculations where the strength of @xmath0-@xmath0 correlation is enhanced by decreasing the soft - parameter @xmath18 .
as shown in the supplementary material @xcite , the reduced spectra become wider for both electrons for a smaller value of @xmath18 .
note that a fat radial momentum distribution is also obtained by the crapola model @xcite .
these results indicate that sdi indeed is influenced by @xmath0-@xmath0 correlation and this influence leaves imprint on the electron momentum spectra . as shown in figs .
3(a ) and 3(b ) , the peaks of the momentum spectra do not change dramatically when @xmath0-@xmath0 correlation is excluded or included .
therefore the mean value of radial momentum is not dramatically influenced by @xmath0-@xmath0 correlation . in the attoclock experiment @xcite , the ionization time
is extracted from the radial momenta .
therefore , the influence of the @xmath0-@xmath0 correlation on the averaged ionization time is not very effective .
similar remarks also can be observed for the 33-fs pulse .
this feature also explains the puzzle why the ionization time can be reproduced with the classical model either including or excluding the @xmath0-@xmath0 correlation @xcite . .
uncorrelated model and revised correlation model are adopted in ( a ) and ( b ) .
the solid lines with dots and the dashed lines represent the vector potential of the laser field and the radial momentum , respectively , width=226 ] the origin of the broadening of the reduced momentum spectra is related to the time intervals when the first and the second ionizations occur .
figure [ fig4](a ) shows the ionization rate as a function of time at 3 pw/@xmath35 for the 7-fs pulse .
the bold and thin solid lines correspond to the first and second electrons without @xmath0-@xmath0 correlation .
because the laser intensity is much higher than the saturation intensity , the ionization is confined in a single cycle around t@xmath40 and t@xmath14 at the ascending part of the pulse , respectively . after ionization ,
the electrons are accelerated by the laser field and finally released with a momentum of p@xmath37 and p@xmath36 as shown in fig . [ fig4](a ) .
but because of no correlation , the second electron is not freed when @xmath41 and the first ionization process is already finished when @xmath42 .
in contrast , when the @xmath0-@xmath0 correlation is accounted for [ fig . [ fig4](b ) ] , the centra of the ionization time ( @xmath43 and @xmath44 ) for both electrons almost do not change .
this is consistent with the above results that the ionization time is not effectively influenced by the @xmath0-@xmath0 correlation .
however , as shown in fig .
4(b ) , when @xmath41 , small percentage of the second electron starts to be released and also small percentage of the first electron is released at @xmath42 .
consequently , the ionization windows of the first and second ionization processes become broader than those excluding the @xmath0-@xmath0 correlation .
the long ionization window broadens the width of the momentum distribution .
moreover , because the first electron quickly moves away after ionization , its correlation with the second electron drops quickly .
only a small percentage of the first and second electrons can be ionized in a correlated way . by increasing the @xmath0-@xmath0 correlation
, larger percentage of the electrons can be ionized in a correlated way and the momentum spectra will be further broadened as shown in the supplementary material @xcite .
therefore , the width of the momentum spectrum , which is closely related to the ionization window , exhibits the faint details of @xmath0-@xmath0 correlation in sdi . in conclusion , sdi of ar subjected to elliptically polarized pulses is investigated using quantum models .
it is shown that the @xmath0-@xmath0 correlation indeed slightly shifts the ionization time , but its influence is not crucial to explain the experiment @xcite .
moreover , uncertainties in the focal volume averaging also influence the apparent ionization times .
it is difficult to decisively reveal the role of @xmath0-@xmath0 correlations based on the ionization timing data alone .
however , our simulations show that the @xmath0-@xmath0 correlation broadens the time windows for the first and second ionizations .
though this effect could not be distinguished in the ionization time of the second electron , it affects the momentum distribution .
the signature of @xmath0-@xmath0 correlation can be identified in the width of the electron momentum distribution . by comparing with the experiment @xcite ,
the width of the electron momentum distribution for 7-fs pulse agrees well with the simulation including fair @xmath0-@xmath0 correlation , which indeed indicates the exist of @xmath0-@xmath0 correlation in sdi .
the subtle features of momentum distribution require further experimental measurements with low noise and high accuracy . |
in 2013 the besiii and belle collaborations reported the observation of a charged state in the charmonium sector called @xmath0 @xcite ( this state was also seen at cleo @xcite ) .
the measured mass and width are @xmath2 soon after its discovery it was proposed that these states are @xmath1-wave hadronic molecules @xcite . later a further analysis in this
framework and different approaches for the substructure like tetraquarks were presented in ref .
@xcite . in this work
, we will assume that the @xmath3 is indeed a hadronic molecule .
a comparison with combinations of charm mesons shows that this states lies close to a specific threshold and thus favors this interpretation .
denoting the wave function with the respective particle name , one can identify their main component to be @xcite @xmath4 first , the constituents are charm mesons whose masses are far above @xmath5 and second , their binding energy is rather small , i.e. below the pion mass @xmath6 or at least of that order .
hence , for the description of these states one can use in a first approximation a pionless effective field theory where the constituents are treated as non - relativistic , point - like particles which only interact via contact interactions . such an short - range effective theory is called eft(@xmath7 ) and is a commonly used tool in nuclear physics @xcite .
its expansion parameter is @xmath8 , where @xmath9 is the scale of the modulus @xmath10 of the internal momenta @xmath11 of the involved particles which are - according to the small binding energy of the molecule , @xmath12 - below the pion mass @xmath6 .
we already remark here that the characteristics of the @xmath3 are such that eft(@xmath7 ) is barely applicable , but it still allows for some first insights into the electromagnetic structure of this exotic particle . in particular , using the charm meson masses from ref .
@xcite the ( averaged ) binding energy @xmath13 can be obtained as @xmath14 \nonumber\\ & = & ( -23.13 \ : \pm \ : 6.01 ) \ : \rm{mev } \label{binding_energy_z_c } \:.\end{aligned}\ ] ] the negative value implies a resonance instead of a charm meson bound state solution . neglecting the imaginary part of the binding momentum @xmath15 ( cf .
[ sec : forma ] ) , we treat the @xmath0 as a virtual state with negative scattering length . the binding momentum obtained from the extracted binding energy , eq .
( [ binding_energy_z_c ] ) , is @xmath16 where the uncertainty is entirely from the one in the binding energy . in leading order ,
the electromagnetic properties are independent of the sign of @xmath15 .
our eft expansion parameter is @xmath17 and since the binding momentum @xmath15 should be a low - energy scale proportional to @xmath9 to provide a proper counting scheme , one concludes that @xmath18 must be smaller than @xmath19 .
this is not the case for our estimate of @xmath15 in eq .
( [ binding_momentum ] ) .
however , one can use the predictions for the @xmath3 as first , _ rough _ model estimations and we will do so in the following .
our predictions can be made more quantitative by explicitly including pions in the theory .
for the charm molecule @xmath3 the analysis of its decay channel shows that the quantum numbers @xmath20 are favored @xcite , which agrees with the @xmath1-wave hypothesis .
thus the @xmath3 can be described by an effective field theory with @xmath1-wave contact interactions similar to the loosely - bound charm molecule @xmath21 @xcite .
the structure of the @xmath3 state can then be calculated analog to the deuteron treated in eft(@xmath7 ) as a baryonic molecule of nucleons in a relative @xmath1-wave with total spin @xmath22 @xcite .
besides this analogy , there is also a conceptual connection to the so - called halo eft developed in refs .
@xcite and in particular to its application in ref . @xcite .
following refs .
@xcite , one can write the effective lagrangian up to next - to - leading order ( nlo ) for the positively charged @xmath3 as @xmath23z \nonumber\\ & & - g \biggl[z^{\dagger } \bigl(d^{+ } \bar{d}^{*0 } + d^{*+ } \bar{d}^{0}\bigr ) \ : + \ : \bigl(d^{+ } \bar{d}^{*0 } + d^{*+ } \bar{d}^{0}\bigr)^{\dagger } z \biggr ] \nonumber\\ & & - \mu_{+ } ( d^+)^{\dagger } \mathbf{u } \cdot \mathbf{b } d^+ \ : - \ : \mu_{*+ } ( d^{*+})^{\dagger } \mathbf{u } \cdot \mathbf{b } d^{*+ } \nonumber\\ & & - \mu_z z^{\dagger } \mathbf{u } \cdot \mathbf{b } z \:.\end{aligned}\ ] ] this lagrangian and our formalism are generic for non - relativistic particles . the kinetic terms for the @xmath24 and @xmath25
mesons simply reproduce the free schrdinger equation which is linear in the time derivative .
moreover , @xmath26 is an auxiliary field describing the molecular state @xmath0 , @xmath27 is the coupling constant for @xmath28 interactions , @xmath29 is a constant and @xmath30 is a phase whose sign depends on the effective range .
we stress that the @xmath3 is not treated as a dynamical degree of freedom .
we use the `` dimeron '' formalism of kaplan @xcite which is convenient to treat non - perturbative contact interactions .
if the auxilliary field is integrated out , the langrangian ( [ z_lagrangian ] ) reduces to one with contact interactions between the @xmath24 and @xmath25 mesons .
if the phase @xmath31 is negative , the @xmath26 field is a ghost but the scattering amplitude for the @xmath24 and @xmath25 mesons is unitary and sensible .
this method is well established in effective field theory treatments of resonant interactions in nuclear and particle physics @xcite .
note also that the spin structure of the @xmath25 and @xmath3 is hidden in eq .
( [ z_lagrangian ] ) since the interactions are spin - independent .
since there are only @xmath1-wave interactions , the spin state of the @xmath3 is always equal to the spin state of the constituent @xmath25 and remains unchanged in the @xmath28 interaction vertex . minimal substitution with electric charge @xmath32 induces an @xmath33 interaction with the constituents and the molecule itself , leading to the feynman diagrams which contribute to the charge form factor .
the magnetic form factor instead comes from the terms in the last line of eq .
( [ z_lagrangian ] ) , where @xmath34 is the magnetic field and @xmath35 is the spin-1 generalization of the pauli vector for spin-1/2 particles , with @xmath36 being the generators of the rotation group .
those terms are proportional to the magnetic moment @xmath37 of the respective charged particle and describe purely magnetic interactions of the spatial part of the vector potential @xmath38 with both the charged @xmath24 mesons and @xmath3 .
other contributions to the lagrangian , for instance those relevant for the quadrupole form factor or effects due to the anomalous magnetic moment of neutral @xmath24 mesons , are at least one order higher and thus not shown .
finally , we note that the @xmath3 binding energy scale of order 20 mev is small compared to the @xmath24-@xmath25 mass splitting of about 140 mev .
our lagrangian is only valid in a small energy region close to the @xmath3 mass and thus heavy quark symmetry is not relevant .
the auxiliary field @xmath26 is not dynamical but due to its coupling to the constituents it is dressed by @xmath24 meson loops , see fig .
[ dyson_zc ] .
full @xmath26 propagator ( thick solid line ) as sum over charm meson loops .
the double solid line represents the bare @xmath26 propagator , the single solid line stands for @xmath39 mesons , the dashed one for @xmath24 mesons and the dots represent higher order terms.,scaledwidth=47.5% ] this leads to a full @xmath26 propagator @xmath40 in terms of the bare propagator @xmath41 and the self - energy @xmath42 . to avoid an unnatural scaling of the scattering length ( similar to that discussed in ref .
@xcite ) , one calculates the self - energy in the pds scheme introduced in refs .
@xcite ; see ref .
@xcite for an alternative solution to this problem . in terms of the pds scale @xmath43 and the mean
reduced mass @xmath44 of the molecule , one obtains @xmath45 \:.\end{aligned}\ ] ] from the full propagator of the @xmath26 field at nlo , @xmath46^{-1 } \:,\end{aligned}\ ] ] one can deduce the scattering length @xmath47 and the effective range @xmath48 of the @xmath3 by matching the scattering amplitude @xmath49 to the effective range expansion ( ere ) @xmath50 at nlo the following relations are obtained : @xmath51 in addition this result fixes the phase introduced in the lagrangian eq .
( [ z_lagrangian ] ) . to get a positive value for the effective range
, the phase @xmath31 must be equal to minus one .
furthermore , we calculate the wave function renormalization constant @xmath26 defined as the residue of the pole of the propagator in eq .
( [ full_propagator_z ] ) . at nlo
one finds @xmath52 where the @xmath3 binding momentum @xmath53 was introduced whose definition is chosen such that one takes care of both , bound _ and _ virtual states .
+ the non - relativistic matrix elements of the electromagnetic current @xmath54 can be expanded in terms of form factors . for a vector particle state @xmath55 with three momentum @xmath11 and linear polarization vector @xmath56 , one can write in the breit frame@xcite @xmath57~ , \label{j^0-current_breitframe}\\ \langle \mathbf{p'},\varepsilon_j\left|j^k\right|\mathbf{p},\varepsilon_i\rangle & = & -i\frac{e}{2 m } \biggl[g_m(q ) ( \delta^k_j q_i - \delta^k_i q_j)\biggr ] \label{j^k - current_breitframe}\:.\end{aligned}\ ] ] here , @xmath58 is the three - momentum transfer with @xmath59 and @xmath60 is the number of space - time dimensions ( here @xmath61 ) .
the charge ( @xmath62 ) , magnetic ( @xmath63 ) and quadrupole ( @xmath64 ) form factors are normalized in the following way @xcite : @xmath65 in terms of the magnetic moment @xmath37 of the particle and its quadrupole moment @xmath66 . according to eq .
( [ j^0-current_breitframe ] ) , the charge form factor is determined by the diagrams shown in fig .
[ g_c_nlo_general_zc ] in which the incoming photons are @xmath33 photons and where the initial and final spin of the @xmath3 are equal .
diagrams contributing to the charge and magnetic form factor of the @xmath3 at nlo .
while the left diagrams in parenthesis are lo , the right diagram is the nlo contribution .
the thick solid line represents the full @xmath26 propagator , the thin solid line stands for @xmath39 and the dashed line for @xmath24 mesons .
the incoming photons are indicated by wavy lines and the application of wave function renormalization factors is indicated by the crosses.,scaledwidth=48.0% ] the charge form factor at nlo of the @xmath3 is then given by @xmath67 \ : .
\end{aligned}\ ] ] this is indeed correctly normalized to one . a second observable besides the form factor itself is the expectation value of the squared charge radius @xmath68
. it can be found from the low - energy expansion of the form factor , @xmath69 . at nlo
one finds @xmath70 if the incoming photons in fig .
[ g_c_nlo_general_zc ] are @xmath38 photons , one can choose the polarizations in eq .
( [ j^k - current_breitframe ] ) and calculate the amplitude of the corresponding diagrams to get the magnetic form factor @xmath63 of the @xmath3 . at nlo
it is given by @xmath71 \:.\nonumber \\ & & \end{aligned}\ ] ] so the difference to the charge form factor are the magnetic moments in front of each term . with the normalization condition in eq .
( [ formfactor_normalization ] ) .
it is possible to identify a relation between the magnetic moment of the molecule @xmath3 and that of its charged constituents @xmath72 and @xmath73 @xmath74 in addition , one can determine the expectation value of the squared magnetic radius @xmath75 by expanding the magnetic form factor up to second order in @xmath76 .
at nlo one finds @xmath77 \ : .
\end{aligned}\ ] ] according to eq .
( [ j^0-current_breitframe ] ) , the quadrupole form factor is determined by @xmath33 photon interactions in fig .
[ g_c_nlo_general_zc ] with a spin change between the initial and the final state .
the @xmath33 interaction itself can not induce such a spin change because @xmath33 is just the time - like component of the @xmath78-vector potential @xmath79 .
thus one needs additional operators in the lagrangian which project on specific polarization states @xmath56 . according to ref .
@xcite such operators first appear at nlo and thus the quadrupole form factor @xmath80 vanishes at leading order .
except for the masses , widths and @xmath81 there are no experimental data available for the @xmath0 , hence one can predict observables only at lo . furthermore ,
the leading order magnetic form factor and the magnetic radius are proportional to the magnetic moments of the constituent mesons . as these are experimentally unknown both are not accessible even at lo .
thus only the charge form factor and the charge radius at lo can be predicted . the results for the scattering length and charge radius of the @xmath3 obtained from the extracted binding momentum , eq .
( [ binding_momentum ] ) are @xmath82 where the uncertainties are entirely from the one in the binding energy , eq .
( [ binding_energy_z_c ] ) . due to the large expansion parameter
, we refrain from estimating the errors due to higher orders .
we reiterate that our predictions for the @xmath3 should thus be considered model estimates .
also , note that the charge radius in eq .
( [ scattering_lengths_and_charge_radii ] ) should be added in quadrature the charge radii of the constituent @xmath83 mesons which are treated as pointlike in our theory .
lo charge form factor @xmath84 of the @xmath0 .
the error band includes only the uncertainty in the binding energy and is depicted in green.,scaledwidth=49.0% ] the charge form factor of the @xmath3 at lo is shown in fig .
[ z_c_g_c_lo_plot ] .
furthermore , there is a the correlation between the leading order charge radius and the binding energy as shown in fig .
[ charge - radius_correlation_z_c_plot ] .
correlation between the expectation value of the squared charge radius @xmath68 and the modulus of the binding energy @xmath13 of the @xmath0 at lo .
the correlation is shown for binding energies within the errors of eq .
( [ binding_energy_z_c ] ) . the central value eq .
( [ scattering_lengths_and_charge_radii ] ) is represented by the cross and the error band is depicted in green.,scaledwidth=49.0% ]
in this work the electromagnetic structure of the exotic meson @xmath0 interpreted as an @xmath1-wave charm meson molecule with spin @xmath22 was investigated in the framework of eft(@xmath7 ) .
similar to the description of the deuteron in this effective theory , the charge , magnetic and quadrupole form factor were formally analyzed up to nlo .
in particular , the charge and magnetic radius were obtained .
due to the poor experimental data on the @xmath3 state only the lo charge radius and charge form factor could be estimated .
furthermore , we derived a relation between the magnetic moment of the molecule and of its constituents , see eq .
( [ moments ] ) .
given the experimental mass of the @xmath0 , the binding momentum of the @xmath0 is a bit too large to allow for a quantitative description of its properties in eft(@xmath7 ) .
thus our results should be viewed as as first , _ rough _ model estimations of the electromagnetic properties of the @xmath0 . to sharpen these predictions , one has to go beyond the pionless eft and include explicit pions . *
acknowledgments * this work is supported in part by the eu integrated infrastructure initiative hadronphysics3 and the dfg through funds provided to the sino - german crc 110 `` symmetries and the emergence of structure in qcd '' .
e. braaten and m. kusunoki , phys .
d * 69 * , 074005 ( 2004 ) [ hep - ph/0311147 ] .
m. t. alfiky , f. gabbiani and a. a. petrov , phys .
b * 640 * , 238 ( 2006 ) [ hep - ph/0506141 ] .
d. l. canham , h .- w .
hammer and r. p. springer , phys .
d * 80 * , 014009 ( 2009 ) [ arxiv:0906.1263 [ hep - ph ] ] .
j. -w .
chen , g. rupak and m. j. savage , nucl .
phys . a * 653 * , 386 ( 1999 ) [ nucl - th/9902056 ] . | the observation of the exotic quarkonium state @xmath0 by the besiii and belle collaborations supports the concept of hadronic molecules .
charmonium states interpreted as such molecules would be bound states of heavy particles with small binding energies .
this motivates their description using an effective theory with contact interactions .
in particular , we focus on the electromagnetic structure of the charged state @xmath0 . using first experimental results concerning spin and parity ,
we interpret it as an @xmath1-wave molecule and calculate the form factors as well as charge and magnetic radii up to next - to - leading order .
we also present first numerical estimations of some of these observables at leading order . |
maxwell s equations@xcite provide a compact and elegant derivation of why a light beam , entering a transparent dielectric material , propagates inside it with a reduced speed . using maxwell s equations ( in gaussian units )
, it is straightforward to derive the wave equations for the electric field @xmath1 in vacuum ( identical equations hold for the magnetic field @xmath2 ) : @xmath3 characterised by dispersiveless propagation with constant speed @xmath4 , the light speed in vacuum .
when , instead , an electromagnetic wave propagates in a dielectric material ( we think of it as homogeneous , isotropic , and non magnetic , and in addition that it responds linearly to the incident radiation ) , a polarization current is induced given by @xmath5 , which makes eq.([eq : ref01 ] ) inhomogeneous through the appearance of a source term @xmath6 : @xmath7 in eq.([eq : ref02 ] ) , @xmath1 and @xmath8 are macroscopic fields of maxwell s equations in dielectric materials , connected by the constitutive relation @xmath9 where @xmath10 and @xmath11 are the frequency dependent electrical susceptibility and index of refraction of the material , which also are macroscopic parameters of the theory .
using the second of the above relations , eq.([eq : ref02 ] ) becomes @xmath12 the source term on the right hand side of eq.([eq : ref03 ] ) reveals an interesting form since it splits into two terms .
a first one is relative to a propagation with the light speed @xmath4 in vacuum and it cancels the corresponding term on the left hand side . a second one re - establishes , after that cancellation , the homogeneous structure of the wave equation in the material with the field that propagates with a reduced ( and dispersive ) speed @xmath0 .
wave propagation in its fully general manifestation is an extremely complex phenomenon for every type of wave motion we consider
. an exceptionally lucid and lively presentation of this complexity is given by feynman in an interview by christopher sykes that is posted on youtube@xcite .
this complexity invites a further investigation into wave phenomena , taking different points of view in order to understand and clarify them at a deeper level .
for example , the previous analysis of light propagation in dielectrics does little to clarify the physical mechanism involved at the microscopic level . for this deeper explanation ,
the principle that we must consider when a light beam enters a dielectric is the atomic nature of matter . a piece of dielectric material is a volume of vacuum space filled with a collection of atoms , each polarized and behaving like a charged harmonic oscillator@xcite ( an oscillating dipole moment ) when acted on by the external incident radiation and by the fields re - radiated by all the other oscillating dipoles of the collection .
this picture , equivalent to the treatment with maxwell differential equations as far as the final result is concerned , is mathematically very different since the atomic description of matter in interaction with radiation naturally leads to a representation through integral equations . from their solutions
we get the expression of the index of refraction @xmath11 of the material ( expressed in terms of the microscopic parameters of the atoms , of their density , and of the frequency of the incident field ) .
in addition , the solutions straightforwardly include an equal but opposite field , generated by the total induced polarization currents , that extinguishes the incident field inside the material ( a fact known in optics as _ extinction theorem_@xcite ) , and its substitution with a new internal field propagating at a reduced speed @xmath0 , despite the fact that the incident beam and the re - radiated waves by the atoms propagate with the vacuum light speed @xmath4 in the interatomic vacuum within the material .
we shall present this point of view , basing our treatment on purely classical electrodynamics , but closely following the neat and concise presentation given by m.l .
goldberger and k.m .
watson in their technical monograph on collision theory @xcite .
let us consider a dielectric medium ( it may be glass ) made of @xmath13 impenetrable identical atoms , kept together by their bonding interactions and distributed randomly within a volume @xmath14 of arbitrary shape with the only restriction that the radius of curvature of its boundary at each point must be large compared to the wavelength @xmath15 of the incident electromagnetic radiation .
we further consider @xmath13 to be a large number and the atomic density @xmath16 to be approximately constant and equal to @xmath17 within @xmath14 .
the medium is placed in the path of a plane electromagnetic wave with wave vector @xmath18 , whose wavelength @xmath19 is much greater than the atomic dimensions ( for example , for visible light we have that the wavelength is @xmath20 the atomic dimensions ) .
we also require that the radiation frequency be significantly less than the frequency corresponding to the lowest excitation energy of the atoms ( @xmath21 ) , so that the interaction of the incident radiation with the atoms may be considered as essentially elastic .
the response of the dielectric material to the impinging radiation is then associated with the scattering of light by the atoms , and we anticipate that the electromagnetic propagation through it will be described by its index of refraction . the microscopic , local electric field ( monochromatic of frequency @xmath22 , with complex time dependence @xmath23 ) is the real part of @xmath24 and has the general structure of the multiple scattering equations @xmath25 where @xmath26 is the incident ( complex ) electric field , and @xmath27 is the field radiated by the @xmath28-th atom , all these fields being evaluated at the position @xmath29 of the @xmath11-th atom in @xmath14 .
this is a complex formula , with the self - consistent signature of multiple scattering problems , since the electric interaction of the @xmath28-th atom at the @xmath11-th atom position , described by the second term on the right hand side of eq.([eq : ref1 ] ) , is itself dependent on the field acting on the @xmath28-th atom , which has exactly the same form of eq.([eq : ref1 ] ) .
having assumed that the incident radiation wavelength is @xmath30 than the atomic dimensions , the dipole approximation applies in this case , and in the second term in eq.([eq : ref1 ] ) we may use the classical scattering formula for @xmath31 , given by@xcite @xmath32,\ ] ] where @xmath33 is the electric dipole moment of the @xmath28-th atom induced by the local field @xmath34 , @xmath35 , and @xmath36 is the gradient operator acting only on the @xmath11-th atom coordinates . with the assumptions of isotropic dielectric and linear response , the induced electric dipole moment of an atom is connected microscopically with the local electric field acting on it by the scalar atomic polarizability , @xmath37 a simple classical formula for the atomic polarizability can be found using the classical charged harmonic oscillator model@xcite .
an easy calculation leads to the result ( for a single resonance frequency ) @xmath38 where @xmath39 are the oscillator characteristic parameters and @xmath22 is the frequency of the electric driving force
. we can now go over to a continuum representation of matter , replacing the sum over @xmath28 in eq .
( [ eq : ref1 ] ) with an integral over the volume @xmath14 of the dielectric except for the exclusion of a sphere @xmath40 with atomic radius and centered at the @xmath11-th atom . since we made the assumptions that atoms are impenetrable , we are guaranteed that only the chosen atom is excluded as required by the sum in eq .
( [ eq : ref1 ] ) .
we thus obtain ( in the following , @xmath29 and @xmath41 should be regarded as continuous vectorial variables ) @xmath42 eq . ( [ eq : ref4 ] ) , together with expressions ( [ eq : ref2 ] ) and ( [ eq : ref3 ] ) , represents the integral equation for the field @xmath43
. its solution will give the electric field at every point inside @xmath14 .
to do this , we proceed according to the arguments of classical macroscopic electrodynamics , defining the macroscopic field in the medium , @xmath44 , as @xmath45 where in this case the integral is over the volume @xmath14 excluding a very small cylindrical , needle - shaped region which includes the point @xmath46 and which is parallel to @xmath44 . subtracting ( [ eq : ref5 ] ) from ( [ eq : ref4 ] ) , and using again this last equation to eliminate the local field @xmath47 , we obtain @xmath48 we thus see that the complex expression of the dipolar interaction among the atoms on the left hand side , that also appears in eq.([eq : ref4 ] ) , is equal to the right hand side combination of the incident field , the internal macroscopic field , and a term that reproduces the one on the left hand side except that the integral now extends to the ( atomic - sized ) volume contained between the sphere @xmath40 and the needle - shaped surface @xmath49 that intersects it . to evaluate this integral , we make a further simplification based on the fact that the wavelength of the incident field is @xmath30 than the atomic dimensions .
this allows us to neglect the term @xmath50 in eq.([eq : ref2 ] ) as compared with @xmath51 ( a quasi - static approximation ) , and to set @xmath52
. we can further consider @xmath53 independent of @xmath41 within @xmath40 , thus replacing @xmath33 with @xmath54 .
then eq.([eq : ref2 ] ) becomes for small @xmath55 : @xmath56 in the final expression of eq .
( [ eq : ref7 ] ) , we have retained only the term that gives a non - zero contribution to the integral on the right hand side of eq .
( [ eq : ref6 ] ) ( see appendix ) .
so the volume integral can be carried out , using standard vector analysis results , transforming it to a surface integral .
this vanishes on the needle - shaped surface of the cylinder , and only receives a contribution from the integration on the surface @xmath40 .
the result is @xmath57 remembering the definition of the polarization vector and using eq .
( [ eq : ref3 ] ) for the electric dipole moment , we can write @xmath58 , and this can be used with the equations ( [ eq : ref4 ] ) , ( [ eq : ref6 ] ) , and ( [ eq : ref8 ] ) to obtain the final result @xmath59.\ ] ] this is the self - consistent solution of the integral equation ( [ eq : ref4 ] ) we were looking for , which evidences how the microscopic response of the material translates into a macroscopic polarization , the source term for all that propagates within the dielectric and also emerges out of it ( internal field , scattered field , as well as , for special geometries with flat interfaces of the dielectric material , refracted , reflected , and transmitted fields ) .
it is also worth noting at this point that no hypotheses have been made about the directions of the incident oscillating electric field and of its wave vector , so that the obtained relation is general .
when we only consider the internal fields , it is immediately evident from ( [ eq : ref9 ] ) that one of the effects of the atomic dipolar re - radiations ( the square bracket in this equation ) is to cancel the incident radiation field from within the dielectric : this is the expression of the _ extinction theorem_. in addition , eq.([eq : ref9 ] ) shows the content of multiple scattering of the atomic dipolar re - radiations as they reproduce the macroscopic polarization also on the right hand side of it , in case of dense materials .
after the cancellation of the incident field , and considering that within an isotropic material the polarization and the electric field vectors are parallel , we obtain @xmath60 where for the second expression the definition of index of refraction has been used .
this is the promised connection between the microscopic parameters and the index of refraction . in case of low density materials ( such as gases ) , we have @xmath61 , and eq .
( [ eq : ref10 ] ) simplifies , providing a simpler relation between the atomic polarizability and the index of refraction .
the equation becomes @xmath62 and , since @xmath63 in the given approximation , it assumes the final form @xmath64 in this case the enhancement effect due to multiple scattering is absent
. since eq.([eq : ref11 ] ) is a very often used expression for the refractive index in the elementary expositions , we would like to add few more comments on it . in the theory of scattering ,
when the above approximation applies , it is common to describe the scattering off a scatterer ( like an atom or , its classical counterpart , a charged harmonic oscillator ) by the scattering amplitude @xmath65 , where the incoming and outgoing wave vectors , in the case of elastic scattering that we considered , satisfy the relation @xmath66 . consider a dielectric material in the form of a very extended but thin slab , with thickness @xmath67 , placed in the plane @xmath68 . from what already assumed , the atoms within it have a very low packaging density ( @xmath61 is satisfied ) . on this slab
a plane electromagnetic wave ( assume unit amplitude , for simplicity ) is incident along the direction @xmath69 .
the electric field computed in a point @xmath70 assumes the form of a superposition of the incident field and of the scattered fields in @xmath69 by all the atoms of the slab@xcite : @xmath71 where in the final results appears the forward ( zero angle ) scattering amplitude @xmath72 , and the expression of the final formula derives from the the assumed approximation .
but the exponent of the last exponential of eq.([eq : ref111 ] ) has a very familiar expression in elementary optics , being equal to @xmath73 , so that , by comparison , we get the well - known expression @xmath74 and , lately , comparing this expression with eq.([eq : ref11 ] ) , we obtain the relation between forward scattering amplitude and atomic polarizability : @xmath75 there exists a remarkable relation ( of general validity for all the scattering processes , not only for light ) , called the _ optical theorem _ , that connects the forward scattering amplitude and the total extinction ( scattering plus absorption ) cross section . for the case considered of elastic scattering ,
this result is easily derived .
having chosen the incident radiation to be of unit amplitude , the measured intensity at a distant forward point @xmath69 is proportional to @xmath76 ) , to get the _ optical theorem _ in this case : @xmath77 and from it , using equations ( [ eq : ref113 ] ) and ( [ eq : ref33 ] ) , to obtain the scattering cross section from a charged oscillating dipole : @xmath78 in eq.([eq : ref116 ] ) , @xmath79 is the classical electron radius , and we used the relation@xcite @xmath80 for the radiation damping of eq.([eq : ref33 ] ) .
this important result , describing the rayleigh scattering , incidentally explains , at an elementary level , the blue color of the sky we see when our line of sight forms a right angle with the white sunlight crossing the atmosphere , as well as its red color at sunrise and sunset . in the opposite case of dense material ,
the relation between the atomic polarizability and the index of refraction is more complex and the formula of lorentz - lorenz is obtained solving eq.([eq : ref10 ] ) for the polarizability : @xmath81
the previous results , in spite of the fact that they are well - known , are only seldom discussed in introductory electromagnetism and optics courses . this , notwithstanding the fact that some of them could easily be deduced from simple experimental observations . in fig.([fig1 ] ) an elementary experiment of optics is shown , that is commonly done to demonstrate snell s laws of reflection and refraction .
a light beam is incident at an angle on a plane air - glass interface , and reflection and refraction beams are produced , all in the same plane of incidence . using his exceptional physical intuition ,
feynman@xcite interprets this simple experimental situation in the light of the atomic model of matter to deduce , with the aid of the energy conservation law , the characteristics of the reflected beam .
he notices that a beam along the dotted line in fig.([fig1 ] ) , which continues the incident beam straight into the dielectric , is not observed .
it should , only if the atoms of the glass were frozen in their positions and could not oscillate under the driving action of the incoming beam , so that the beam can propagate along the dotted line into the interatomic vacuum with speed @xmath4 .
but the atoms do polarize and oscillate driven by the incoming primary field and by the secondary re - radiated fields by the multiple scattering atomic actions .
the total induced polarization current must thus generate , in addition to the refracted beam , another vacuum beam that exactly cancels the incoming beam inside the dielectric .
this is the way feynman uses the _ extinction theorem _ ( without ever naming it ) , just running the argument backwards : a vacuum beam inside is not observed , so , by superposition , there must be an internally generated vacuum beam that extinguishes it .
equally interesting is to study the role of the polarization direction ( perpendicular or parallel to the plane of incidence ) of the incident beam in this experiment . as is well - known
, these two electric field polarizations lead to different expressions for the reflection ( and refraction ) coefficients .
feynman approaches the problem of deducing the different contributions for the two cases according to the idea that only the fields radiated by the projections of the oscillating dipoles perpendicular to the propagation directions of both reflection and extinction fields are non - zero .
he carries out the main calculation steps , also offering hints to arrive at the final answers that he writes down . a detailed derivation along these lines of thought , with some further interesting extensions ,
have recently been published@xcite .
finally , there remains one more question to be noted about the experiment in fig.([fig1 ] ) .
the standard derivation for the reflection and transmission coefficients , for both perpendicular and parallel polarizations , is by matching boundary conditions at the air - glass interface for the macroscopic fields of maxwell s equations , and of course that works very well except that it does not clarify an important point .
in fact , the reflected and transmitted fields appear , in that derivation , to only depend on what happens at the interface , a fact that seems to be also supported by examining the experimental situation in fig.([fig1 ] ) .
however , this is not what we understand with our microscopic calculation , which suggests instead a different physical interpretation .
what generates all of the internal and external fields ( except the incident one ) is the total polarization current that is set up into the dielectric material .
if the material has a plane interface , as in fig.([fig1 ] ) , then the fields radiated by all the atomic dipoles in @xmath14 , coherently organise themselves in order to generate ( in addition to the extinction field ) fields that propagate along the refraction the reflected directions .
this sounds more likely as a bulk effect rather than a surface effect .
the following standard vector identities involving the operator @xmath82 hold : @xmath83 @xmath84 if @xmath85 is a constant vector , then the terms that apply derivatives to @xmath85 are zero , obtaining @xmath86 @xmath87 the term @xmath88 can be eliminated by the two previous expressions , getting @xmath89 the complete expression for the eq.([eq : ref7 ] ) is thus : @xmath90 the volume integral of eq .
( [ eq : ref6 ] ) is thus composed by the three terms that appear in eq.([eq : ref13 ] ) , and it can be evaluated using the geometry shown in fig.([fig2 ] ) . ) . on the left , the differential vectorial area @xmath91
is shown which as usual points to the outside of the closed surface @xmath40 , centered at the position of the @xmath11-th atom .
on the right , the polar coordinate system is shown , with the dipole moment @xmath54 chosen to be in the @xmath69-direction .
note that the unit vector pointing to the outside is @xmath92.,width=529 ] the integral of the first term of ( [ eq : ref13 ] ) is converted to a surface integral .
the integral over the needle - shaped surface of the cylinder gives a negligible contribution , while the one on the surface @xmath40 results in the contribution reported in ( [ eq : ref8 ] ) .
this can be seen using @xmath93 and arbitrary choosing @xmath94 . with these choices ,
we thus obtain @xmath95 the second term of ( [ eq : ref13 ] ) is easily seen to give a zero contribution to the integral : @xmath96 finally , the third term of ( [ eq : ref13 ] ) is not contributing since the origin is not included in the integration domain : @xmath97 this demonstrates the result reported in eq .
( [ eq : ref8 ] ) . | we present a pedagogic derivation of the electromagnetic field established in a dielectric material by an impinging external field .
we consider the problem from the point of view of the physical mechanism involved at the microscopic level .
the internal field emerges when the material is thought of as an assembly of atoms in vacuum , each of them being polarized by the external incident field and by the re - radiated fields of all the other polarized atoms of the material . in this way
, each atom becomes itself a source of secondary radiation that adds and interferes with all the other internal fields ( including the internal extension of the externally impinging field ) , contributing to build up the total internal field within the dielectric material as well as the externally scattered field .
this picture naturally leads to a connection between the microscopic properties of the material and its index of refraction , that describes the dielectric response to the applied electromagnetic field .
calculations also show that the incident radiation is extinguished inside the dielectric by an equal but opposite field generated by the total induced polarization currents , and is substituted by a macroscopic electric field which propagates with the new speed @xmath0 . |
dark matter ( dm ) is one of the defining issues and long - standing mysteries of modern physics .
nowadays , there is an overwhelming amount of data , coming from cosmological , large - scale - structure , and galactic - scale observations , which lead us to conjecture that more than 80% of the total matter content of the universe is most probably in the form of a not - yet - identified non - baryonic component , dm .
dm has been inferred so far through its gravitational properties only .
there are however several scenarios in which dm is expected to couple to ordinary matter , with the currently most investigated scheme postulating weakly interacting massive particles ( wimps ) .
the wimp scenario ( see , e.g. , ref .
@xcite for a review of key wimp properties ) is particularly appealing since it naturally predicts a dm relic density matching the observed cosmological dm density , and numerous extensions to the standard model of particle physics feature viable wimp dm candidates ( the lightest neutralino in supersymmetry being the prime example ) .
wimps have masses in the gev - tev regime and can annihilate in pairs into lighter particles with a rate of the order of the weak interaction rate ( with the canonical value needed to match the observed dm relic density for a thermally produced candidate being @xmath0 ) .
indirect detection studies have mainly been focused on the search for a wimp - induced component in the neutrino and local antimatter cosmic - ray fluxes , and for an excess in the multi - wavelength galactic or extra - galactic emissions ( see , e.g. , @xcite ) .
electrons and positrons can be directly or indirectly injected by wimp annihilations . in particular , and with the exception of wimp models annihilating exclusively into neutrinos , a sizable final branching ratio of annihilation into @xmath1 is a general feature of wimp models ( see e.g. fig . 4 in ref .
emitted in environment with background magnetic fields , high - energy electrons and positrons give rise to a synchrotron radiation extending typically from the radio to the infrared bands .
dwarf spheroidal ( dsph ) galaxies have been recognized as optimal laboratories for indirect dm searches @xcite . they are not only the closest ( other than the galaxy itself ) and most dm dominated objects in the local universe , but they are also the faintest and most metal - poor stellar systems known ( see , e.g. , ref .
@xcite for a recent review ) .
the presence of gas is tightly constrained by limits on hi emission @xcite , confirming the lack of recent star formation .
no evidence for diffuse thermal or non - thermal emissions at x - rays and gamma - rays has been reported , implying very low quantities of either hot ionized and relativistic plasmas .
the expected large dm signal , on one side , and low astrophysical background , on the other , make local dsphs an ideal target where to search for a wimp signature .
various attempts have been pursued in this context to detect wimp - induced prompt emission of gamma - rays or a radiative emission ( inverse compton ( ic ) scattering in the x- and gamma - ray bands and synchrotron radiation at radio frequencies ) associated with wimp - induced electrons and positrons .
no evidence of diffuse signal has been obtained so far at any relevant frequency and this only allowed to set upper limits on the dm annihilation / decay rate for a wide range of wimp masses .
the greatest effort for dm searches in dsphs has been so far undertaken by means of observations with gamma - ray telescopes .
recently , the fermi - lat collaboration have reported on 4-year observations of 25 dsph satellite galaxies of the milky - way ( mw ) @xcite .
none of the targets have been detected with presented gamma - ray flux upper limits extending between 500 mev and 500 gev .
constraints on the dm annihilation cross section have been derived from a joint likelihood analysis of a subset of 15 dsphs . at 95% c.l . and assuming an nfw dark matter distribution
, the bounds constrain the `` thermal '' annihilation rate @xmath0 for masses @xmath2 gev in the cases of @xmath3 and hadronic final states ( while being an order of magnitude weaker for lighter leptonic final states ) .
this study updates results from a 2-year analysis involving 10 dsphs @xcite , with bounds compatible within about a factor of 2 .
tev searches towards the most promising dsphs have been performed in the past ten years with ground - based cherenkov telescopes , such as whipple @xcite , magic @xcite , veritas @xcite , and hess @xcite .
this technique is most sensitive to heavy wimps ( with mass above 100 gev or so ) and the derived exclusion limits are at the level of @xmath4 for dm annihilating into quarks and @xmath3 .
recent analyses making use of gamma - ray data to place constraints on the wimp dm parameter space include @xcite ( see also references therein ) .
the population of non - thermal electrons and positrons induced by dm annihilations can be tested by means of x - ray observations of ic scattering on the cmb .
future observations can have interesting prospects in this respect @xcite .
archival data of the xmm - newton telescope in the fields of ursa minor , fornax , and carina dsphs were used by ref .
@xcite to constrain the wimp parameter space .
bounds are at the level of @xmath5 for @xmath6 gev , assuming a mw - like spatial diffusion and an nfw dm profile .
radio observations of dsph galaxies have been also very recently attempted through the green bank telescope ( gbt ) @xcite , aiming to test the prediction of @xcite .
the fovs of draco , ursa major ii , coma berenices , and willman i were mapped at 1.4 ghz with a resolution of 10 arcmin and a sensitivity of 7 mjy / beam ( after discrete source subtraction ) .
no significant emission was detected from the dsphs , and ursa major ii and willman 1 were found to be the cleanest cases to be exploited for setting bounds on the wimp parameter space . for a scenario where the dm annihilates into @xmath7 , the dm spatial distribution follows an nfw profile , the magnetic field strength is @xmath8 g , and the cosmic - ray particle diffusion is described as in the case of the galaxy , ref .
@xcite found @xmath9 for @xmath10 gev at 95% c.l .
( for a more extended discussion , see ref .
@xcite ) . in this paper ( paper iii ) , we present the results of the search for wimp radio signatures in six dsphs ( carina , fornax , sculptor , bootesii , hercules and segue2 ) making use of observations performed with the australia telescope compact array ( atca ) operating in the 1.1 - 3.1 ghz band .
in section [ sec : obs ] , the observing setup and data reduction process are summarized ( see paper i @xcite and paper ii @xcite for further details ) .
essentially , the dsph targets were observed with both a compact core and long baselines , with sensitivity below 0.1 mjy ( 0.05 mjy ) for a synthesized beam of 1 arcmin ( 6 arcsec ) .
we describe the models adopted for the expected wimp signal ( involving both microscopic properties and the dm spatial profile ) and for the dsph interstellar medium in section [ sec : mod ] .
the possibility of a detection of point - like emissions induced by wimps is investigated in section [ sec : detect ] .
after subtracting point - sources , no firm evidence for an extended emission over the dsph size ( few arcmin ) was obtained . in section [ sec : bound ] , we present the associated bounds on the wimp parameter space , for different final states of annihilation / decay and for different astrophysical assumptions . in section [ sec : compar ] , these constraints are compared with the bounds on wimps in dsphs obtained by other experiments summarized above .
an outlook of the possibilities offered by the next - generation radio telescopes , which can significantly improve the sensitivity to wimp signatures , is described in section [ sec : prospects ] .
section [ sec : concl ] summarizes our conclusions .
the observations employed in this paper were performed during july / august 2011 with the six 22-m diameter atca antennae operating in the frequency range 1.1 - 3.1 ghz .
the targets included three classical dsphs ( cds ) , carina , fornax , and sculptor , and three ultra - faint dsphs ( uds ) , bootesii , hercules and segue2 , for a total of 123 hours of observing time . the array configuration was formed by a core of five antennae with maximum baseline of about 200 m , and a sixth antenna located at about 4.5 km from the core .
details about the observing setup are provided in paper i. the data were reduced using the _ miriad _ data reduction package @xcite , following a standard calibration scheme . since we performed interferometric observations , data are collected in terms of spatial coherence of radiation measured at two points , or , in other words , in terms of correlated visibilities .
they consist of complex quantities made by the amplitude and phase of the cross - product of measured voltages from two antennae .
the visibility function is a fourier transform of the sky brightness distribution which can be inverted to produce an image of the sky .
the visibility plane is also sometime called @xmath11-plane with @xmath12 and @xmath13 being its coordinates . because of the duality between image and visibility planes , short distances in an image correspond to large spacings in the @xmath11-plane , and viceversa . at the central wavelength of observation ( @xmath14 cm ) ,
such relation can be seen as @xmath15 arcmin , with @xmath16 being the angle in the sky - image and @xmath17 being the baseline length . in the imaging process , we produced two types of maps.regis/c2499.html . ] first , data were imaged in high - resolution maps ( where the short baselines of the core are down - weighted ) which can probe scales from few arcsec to few tens of arcsec , and have an rms noise of 30 - 40 @xmath18jy . in the second set of maps ,
we apply a gaussian taper of 15 arcseconds ( which down - weighs long baselines involving the sixth antennae ) to the data before fourier inversion . the synthesized beam of these maps becomes about 1 arcmin , and the largest scale which can be well imaged is around 15 arcmin . because of confusion limitation , the rms noise raises up to 0.1 - 0.15 mjy .
the main properties of the maps are summarized in table 1 of paper ii .
details about the data reduction can be found in paper i. the tapered maps have a better sensitivity for testing the presence of a truly diffuse emission on scales above few arcmin , because of their larger beam .
for these maps , point sources ( i.e. , discrete sources in the high resolution maps ) are a `` background '' that can be subtracted .
we perform this subtraction before fourier inversion ( namely , in the visibility plane ) with the clean component of the corresponding high - resolution map used as the input source model .
this procedure is described in sec .
3.1.1 of paper ii , and allows to bring the rms of the tapered map down by a factor of few ( see fig . 5 of paper ii ) .
we will refer to these resulting images as map1 .
however , some residuals are left and they are mostly due to a non - perfect reconstruction of the synthesized beam shape . in order to remove them , a further subtraction can be performed in the final image of the tapered maps by removing point - like structures ( which can be , in fact , remnant of a non - perfect subtraction in the visibilities ) , as described sec .
3.1.2 of paper ii .
the latter procedure ( whose output will be called map2 ) might however be somehow risky since the beam of the map ( and thus the size of point - sources ) is about 1 arcmin , while the emission we are looking for is just a factor of few larger in size . in order to be conservative , when deriving bounds , we will mostly refer to maps with sources subtracted in the visibility plane only . to understand the impact of different imaging / subtraction methods on the intensity bounds , see also fig . 6 in paper ii . in fig
[ fig : fake ] , we show an example of the expected theoretical signal ( left panel ) and how it would have been reconstructed in our maps ( right panel ) . for illustrative purposes , we focus on the example of the fornax dsph . the simulated map is obtained by fourier transforming the theoretical emission of the left panel , and adding the outcome to the actual data in the visibility plane .
this procedure has been performed by means of the task uvmodel in _ miriad_. after the simulated uv - data are derived , we reduced and imaged them following the procedure adopted to build the original maps ( described in section iii of paper i ) .
[ fig : fake]b can be compared with the original map in fig . 1 of paper ii . in section iiib of paper ii , we demonstrated that the kind of signals we are going to discuss in the rest of the paper can be actually detected ( within expected error bars ) in our maps , by following the analysis pipeline we adopted .
in the standard wimp scenario , wimp particles are stable but can annihilate in pairs injecting given species of particles ( here we are interested in @xmath19 ) . the dm wimp source scales with the number density of wimp pairs locally in space , i.e. assuming a smooth ( without substructures ) , spherically symmetric , and static dark matter distribution , with @xmath20 , where @xmath21 is the halo mass density profile at the radius @xmath22 , and @xmath23 is the mass of the dm particle .
the source term associated to the production of @xmath19 is given by : q^a_e(e , r)=_a v ( e ) , [ eqq ] where @xmath24 is the velocity - averaged annihilation rate , and @xmath25 is the number of electrons / positrons emitted per annihilation in the energy interval @xmath26 , obtained by weighting spectra for single annihilation channels over the corresponding branching ratio .
another possibility is that dm particles have a long but finite lifetime , and electrons and positrons are injected in dm decays . in this case
the source function takes the form : q_e^d(r , e)=_d ( e ) , [ eqq2 ] where @xmath27 is the decay rate and @xmath28 is the number of electrons / positrons emitted per decay in @xmath26 .
the dm density profile @xmath29 in dsph galaxies is typically reconstructed from star kinematic data , as mentioned below .
all the other quantities in eqs .
[ eqq ] and [ eqq2 ] are instead related to the particle physics properties of dm , and they are the main subject of investigation in this paper . [ cols="<,^,^,^,^,^,^,^ " , ] n - body numerical simulations suggest the dm distribution in structures to be clumpy .
therefore , one could explore the possibility to detect single clumps in our maps . in
the source catalogue reported in paper i , we have 1392 detected sources . even filtering the catalogue by discarding identified sources ( i.e. , with an association to a known astrophysical source at other wavelength ) and sources lying outside the virial radius , we are still left with a large number of sources and it would be rather speculative to analyze all of them in terms of dm emission from subhalos .
moreover , although it is in principle not excluded to have small clumps appearing as point - sources , it is very likely that spatial diffusion would make their size much broader . in this case
, they will therefore rather contribute to the global extended emission of the dsph . assuming a certain subhalo distribution
, one can in principle compute the additional component of the dsph diffuse emission given by dm annihilations in substructures .
on the other hand , the mass function of subhalos in dwarf galaxies is still very uncertain .
n - body simulations do not currently reach the resolution to study the subhalo distribution for main halo mass @xmath30 , as would be required in the case of dsphs .
we conservatively decide not to include a substructure boost to the total diffuse emission of annihilating dm . in summary
, we neglect any possible contribution from subhalos . in paper ii , we tested different spatial profiles for the diffuse emission in dsphs and , from the analysis of single dsphs , found no significant evidence . here
we re - perform the test in the ave and pes scenarios , by setting the halo parameters @xmath31 and @xmath32 to the value in table [ tab : dmprof ] , and fitting all the dsph simultaneously with two free parameters ( @xmath33 and @xmath34 ) , since dm microscopic properties are considered to be universal . similarly to the results in paper ii , for the maps with point - sources subtracted in the visibility plane only ( map1 ) , we obtain a significant statistical evidence , at the level of 12-@xmath35 for both the ave and pes scenarios . since our measurement concerns only one frequency bin and the injection spectrum
does not have a great impact on the source morphology ( as we will better discuss below ) , the fit is not able to break the degeneracy between mass and annihilation rate ( which both enters in the normalization of the signal ) .
very mild preference for large masses is found with best - fit around @xmath36 tev and @xmath37 ( @xmath38 ) for annihilations into @xmath7 .
however , as discussed in paper ii , the model is actually fitting a residual emission from sources ( due to a non - perfect subtraction ) , rather than a truly diffuse emission . after removing the remaining sources in the image plane of maps
having already undergone source subtraction in the fourier space ( map2 ) , the signal case is no longer preferred with respect to the null hypothesis . therefore no compelling evidence has been found . in section
[ sec : bound ] , we proceed to compute constraints on @xmath24 and @xmath23 . to further explore the possibility of a detection
, one can perform a stacking of the 6 dsph maps .
the stacking is not totally well - defined since we expect a different size for each dsph , and there is not robust scaling factor that can be adopted .
however , the prediction of an excess in the inner few arcmin is common to all the cases .
we simply stacked together the 6 dsphs without any rescaling and fit the resulting map with a phenomenological form for the emissivity following either an nfw or a burkert profile ( with free @xmath31 and overall normalization of the signal ) .
the operation is performed for the two types of source - subtracted maps ( i.e. , map1 and map2 ) .
the results are similar as in the case of the combined fit discussed above , namely , a significant detection for the case with sources subtracted in the visibility plane only ( due to remnants of point - sources ) which disappears when considering maps with sources subtracted also in the image plane .
before deriving constraints on the wimp parameter space , we discuss the degree of uncertainty associated to the dm profile . in fig .
[ fig : boundr0rho0]a , the impact of the profile shape on the bounds is shown .
we report the ratios of the annihilation cross - sections for the einasto and burkert profiles over the nfw case .
they are computed taking @xmath32 and @xmath31 from table [ tab : dmprof ] .
the absence of a central cusp in the burkert case makes constraints weaker , while the einasto profile provides the most stringent constraints . for this plot
, we consider the no - diffusion case and the tapered map with sources subtracted in the visibilities . the overall uncertainty due to the profile shape is between a factor of 2 to 15 , depending on the dsph .
note that this conclusion depends on the diffusion scenario adopted and on the type of map considered .
for example , the presence of sources , especially near the center , and a noise which varies from the center to the outer part of the map , can have a significant impact on the ratio of the bounds .
this is because a cuspy scenario is obviously more constrained by the inner data , while a more extended emission involves a comparison with the full map .
however , for all the maps and diffusion scenarios considered here , the overall uncertainty stays roughly within one order of magnitude .
note that in the case of hercules ( and to a lesser extent of carina ) the variation is quite limited .
this is because the size of the dsph is relatively small , and the stellar velocity dispersion is relatively well - known .
the latter allows a more precise determination of the dm profile with respect to the other uds ( bootesii and segue2 ) .
fornax and sculptor , due to their larger size , have instead a larger area involved in the fit which makes the dependence on the shape stronger than in the hercules case ( which , for the diffusion scheme and type of maps adopted in this plot , nearly resembles a point - source ) . in fig .
[ fig : boundr0rho0]b , we show the uncertainty on the cross - section bounds arising from the uncertainty on @xmath32 and @xmath31 . in order to derive the allowed range ,
we consider the 1-@xmath35 band for @xmath39 and @xmath40 reported in table 2 of @xcite and assume they are fully correlated ( we need to make an approximation since the covariance is not provided ) . in other words , we took the 1-@xmath35 upper values of both @xmath39 and @xmath40 , translate them into @xmath32 and @xmath31 and compute the upper value of @xmath24 ( shown in fig .
[ fig : boundr0rho0]b ) , and similarly for the lower value , taking the 1-@xmath35 lower values of @xmath39 and @xmath40 . to check our approximation
, we computed the so - called j - factor , which is given by @xmath41 . taking @xmath42 sr ,
our findings matches with the results reported in table i of @xcite , which were computed with a full implementation of the method of @xcite .
[ fig : boundr0rho0]b shows that the uncertainty related to the observationally allowed range of @xmath32 and @xmath31 is within about one order of magnitude .
again , this factor varies depending on the diffusion scenario adopted , but we checked that such variation turn always out to be relatively small . the results of fig .
[ fig : boundr0rho0 ] thus shows that the uncertainty related to the dm profile has a smaller impact on the bounds than the description of dsph magnetic properties . on the other hand
, we should bear in mind that the analysis of @xcite , similarly to other works in the field , includes a number of priors , which somehow restrict the allowed ranges for the parameters , while more conservative and broader ranges could be in principle considered . in order to bracket all the uncertainties
associated to the required astrophysical inputs , we set three benchmark scenarios , opt , ave , and pes , which are listed in table [ tableb ] and discussed in section [ sec : dsph ] .
we compute bounds on the wimp microscopic properties in the @xmath43 plane for different final states of annihilations / decays .
the likelihood is computed as described in in section 5.1 of paper ii and focusing on the central dsph region , taking a radius of 30 arcmin ( 20 arcmin ) for the cds ( uds ) .
the relevant region for the statistical test will be actually related to the source area , since the part of the likelihoods outside this region cancels out in the likelihood ratio ( being identical between different models ) .
the source region can be significantly smaller than 20 - 30 arcmin , and at most corresponding to the largest achievable scale of the maps , which is of the order of 15 arcmin ( 30 arcsec ) when considering low ( high ) resolution maps , see paperii .
on top of single target analyses , we also perform a combined analysis where the likelihood is given by the product of the likelihoods of the each dsph .
as already mentioned , the uncertainty in the centroid position of the dsphs considered here is typically estimated to be below arcmin level . on top of that , the center of the spherical dm distribution might be slightly offset with respect to the stellar distribution . however , since our sensitivity is rather homogeneous on scales comparable to the dsph size , we do not expect a significant variation of the bounds due to possible misalignment between the assumed center of the spherical distribution of our models and the real dsph dm center . .
the annihilation final state is taken to be @xmath44 .
the bounds in the left panel have been obtained taking the most constraining map between the tapered and un - tapered maps . in the central and right panels ,
the bounds come from the tapered map with source subtracted in the visibility plane ( for the dotted lines , we further subtract sources in the image plane ) .
, scaledwidth=80.0% ] .
the annihilation final state is taken to be @xmath44 .
the bounds in the left panel have been obtained taking the most constraining map between the tapered and un - tapered maps . in the central and right panels ,
the bounds come from the tapered map with source subtracted in the visibility plane ( for the dotted lines , we further subtract sources in the image plane ) .
, scaledwidth=80.0% ] .
the annihilation final state is taken to be @xmath44 .
the bounds in the left panel have been obtained taking the most constraining map between the tapered and un - tapered maps . in the central and right panels ,
the bounds come from the tapered map with source subtracted in the visibility plane ( for the dotted lines , we further subtract sources in the image plane ) .
, scaledwidth=80.0% ] as clear from fig .
[ fig : dmsize ] , the scenario with loss at injection place can be probed by both the tapered and high - resolution maps , if the profile has an nfw or einasto shape . for bounds on the opt scenario
, we will use the map which provides the strongest constraint ( since , in this case , low and high resolution maps provide comparable bounds ) , and we do not subtract point - sources . for the other two benchmark scenarios , ave and pes , the emission is on scales @xmath45 , so the bounds can be only obtained from the tapered maps . in these cases ,
as already mentioned , we can consider two types of approaches and maps .
the more conservative approach relies on maps ( map1 ) where the subtraction of sources is performed in the visibility plane only and involves sources from the un - tapered maps ( we remind that the un - tapered maps have a synthesized beam of about 6 arcsec and largest detectable scale of 30 arcsec ) . in the more challenging approach , we consider map2 which are built from map1 by further subtracting point - like sources in its image plane . in the tapered maps ,
the synthesized beam is @xmath46 , so here for point - like structures we refer to sizes @xmath47 .
results are shown in figs .
[ fig : ann1 ] , [ fig : ann2 ] , and [ fig : dec]a . in fig .
[ fig : ann1 ] , we consider annihilation into @xmath7 and derive 95% c.l .
bounds in the three benchmark astrophysical models .
left panel shows that the opt scenario is indeed very optimistic .
the large value of the equipartition magnetic field and the cuspy @xmath1 distribution ( resulting from the cuspy dm distribution and the assumed radiative loss at injection place ) make the constraints extremely strong .
the `` thermal '' annihilation rate @xmath48 is confidently excluded by all the six dsph bounds .
the caveat here is , in addition to the optimistic assumptions , that the dm center could be offset with respect to the optical center , and one of the detected sources in table [ tab : sources ] can be actually induced by wimp annihilations , as discussed in section [ sec : ps ] .
the behaviour of the curves ( bounds increasing with mass ) can be easily understood from eq .
[ eqq ] , and is due to the fact that for a given energy density the number density decreases as the mass increases . for the specific case of fig .
[ fig : ann1]a , fornax and segue2 bounds come from the high resolution maps , while the other bounds arise from the low resolution ones .
as expected , the combined analysis slightly improves the bound obtained in the most constraining case . the closest dsphs ( segue2 and bootesii )
provide the strongest bounds .
the central panel of fig .
[ fig : ann1 ] focuses on the ave scenario .
note the upturn at low masses occurring in all the dsphs .
this is due to the fact that low mass wimps induce @xmath1 whose distribution might peak at energies below the value corresponding to the synchrotron power peak , which , at 2 ghz , is given by @xmath49 ( where @xmath50 is the magnetic field in @xmath51 ) . the significantly larger magnetic field in the opt scenario was preventing this upturn . in the ave case ,
the bounds from the cds becomes at the same level as for the segue2 and bootesii dsphs .
this is due to two reasons .
first , the formers have a significantly larger size , and in turn a larger confinement time , which makes the depletion of @xmath1 due to spatial diffusion less effective .
second , cds have experienced a more significant star formation , thus have a larger magnetic field @xmath52 ( see table 2 in paper ii ) . as already noticed in paper ii , if the source subtraction in the visibility plane was not completely successful ( as especially for the fornax and sculptor fovs ) , the bounds that can be derived from map1 and map2 can significantly differ , up to two orders of magnitude .
this can be noticed comparing the dashed and solid lines in fig .
[ fig : ann1]b .
while in map2 the rms noise is only a factor of few lower than in map1 , the difference in the derived bounds is much larger for dsphs with significant detection of extended emission in map1 ( due to residuals from point - source subtraction ) .
indeed , in this case , the constrained flux has to obviously be above the best - fit flux , which is significantly above the rms . on the contrary , in map2
no significant deviation from the no - signal case has been found , which means that the best - fit flux for the diffuse emission is close to zero . for similar reasons
, the combined analysis does not improve the constraints in the case of map1 . indeed , the likelihood of the cases with large residuals ( in particular , of the sculptor dsph where the detection is at 11-@xmath35 , see table 3 in paper ii ) significantly reduces if a negligible flux is considered , and this drives down the global likelihood for low values of @xmath24 .
this discussion highlights the importance of subtraction of sources .
the confusion limit is indeed one of the greatest obstacles to be overcome in this kind of studies . in this respect ,
future radio telescopes with high sensitivity and spatial resolution ( like , e.g. , the ska and its precursors ) will be crucial for the study of the dm nature ( see discussion in @xcite ) . in the right panel of fig .
[ fig : ann1 ] , we show the constraints in the pes scenario .
the discussion is similar to the one reported for the ave case .
the effect of spatial diffusion is even more pronounced and the larger cds become more promising than the smaller uds , even though the latter are closer ( see also fig .
11 in paper ii ) .
the fornax dsph is the most constraining case .
note that this trend is the opposite with respect to what found in the opt scenario ( and typically also in gamma - ray searches ) . in fig .
[ fig : ann2 ] , we show bounds on the wimp annihilation cross section for few different final states of annihilation , @xmath7 and @xmath53 in the left panel , and @xmath3 and @xmath54 in the right panel . in order to be conservative , we choose to show bounds making use of map1 . assuming the residuals in map1 are not due to dm annihilations
, we can consider all the annihilation rates which are incompatible with at least one target to be excluded .
the case of @xmath53 is very similar to the @xmath7 case discussed in fig .
[ fig : ann1 ] ( and reported in fig .
[ fig : ann2 ] for completeness ) .
indeed , the two @xmath1 yields are very similar for electron / positron energy below @xmath55 ( with @xmath56 gev to have a kinematic allowed production of @xmath53 ) , see e.g. fig . 4 in ref .
@xcite , namely in the relevant energy range for the synchrotron production .
annihilations in @xmath3 and @xmath54 induce instead harder spectra of @xmath1 .
bounds for the leptonic channels are thus more stringent than in the hadronic case at low wimp masses , while the opposite picture occurs in the tev range .
[ fig : dec]a shows 95% c.l .
constraints in the decaying - dm parameter space for the @xmath7 and @xmath54 final states . again
the softer @xmath1 spectrum of the hadronic case makes the latter more constraining than the leptonic channel for high wimp masses and viceversa at low mass .
the bounds in the ave and pes scenarios differ by a factor which is of the same order of the findings in the annihilating case .
the curve of the opt scenario is instead much closer to the other two benchmark models , and , overall , the uncertainty is significantly reduced .
the dependence on the square of the dm density profile in the annihilating case ( with respect to a linear dependence for decaying dm ) amplifies the effect of a cuspy initial distribution , if this is not flattened by spatial diffusion as in the ave and pes scenarios .
the huge overdensity of the injected @xmath1 makes also the equipartition magnetic field large at the center , and these two points explain why the large discrepancy between opt and ave found for annihilating dm is not present in fig .
[ fig : dec]a . in the case of decaying dm
, there is no natural scale to aim as for the `` thermal '' annihilation rate in the annihilating dm framework .
however , much interest was recently devoted to decaying dm models in connection to the possibility of explaining the measured raise in the local cosmic - ray positron fraction at high energy .
the green region in fig .
[ fig : dec]a shows the best - fit to the pamela excess in the case of @xmath54 final state ( see , e.g. , ref @xcite ) , which is the only viable interpretation among the final states considered here . .
in the ave and pes cases , bounds are derived from the tapered maps with sources conservatively subtracted in the visibility plane only . left : annihilations into @xmath7 ( solid ) and @xmath53 ( dotted ) .
we show also the bounds from the analysis of dsphs in refs .
@xcite ( light - green ) and @xcite ( dark - green ) at gamma - ray frequencies , and in ref .
@xcite ( orange , ave scenario ) in the x - ray band , for the @xmath7 finale state .
right : annihilations into @xmath3 ( solid ) and @xmath54 ( dotted ) .
we show also the bounds from the analysis of dsphs in refs .
@xcite ( light - green ) and @xcite ( dark - green ) at gamma - ray frequencies , and in ref .
@xcite ( orange , ave scenario ) in the x - ray band , for the @xmath3 finale state.,scaledwidth=80.0% ] . in the ave and pes cases ,
bounds are derived from the tapered maps with sources conservatively subtracted in the visibility plane only . left : annihilations into @xmath7 ( solid ) and @xmath53 ( dotted ) .
we show also the bounds from the analysis of dsphs in refs .
@xcite ( light - green ) and @xcite ( dark - green ) at gamma - ray frequencies , and in ref .
@xcite ( orange , ave scenario ) in the x - ray band , for the @xmath7 finale state .
right : annihilations into @xmath3 ( solid ) and @xmath54 ( dotted ) .
we show also the bounds from the analysis of dsphs in refs .
@xcite ( light - green ) and @xcite ( dark - green ) at gamma - ray frequencies , and in ref .
@xcite ( orange , ave scenario ) in the x - ray band , for the @xmath3 finale state.,scaledwidth=80.0% ]
very few studies have so far attempted at deriving particle dm bounds from radio data in the dsph fields ( as proposed in @xcite ) .
the analysis most directly comparable to ours has been recently presented in refs .
it is based on radio observations of the ursa major ii and willman 1 fovs at 1.4 ghz with the gbt telescope . the reported sensitivity ( 7 mjy / beam after subtracting discrete sources ) is about two orders of magnitude weaker than the rms noise level of the atca observations considered in this work ( at 2 ghz ) . for peaked spatial distributions ,
our bounds are therefore about two orders of magnitude more constraining . for shallower emission profiles ,
the bound scaling is , on the other hand , lower than the ratio of the sensitivities since the beam of the gbt observations ( 10 arcmin ) is much larger than the atca beam ( 1 arcmin ) , and so the former telescope can integrate more signal . in the cases with significant diffusion considered in refs .
@xcite , the current work improves wimp constraints by about one order of magnitude
. the actual ratio might be actually slightly larger since in our modeling the magnetic field ( diffusion coefficient ) decreases ( increases ) exponentially outside the stellar region while in @xcite they are taken to be constant . for a given diffusion scheme , we can directly compare our radio constraints with x - ray bounds arising from ic scattering on the cmb in dsphs .
the only assumption which is present in the radio estimate while being not crucial in the x - ray case concerns the magnetic field strength .
focusing on the ave scenario , we found that the bounds at @xmath10 gev derived in this work are about four orders of magnitude stronger than those derived in ref .
@xcite ( which stem from x - ray observations of the fornax dsph ) .
they are shown in fig .
[ fig : ann2 ] in orange for @xmath7 and @xmath3 final states , after rescaling the constraints presented in their fig . 4 for the diffusion model of the ave scenario ( for simplicity , we employed a linear scaling , correcting from a diffusion coefficient of @xmath57 @xmath58/s to @xmath59 @xmath58/s , see also fig .
3b in ref .
this means that , with current data , radio observations have a significantly better constraining power than x - rays provided a value of the magnetic field strength @xmath60 g .
finally , we can compare results of fig .
[ fig : ann1 ] with bounds obtained at gamma - ray frequencies .
the radio signal is , however , much more uncertain , due to the unknown description of spatial diffusion and magnetic field , while neither ingredient is needed in the computation of the prompt gamma - ray signal . the current , most constraining dm bounds from the analysis of dsphs
are obtained by the fermi - lat @xcite ( below about 1 tev ) and magic @xcite ( in the tev regime ) collaborations . for definiteness , we focus again on the comparison with the ave scenario .
the fermi - lat bounds are slightly more constraining than fig .
[ fig : ann1 ] for @xmath7 , while of the same order of magnitude for the @xmath3 channel . in the latter case
, the magic observations significantly improve the constraints in the region above few hundreds of gev .
notice that the ic emission is not included in the work of both @xcite and @xcite , while ( within the ave scenario ) is expected to slightly improve constraints in the tev regime for the @xmath3 channel .
radio astronomy is entering a golden era . in the next few years , the square kilometre array ( ska ) project will be completed in its phase-1 building the world s largest radio telescope and the phase-2 will further improve its sensitivity and spectral range .
its pathfinders , askap and meerkat , are in the final constructing phase .
the first continuum images have been already produced by a six antenna test prototype of askap , the boolardy engineering test array ( which will be part of the full askap array ) , and by a seven antenna precursor of meerkat , the kat-7 .
the european new - generation radio interferometer , the low - frequency array ( lofar ) , has recently started operations and will offer an unprecedented coverage of the low - frequency range from 10 to 240 mhz . since the end of 2012 , the karl g. jansky very large array ( jvla ) project has been completed significantly improving the sensitivity of current radio telescopes . all these instruments can be fruitfully employed for the dm search strategy discussed in this work . in order to provide quantitative predictions , we focus on few observational setups which will be available in the years to come .
first , one can realize a survey similar to the one presented here , but with the interferometric jvla array in the northern sky ( so also probing different targets ) .
the rms sensitivity for an integration time of 1 hour per pointing is about @xmath61jy in the 1 - 2 ghz band , and the d - configuration is the best - suited one for the detection of diffuse emissions .
the d - configuration consists of 27 antennae with dish diameter of 25 meters arranged along the three arms of a y - shape and has a maximum and minimum baseline of 1 km and 35 m , respectively . on the other hand , observations in other configurations with higher resolution ( longer baselines ) are also needed , in order to provide an accurate mapping of point - sources for the subsequent subtraction ; indeed the confusion limit of d - configuration is around @xmath62jy .
an overall gain of at least a factor of 2 in the sensitivity can be confidently expected .
one of the key science projects of askap is the evolutionary map of the universe ( emu ) .
emu is a survey project which is about to commence operations and will perform a deep continuum survey at ghz - frequency covering the entire southern sky with rms sensitivity of @xmath61jy .
askap will have 36 independent beams giving a total field of view of 30 square degrees , and with a resolution of about 10 arcsec . with the emu data at hand
, it will be possible not only to improve the sensitivity with respect to the project discussed here , but also to dramatically enlarge the number of observed dsphs .
indeed , 14 known dsph milky - way satellites will be within the covered area , and we expect this number to be further increased by forthcoming optical surveys , in particular of the southern sky ( e.g. , dark energy survey @xcite , skymapper @xcite , large synoptic survey telescope @xcite ) .
indeed , similarly , in the past few years , the sloan digital sky survey data have more than doubled the number of known dsph satellites in the northern sky .
putting together the increase in sensitivity and the larger dsph sample , we can foresee a gain in the constraining power of a factor of 5 - 10 .
the south african meerkat radio telescope is currently being built , with science operations in the final 64-antenna configuration expected to start in 2017 .
its fov will be more limited than the askap one , but will have higher survey speed : deep observations of the most promising dsph targets are probably the best strategy to pursue with such a radio telescope . with few hours of integration time over the region of a single dsph , a @xmath63jy rms level can be achieved at ghz - frequency , increasing the sensitivity to diffuse emission in dsph by a factor of @xmath64 with respect to the present observations .
we do not discuss in details the prospects for the ska precursors working at frequencies much lower than in the observations presented here , such as lofar and the murchison widefield array ( mwa ) .
indeed , although similar improvements as for the other pathfinders can be foreseen , a simple rescaling of the computed bounds can not be adopted and a dedicated study would be in order . finally , by 2020,the ska-1 mid telescope array should be deployed and fully operational , increasing the sensitivity of its precursors by two - orders - of - magnitude . the ska-1 mid - band ( 350 -1050 mhz )
will probably be the most promising frequency range for the majority of wimp models .
the full ska-2 phase will bring another factor @xmath65 increase in sensitivity and an extended frequency range up to at least 25 ghz .
such frequency coverage increase will also contribute to constrain the dm particle mass by looking at the expected spectral cut - off at @xmath66 ( see , e.g. , ref .
@xcite ) .
more in particular , the ska expected noise can be estimated with : _
rms = k_b , [ eq : ska ] where @xmath67 is the bandwidth , @xmath68 is the observing time , @xmath69 is the boltzmann s constant , @xmath70 is the system noise temperature , and @xmath71 is the effective area . taking @xmath72 @xcite and a bandwidth of 300 mhz at ghz frequency , one obtains , from eq .
[ eq : ska ] , @xmath73 njy for 10 hours of integration time .
this is about a @xmath74 factor of gain in sensitivity with respect to the observations presented here .
a further improvement by a factor of 2 can be confidently foreseen due to the larger number of accessible dsph satellites .
the ska will also have the great advantage to be able to determine the dsph magnetic field via faraday rotation measurements ( and possibly also polarization ) , provided its strength is around the @xmath18 g level ( which is the expected level based on sfr arguments , see discussion above ) .
this will make the predictions for the expected signal much more robust and obtainable with a single experimental configuration , which can simultaneously measure both the dm - induced synchrotron emission and the magnetic field .
the prospects of detection / constraints of the wimp particle properties are shown in fig .
[ fig : dec]b for the observing scenarios outlined above , and in the ave and pes cases .
note that we can progressively close in on the full parameter space , even in the pes case , up to tev masses .
there are two main caveats in the presented forecasts .
the first stems from the fact that , for an extended emission , the confusion issue becomes stronger and stronger as one tries to probe fainter and fainter fluxes .
the source subtraction procedure becomes thus crucial and this can affect the estimated sensitivities . the extent of the impact of this effect on the actual sensitivity is hardly predictable at the present time , especially for the ska , since it will depend on the properties of the detected sources ( in an unknown flux density range ) , the efficiency of deconvolution algorithms , and the accuracy of the telescope beam shape .
the second caveat is that bringing down the observational threshold , one can start to possibly probe the very low - level of expected non - thermal emission associated to the ( very - low ) star formation in dsph .
a dm contribution should be then disentangled from such astrophysical background . on the other hand ,
the superior angular resolution of the ska will allow to precisely map such two putative emissions , and to correlate them with the stellar or dm profiles ( obtained instead via optical and kinematic measurements ) .
it is clear that a full use of dsph as dm laboratories will require a synergy between optical observations ( deep , large - area photometric searches for dsph identification and spectroscopic and astrometric follow - ups to derive structural properties ) , and multi - frequency observations of non - thermal emissions from radio to gamma - rays . as a function of the wimp mass ,
for the three scenarios opt ( blue ) , ave ( red ) , and pes ( black ) listed in table [ tableb ] , and decays into @xmath7 ( solid ) and @xmath54 ( dotted ) .
the green circle shows the best - fit to the pamela positron fraction for @xmath54 final state ( see , e.g. , ref @xcite ) .
right panel : prospects of detection for wimps annihilating into @xmath7 in the ave and pes scenarios .
see text for details about the selected experimental configurations.,scaledwidth=80.0% ] as a function of the wimp mass , for the three scenarios opt ( blue ) , ave ( red ) , and pes ( black ) listed in table [ tableb ] , and decays into @xmath7 ( solid ) and @xmath54 ( dotted ) .
the green circle shows the best - fit to the pamela positron fraction for @xmath54 final state ( see , e.g. , ref @xcite ) .
right panel : prospects of detection for wimps annihilating into @xmath7 in the ave and pes scenarios .
see text for details about the selected experimental configurations.,scaledwidth=80.0% ]
in this paper , we presented a search for a particle dm signal performed observing six dsph galaxies ( carina , fornax , sculptor , bootesii , hercules and segue2 ) at radio frequency .
a total of 123 hours of observing time with the atca telescope operating at 16 cm wavelength were employed .
the rms sensitivity of the resulting maps is below 0.05 mjy / beam for all the targets of our observations .
the synchrotron radiation from high - energy electrons and positrons originated from dm annihilations or decays is expected to produce a diffuse signal in dsph satellites of the mw on scales of few arcmin .
we have presented here the first project dedicated to the wimp search making use of radio interferometers , that could be considered as a pilot experiment for the next generation high - sensitivity and high - resolution radio telescope arrays like the ska .
a very recent campaign using instead the gbt single - dish radio telescope was presented in @xcite , and similar studies can be conducted with the effelsberg 100-m single - dish telescope which has the capability of also detecting extremely weak polarised emission . for the particle dm search we are interested in , the use of multiple array detectors having synthesized beams of arcmin size has a number of advantages with respect to single - dish observations .
first , the large collecting area can allow to increase the sensitivity .
the best beam choice for the detection of a diffuse emission requires a large synthesized beam ( in order to maximize the integrated flux ) , but still smaller than the source itself to be able to resolve it .
a good angular resolution is also crucial in order to distinguish between a possible non - thermal astrophysical emission and the dm - induced signal , which clearly becomes very hard if the dsph is not well - resolved .
the possibility of simultaneously detecting small scale sources with the long - baselines of the array allows to overcome the confusion limit . in the case of arcmin beams
, the confusion level can be easily reached with observations lasting for few tens of minutes , even by current telescopes .
a source subtraction is thus a mandatory and crucial step of the analysis .
finally , single dish telescopes face the additional complication related to galactic foreground contaminations , which are instead subdominant for the angular scales typically probed by telescope arrays at ghz frequency .
note that similar arguments apply also for the comparison of capabilities of gamma - ray telescopes versus radio interferometers for what concerns the wimp search in dsphs .
no evidence for an extended emission over a size of few arcmin has been detected with current observations ( see also discussion in paper ii ) .
we derived bounds on the wimp annihilation / decay rate as a function of the mass for different final states of annihilation / decay .
they are comparable to the best limits obtained with gamma - ray observations and are much more constraining than what obtained in the x - ray band or with previous radio observations . in section [ sec : mod ] , we described how to model the expected wimp signal , discussing the involved uncertainties .
they are mostly given by the shape of the dm profile and the dsph magnetic properties .
dsphs are poorly known systems and this reflects into a large uncertainty in the predicted signal .
we define an optimistic and a pessimistic scenarios to bracket such uncertainty .
the associated bounds on the annihilation rate varies by nearly 7 orders of magnitude in the annihilating dm case ( while the variation of the decay rate is more limited in the case of decaying dm ) . on the other hand , even for the pessimistic scenario , constraints are not dramatically far from the `` thermal '' annihilation rate , and forthcoming observations can be able to cover the full wimp parameter space .
for an average scenario , where we derive a magnetic field from the sfr of the dsph ( averaged over its history ) , and assume a mw - like spatial diffusion and an nfw profile for the dm distribution , the bounds on the annihilation rate are around the `` thermal '' value for @xmath10 gev and leptonic channels , while a factor of 10 above it for hadronic or gauge - boson final states .
as mentioned , these limits are significantly more constraining the bounds from ic in the x - ray band , and comparable to the fermi - lat constraints from dm - induced prompt emission of gamma - rays .
we also investigate the possibility that point - sources detected in the proximity of the dsph optical center might be associated to the emission from a dm cuspy profile .
this possibility is however likely only in the loss at injection scenario while spatial diffusion should in any case flatten the @xmath1 distribution , making the source extended rather than point - like .
we found no source spatially coincident with the dsph center . on the other hand
, we have 7 viable candidates with small displacements ( around 1 or 2 arcmin ) and with no solid identification at other wavelengths .
these sources can deserve further investigation since we found that the wimp scenario can fit the point - like emission with annihilation rates consistent with existing bounds . to conclude
, we demonstrate that radio interferometric observations are a suitable strategy to search for a wimp - induced diffuse emission in dsphs .
the ska and its precursor will be able to progressively probe a signal from wimp scenarios with `` thermal '' annihilation rate and masses up to few tev , irrespective of astrophysical assumptions .
we thank piero ullio for insightful discussions during the early stages of the project and rainer beck for useful comments .
s.c . acknowledges support by the south african research chairs initiative of the department of science and technology and national research foundation and by the square kilometre array ( ska ) .
is partly supported by the us department of energy , contract de - fg02 - 04er41268 .
m.r . acknowledges support by the research grant _ tasp ( theoretical astroparticle physics ) _ funded by the istituto nazionale di fisica nucleare ( infn ) . | we performed a deep search for radio synchrotron emissions induced by weakly interacting massive particles ( wimps ) annihilation or decay in six dwarf spheroidal ( dsph ) galaxies of the local group .
observations were conducted with the australia telescope compact array ( atca ) at 16 cm wavelength , with an rms sensitivity better than 0.05 mjy / beam in each field . in this work ,
we first discuss the uncertainties associated with the modeling of the expected signal , such as the shape of the dark matter ( dm ) profile and the dsph magnetic properties .
we then investigate the possibility that point - sources detected in the proximity of the dsph optical center might be due to the emission from a dm cuspy profile .
no evidence for an extended emission over a size of few arcmin ( which is the dm halo size ) has been detected .
we present the associated bounds on the wimp parameter space for different annihilation / decay final states and for different astrophysical assumptions . if the confinement of electrons and positrons in the dsph is such that the majority of their power is radiated within the dsph region , we obtain constraints on the wimp annihilation rate which are well below the thermal value for masses up to few tev . on the other hand
, for conservative assumptions on the dsph magnetic properties , the bounds can be dramatically relaxed .
we show however that , within the next 10 years and regardless of the astrophysical assumptions , it will be possible to progressively close in on the full parameter space of wimps by searching for radio signals in dsphs with ska and its precursors . |
[ cja ] in previous works , we have systematically studied the bernstein problem for complete minimal submanifolds of higher codimension in euclidean space ( see @xcite ) . in particular
, we could prove that a complete minimal submanifold in euclidean space is affine linear if it does not deviate too much from a linear subspace in the sense that a certain function @xmath0 defined in terms of jordan angles is bounded by 3 .
it is natural to ask whether that number is optimal .
now , there is the lawson - osserman s counterexample @xcite to the higher codimension bernstein problem for which @xmath0 is identically 9 .
the aim of the present paper then is to understand this example in geometric terms , in particular in terms of jordan angles . here , the jordan angles between two linear subspaces @xmath1 and @xmath2 are the critical values of the angle @xmath3 between the nonzero vectors @xmath4 in @xmath1 and their orthogonal projection @xmath5 in @xmath2 .
when these jordan angles are constant for all the normal spaces of some submanifold @xmath6 of euclidean space and a fixed linear reference subspace , we say that @xmath6 has constant jordan angles .
this is the fundamental concept of our paper , and we abbreviate it as cja . for a precise statement , refer to definition [ def - cja ] below .
now it turns out the lawson - osserman s counterexample has cja relative to the imaginary quaternions when viewed as a subspace of the imaginary octonians .
harvey - lawson @xcite showed that the lawson - osserman s cone is a four dimensional coassociative submanifold in @xmath7 which can be identified with the imaginary octonians .
therefore , we study such coassociative submanifolds with cja and find that a coassociative graph with cja relative to the imaginary quaternions and at most two different normal jordan angles either is affine linear or a translate of a portion of the lawson - osserman s cone . for more precise statements ,
we now develop some notation and technical concepts .
let @xmath1 and @xmath8 be @xmath9-dimensional subspaces ( i.e. @xmath9-planes ) in @xmath10 .
the _ jordan angles _ between @xmath1 and @xmath8 are the critical values of the angle @xmath3 between a nonzero vector @xmath4 in @xmath1 and its orthogonal projection @xmath5 in @xmath8 as @xmath4 runs through @xmath1 .
this concept was firstly introduced by jordan @xcite in 1875 , and they are also called _ principal angles _ in some references , e.g. @xcite . if @xmath3 is a nonzero jordan angle between @xmath1 and @xmath8 determined by a unit vector @xmath4 in @xmath1 and its projection @xmath5 in @xmath8
, then @xmath4 is called an _ angle direction _ of @xmath1 relative to @xmath8 , and the @xmath11-plane spanned by @xmath4 and @xmath5 is called an _ angle 2-plane _ between @xmath1 and @xmath8 ( see @xcite ) .
denote by @xmath12 the orthogonal projection of @xmath10 onto @xmath8 and by @xmath13 the orthogonal projection of @xmath10 onto @xmath1 .
then for any @xmath14 and @xmath15 , @xmath16 and moreover @xmath17 holds for every @xmath18 , which implies @xmath19 is a nonnegative definite self - adjoint transformation on @xmath1 . for any nonzero vector @xmath14
, @xmath20 hence @xmath3 is a jordan angle between @xmath1 and @xmath8 if and only if @xmath21 is an eigenvalue of @xmath19 , and @xmath4 is an angle direction with respect to @xmath3 if and only if @xmath4 is an eigenvector associated to the eigenvalue @xmath22 , i.e. @xmath23 therefore , all the angle directions with respect to @xmath3 constitute a linear subspace of @xmath1 , which is called an _ angle space _ of @xmath1 relative to @xmath8 and we denote it by @xmath24 . in particular , @xmath25 the dimension of @xmath24 is called the _ multiplicity _ of @xmath3 , which is denoted by @xmath26 . if we denote by @xmath27 the set consisting of all the jordan angles between @xmath1 and @xmath8 , then @xmath28 and the angle spaces are mutually orthogonal to each other .
hence @xmath29 the jordan angles between two @xmath9-planes completely determine their relative positions .
more precisely , one can conclude that : @xcite[position ] let @xmath30 and @xmath31 be any two pairs of @xmath9-planes in @xmath10 .
if @xmath32 and the multiplicities of the corresponding jordan angles are equivalent , then there exists a rigid motion of @xmath10 , carrying @xmath33 onto @xmath31 , respectively . and vice versa .
similarly , let @xmath34 denote the set consisting of all the jordan angles between @xmath8 and @xmath1 , then @xmath35 if and only if @xmath21 is an eigenvalue of @xmath36 .
denote by @xmath37 the angle space of @xmath8 relative to @xmath1 associated to @xmath3 , then @xmath38 if and only if @xmath39 , and @xmath40 let @xmath41 and @xmath42 be the orthogonal complements of @xmath1 and @xmath8 , and denote by @xmath43 and @xmath44 the orthogonal projections of @xmath10 onto @xmath41 and @xmath42 , respectively . as above , the set consisting of all the jordan angles between @xmath41 and @xmath42
is denoted by @xmath45 , @xmath46 denotes the angle space associated to @xmath47 , and @xmath48 denotes the multiplicity of @xmath3 .
the following lemma reveals the close relationship between @xmath27 , @xmath49 and @xmath45 .
[ jordan](@xcite ) let @xmath50 be @xmath9-planes in @xmath10 , then @xmath51 and the multiplicities of each corresponding jordan angles are equivalent . if we denote @xmath52 for each @xmath53 , then @xmath54 whenever @xmath55 , and @xmath56 for any @xmath57 $ ] , @xmath58 if and only if @xmath53 , and @xmath59 , @xmath60 .
moreover , for every @xmath61 , there exists an isometric automorphism @xmath62 , such that \(i ) @xmath63 , @xmath64 ; \(ii ) @xmath65 ; \(iii ) for any nonzero vector @xmath66 ( @xmath67 ) , @xmath68 ( @xmath69 ) lies in the angle 2-plane generated by @xmath4 ( @xmath0 ) ; more precisely , @xmath70 * remark .
* let @xmath1 and @xmath8 be a pair of intersecting planes in @xmath71 , then @xmath72 , where @xmath3 is the dihedral angle between @xmath1 and @xmath8 .
@xmath73{cja.jpg}\ ] ] denote by @xmath74 their line of intersection and @xmath75 the origin of @xmath71 .
choose @xmath76 , such that @xmath77 is a unit vector orthogonal to @xmath1 .
through @xmath78 , draw a perpendicular line to @xmath8 , intersecting @xmath8 at @xmath79 , @xmath1 at @xmath80 .
denote @xmath81 , then @xmath82 , @xmath83 and @xmath84 .
denote @xmath85}m_\th=\sum_{\th\in \text{arg}(p^\bot , q_0^\bot)\cap ( 0,\pi/2]}m_\th^\bot\ ] ] then @xmath86 if and only if @xmath87 , and @xmath88 .
similarly @xmath89 if and only if @xmath90 , and @xmath91 .
let @xmath6 be an @xmath92-dimensional submanifold in @xmath10 and @xmath8 be a fixed @xmath9-plane in @xmath10 .
denote by @xmath93 and @xmath94 the tangent bundle and the normal bundle along @xmath6 , respectively . for
any @xmath95 , denote by @xmath96 ( @xmath97 ) the set consisting of all the jordan angles between @xmath98 ( @xmath99 ) and @xmath8 ( @xmath42 ) , which are called _ normal ( tangent ) jordan angles _ at @xmath100 .
let @xmath3 be a @xmath101$]-valued smooth function on @xmath6 , if @xmath102 ( @xmath103 ) , we say @xmath3 is a _
normal ( tangent ) jordan angle function _ of @xmath6 relative to @xmath8 .
denote by @xmath104 ( @xmath105 ) the set consisting of all the normal ( tangent ) jordan angle functions of @xmath6 relative to @xmath8 .
if @xmath3 is a smooth function on @xmath6 that is nonzero everywhere , then lemma [ jordan ] implies @xmath106 if and only if @xmath107 .
denote @xmath108 let @xmath12 and @xmath44 be orthogonal projections onto @xmath8 and @xmath42 , @xmath109 and @xmath110 denote orthogonal projections onto @xmath99 and @xmath98 , respectively . then @xmath111 if and only if @xmath112 and similarly @xmath113 if and only if @xmath114 let @xmath115 , @xmath116 for every @xmath95 , then @xmath117 and @xmath118 are both @xmath119-valued functions on @xmath6 . based on @xcite ,
one can easily deduced that [ dis](@xcite ) let @xmath3 be a normal ( tangent ) jordan angle function of @xmath6 relative to @xmath8 .
if @xmath117 ( @xmath118 ) is a constant function on @xmath6 , then @xmath120 ( @xmath121 ) is a smooth subbundle of @xmath94 ( @xmath93 ) . in this case , @xmath120 ( @xmath121 ) is said to be a _
normal ( tangent ) angle space distribution _ associated to @xmath3 .
a curve @xmath122 , all of whose tangent vectors belongs to a tangent angle space distribution , i.e. @xmath123 for every @xmath124 , is called an _ angle line _ of @xmath6 .
more generally , an _ angle surface _ is a connected submanifold @xmath125 of @xmath6 , such that for any @xmath126 , @xmath127 .
now we can formulate the definition of _ submanifolds with constant jordan angles ( cja ) _ , the main subject of this paper .
[ def - cja ] let @xmath6 be an @xmath92-dimensional submanifold of @xmath10 and @xmath8 be a fixed @xmath9-plane .
if every normal jordan angle function of @xmath6 relative to @xmath8 is a constant function , and @xmath117 is constant on @xmath6 for each @xmath106 , then we say @xmath6 has * constant jordan angles ( cja ) * relative to @xmath8 . with the aid of lemma [ jordan ] and proposition [ position ] , one can obtain equivalent definitions of submanifolds with cja . for any @xmath92-dimensional submanifold @xmath6 of @xmath10 and a fixed @xmath9-plane @xmath8 , the following statement
are equivalent : \(i ) @xmath6
has cja relative to @xmath8 ; \(ii ) every tangent angle function of @xmath6 relative to @xmath42 is constant , and @xmath118 is constant ; \(iii ) @xmath96 ( @xmath97 ) is independent of @xmath95 , and the multiplicity of each normal ( tangent ) jordan angle is constant ; \(iv ) the relative position of @xmath98 ( @xmath99 ) and @xmath8 ( @xmath42 ) is independent of @xmath95 .
* remarks : * * let @xmath128 be an arc - length parameterized curve in @xmath71 . if @xmath128 is a _ constant angle curve _ , i.e. the unit tangent vector at every point makes a constant angle with a fixed straight line in @xmath71 , then @xmath128 is a helix , and vise versa .
let @xmath125 be a smooth surface in @xmath71 , if the normal vector at every point makes a constant angle with a fixed straight line in @xmath71 , then @xmath125 is said to be a _ constant angle surface _ in @xmath71 .
a surface @xmath125 in @xmath71 is a constant angle surface if and only if it is locally isometric to either a cylinder , a right circular cone , or the tangential developable of a helix . moreover ,
if we additionally assume @xmath125 to be complete , then @xmath125 has to be a cylinder .
recently , many geometers are interested in constant angle surfaces in other ambient spaces , e.g. @xmath129 @xcite , @xmath130 @xcite , heisenberg group @xcite , minkowski space @xcite and product spaces @xcite .
our notion is a natural generalization of the classical constant angle curves and surfaces . * if @xmath131 is a hypersurface of @xmath132 , then @xmath6 has cja if and only if @xmath6 is a helix hypersurface @xcite .
hence the concept of submanifolds with cja is a natural generalization of helix hypersurfaces to higher codimensional cases .
helix hypersurfaces are closed related to the shadow problem ( see @xcite ) formulated by h. wente , and another interesting motivation for the study of helix hypersurfaces comes from the physics of interfaces of liquid crystal ( see @xcite ) .
* let @xmath125 be a surface in @xmath133 , then @xmath125 has cja if and only if @xmath125 is a surface in @xmath133 with _ constant principal angles with respect to a plane_. this concept was introduced by bayard - di scala - castro - hernndez in @xcite . in this paper , the authors established a local existence theorem and classified all the complete surfaces in @xmath133 with constant principal angles
. denote @xmath134 then @xmath135 is a constant @xmath119-valued function on @xmath6 . as shown above ,
@xmath136 ( @xmath137 ) if and only if @xmath138 ( @xmath139 ) , and the multiplicity of @xmath140 equals @xmath141 ( @xmath142 ) .
let @xmath143 be the numbers of distinct jordan angles .
note that @xmath144 whenever @xmath145 , @xmath146 whenever @xmath147 , and otherwise @xmath148 .
in conjunction with lemma [ jordan ] and lemma [ dis ] , @xmath94 and @xmath93 have the following vector bundle decompositions @xmath149 in particular , if @xmath150 , then there exists a smooth mapping @xmath151 , where @xmath152 such that : ( i ) @xmath153 keeps each fiber invariant ; ( ii ) the length of each vector in @xmath154 is invariant under @xmath153 ; ( iii ) @xmath155 ; ( iv ) @xmath156 , @xmath157 ; ( iv ) for any @xmath158 and @xmath159 , @xmath160 @xmath153 is called the _ anti - involution _ associated to @xmath3 .
the concept of cja submanifolds that we have just introduced arises from our systematic investigation of the bernstein problem in higher codimension .
we now wish to explain this connection . the classical bernstein theorem @xcite states that any entire minimal graph in @xmath71 has to be affine linear .
this result has been extended by j. simons @xcite to such entire minimal graphs in @xmath132 for @xmath161 , whereas bombieri - de giorgi - giusti @xcite constructed counterexamples in higher dimensions .
but for any dimensions , there is a weak version of the bernstein type theorem , obtained by j. moser @xcite who proved that any entire solution @xmath162 to the minimal surface equation @xmath163 has to be affine linear , provided that @xmath164 is a bounded function .
@xmath0 is a significant quantity here for various reasons .
firstly , the boundedness of @xmath0 ensures that ( [ min - eq ] ) is a uniformly elliptic equation , so that a bernstein type result can be obtained by moser s iteration .
secondly , for any @xmath162 , @xmath165 is a global coordinate chart of the graph of @xmath166 , and a straightforward calculation shows that the volume form of @xmath167 is @xmath168 , i.e. @xmath0 equals the radio of the volume form of @xmath167 and the coordinate plane .
thirdly , @xmath0 has a close relationship with jordan angles .
a direct computation shows @xmath169 is a unit normal vector field on @xmath167 .
thus the angle between @xmath170 and the @xmath171-axis is @xmath172 , which is smaller than an acute angle whenever the @xmath0-function is bounded .
therefore , moser s theorem can be restated as : let @xmath6 be a complete minimal hypersurface in @xmath132 and @xmath173 . if the angle between the normal vector and @xmath171-axis is smaller than @xmath174 everywhere , then @xmath6 has to be an affine @xmath92-plane .
now we consider an @xmath92-dimensional entire minimal graph @xmath6 in @xmath10 , generated by a smooth vector - valued function @xmath175 @xmath176 then @xmath166 satisfies the minimal surface equations @xmath177 here @xmath178 is the induced metric on @xmath6 , @xmath179 denotes the inverse matrix of @xmath180 , and @xmath181 is the volume form of @xmath6 .
more precisely , @xmath182^{1/2}.\ ] ] similarly to the case of codimension 1 , the @xmath0-function has a close relationship with jordan angles . at any point @xmath95 ,
denote by @xmath183 the jordan angles between @xmath98 and the coordinate @xmath9-plane , then a calculation shows ( see @xcite@xcite ) @xmath184 we note that @xmath185 is the inner product of the normal @xmath9-plane and the coordinate @xmath9-plane . here
all the @xmath9-planes are viewed as vectors in a euclidean space of larger dimension , via plcker embedding ( see @xcite ) .
it is natural to ask whether moser s theorem can be generalized to the higher codimensional case . in other words ,
given an entire minimal graph @xmath186 with @xmath175 , does the boundedness of the @xmath0-function ensure that @xmath6 has to be an affine @xmath92-plane ?
the answer is yes for the cases of dimension 2 @xcite@xcite and dimension 3 @xcite@xcite , but it is no for dimension 4 , according to the works of lawson - osserman @xcite and harvey - lawson @xcite . let us explain the geometric reason for this fact .
let @xmath187 and @xmath188 denote the octonions and quaternions , respectively .
we have @xmath189 , with @xmath190 a unit element orthogonal to @xmath188 , and for any @xmath191 , @xmath192 denote @xmath193 .
assume @xmath194 is a fixed unit element , then @xmath195:q\in \text{sp}_1,r\in \bbb{r}^+\}\ ] ] is a @xmath196-dimensional cone in @xmath197 , which is the graph of the function @xmath198 @xmath199 here @xmath200 and @xmath201 .
note that @xmath202 is a _ cone - like function _
, i.e. @xmath203 for any @xmath204 and @xmath205 .
it was discovered by lawson - osserman @xcite that @xmath202 is a lipschitz solution to the non - parametric minimal surface equations that is not @xmath206 , and a straightforward calculation shows the @xmath0-function is always @xmath207 on @xmath208 .
afterwards , harvey - lawson @xcite constructed a family of @xmath196-dimensional entire minimal graphs in @xmath197 ; the tangent cone at infinity of each one is just the lawson - osserman s cone , and the @xmath0-function takes value in @xmath209 .
therefore , moser s theorem can not generalize to all higher codimensional cases .
now we further explore the geometric properties of lawson - osserman s cone via jordan angles .
[ lo - cone ] lawson - osserman s cone @xmath208 is a @xmath196-dimensional submanifold in @xmath210 with cja relative to @xmath211 , and @xmath212 , @xmath213 .
* proposition [ lo - cone ] was firstly proved in the appendix of @xcite , and the calculation was based on the complex form of the hopf map from @xmath214 to @xmath215 .
now , we shall give another proof , which is based on the fact that @xmath208 is a @xmath216-invariant manifold and has a close relationship with the argument in section [ co - cja ] .
denote @xmath217 @xmath218.\ ] ] let @xmath219 be an arbitrary point in @xmath208 .
we shall compute the jordan angles between @xmath220 and @xmath221 .
let @xmath222 be the lie algebra associated to @xmath216 , which can be seen as the linear space constisting of right - invariant vector fields on @xmath216 .
it is well - known that @xmath222 is isomorphic to @xmath223 , and the isomorphism is given by @xmath224 @xmath225 as a matter of convenience , @xmath226 and @xmath227 are regarded to be same in the sequel .
then at @xmath228 , @xmath229 with @xmath230 and @xmath231\\ & = \f{d}{dt}\big|_{t=0}r_0\big[(\sqrt{5}/2)e^{tb}(q_0 a\bar{q}_0)e^{-tb}+(\bar{q}_0 e^{-tb})e\big]\\ & = \f{d}{dt}\big|_{t=0}r_0\big[(\sqrt{5}/2)e^{tb}a_1e^{-tb}+e^{-tb}\ep\big]\\ & = r_0\big[(\sqrt{5}/2)(ba_1-a_1 b)-b\ep\big ] . \endaligned\ ] ] let @xmath232 be a unit vector in @xmath211 that is orthogonal to @xmath233 and denote @xmath234 . then @xmath235 is an orthonormal basis of @xmath211 , satisfying @xmath236 , @xmath237 , @xmath238 and @xmath239 , then @xmath240 denote @xmath241 then @xmath242 is an orthonormal basis of @xmath220 .
let @xmath12 , @xmath44 be the orthogonal projections of @xmath243 into @xmath211 and @xmath221 , respectively , then @xmath244 which implies @xmath245 and hence @xmath246 is a tangent angle direction associated to @xmath247 .
note that @xmath246 is the direction of the ray going through @xmath228 .
similarly , one can prove that @xmath248 are both tangent angle directions associated to @xmath249 , and @xmath250 is a tangent angle direction associated to @xmath140 .
since @xmath228 can be taken arbitrarily , @xmath208 has cja relative to @xmath211 , and @xmath251 , @xmath252 .
moreover , an arbitrary angle line with respect to @xmath253 is a ray of @xmath208 , and vise versa . in @xcite ,
lawson - osserman raised the following question : what is the largest constant @xmath80 such that an entire minimal graph of arbitrary dimension and codimension with @xmath254 has to be affine linear ?
up to now , the best positive answer to this question in a successive series of achievements by several mathematicians ( see @xcite , @xcite , @xcite , @xcite ) is gotten in @xcite , which says that for any entire minimal graph @xmath255 with @xmath175 , if @xmath256 , then @xmath6 has to be an affine @xmath92-plane .
but there is still a large quantitative gap between 3 and 9 , that is , between known bernstein type theorems and the counterexamples .
lawson - osserman s problem can be viewed as the first gap problem of the @xmath0-function for entire minimal graphs of higher codimension .
to study the gap phenomena of the @xmath0-function , it is natural to consider minimal graphs whose @xmath0-function is constant . observing that the @xmath0-function is a function of all jordan angle functions ( see ( [ w ] ) ) , the @xmath0-function on any minimal graph with cja relative to the coordinate plane is constant .
proposition [ lo - cone ] shows the lawson - osserman s cone @xmath208 has cja relative to the imaginary quaternions , but unfortunately it is not a complete submanifold .
so one can propose the following problems : do there exist nonflat entire minimal graphs whose @xmath0-function is constant ?
[ problem1 ] let @xmath257 and @xmath258 be sets consisting of some real numbers strictly bigger than 1 .
@xmath259 if and only if there exists a nonlinear cone - like map @xmath260 , such that @xmath261 is a minimal graph whose @xmath0-function always equals @xmath262 . similarly , @xmath263 if and only if there exists a nonlinear cone - like map @xmath260 , such that @xmath261 is a minimal graph with cja relative to @xmath264 . are @xmath257 and @xmath258 discrete sets ? [ problem2 ]
let @xmath265 and @xmath266 be sets consisting of some real numbers strictly bigger than 1 .
@xmath267 if and only if there exists a nonlinear vector - valued function @xmath268 ( @xmath269 is an open domain ) , such that @xmath261 is a minimal graph whose @xmath0-function always equals @xmath262 .
similarly , @xmath270 if and only if there exists a nonlinear vector - valued function @xmath268 , such that @xmath271 is a minimal graph with cja relative to @xmath264 . are @xmath265 and @xmath266 discrete sets ?
does any minimal graph in euclidean space with constant @xmath0-function have to be a submanifold with cja ?
obviously @xmath272 , @xmath273 , @xmath274 , @xmath275 and problem [ problem2 ] can be viewed as a local version of problem [ problem1 ] . for problem [ problem1 ] ,
the known facts include @xmath276\notin s_v$ ] ( see @xcite ) and @xmath277 .
problem [ problem1 ] is quite similar to chern s conjecture , intrinsic rigidity problem in the theory of minimal submanifolds , which claims that if the squared length of the second fundamental form ( denoted by @xmath278 ) of a compact minimal submanifold in the unit euclidean sphere is constant , then the value should be contained in a discrete set ( see @xcite ) . if @xmath6 is an @xmath92-dimensional cone in @xmath10 , then the intersection of @xmath6 and the unit sphere gives an @xmath279-dimensional submanifold @xmath280 in @xmath281 .
@xmath6 is said to be the cone generated by @xmath280 , i.e. @xmath282 . as pointed out by j. simons @xcite , the geometric properties of @xmath280 are closed related to those of the cone @xmath283 .
firstly , @xmath283 has parallel mean curvature in @xmath10 if and only if @xmath280 is a minimal submanifold in @xmath281 ( see @xcite p.64 ) . noting that @xmath283 is a linear subspace if and only if @xmath280 is a totally geodesic subsphere , the bernstein problem for minimal submanifolds in euclidean space can be transferred to the spherical bernstein problem for minimal submanifolds in the sphere , in the framework of the geometric measure theory ( see @xcite , @xcite ) .
for any @xmath284 , denote by @xmath285 and @xmath286 the tangent @xmath279-plane and the normal @xmath9-plane of @xmath280 at @xmath100 , respectively , then @xmath287 along the ray going through @xmath100 , the tangent @xmath92-planes and the normal @xmath9-planes of @xmath283 are both constant , and @xmath288 here @xmath289 denotes the position vector of @xmath100 in @xmath10 .
let @xmath8 be a fixed @xmath9-plane in @xmath10 , if @xmath290 is independent of @xmath284 , and the multiplicity of each normal jordan angle is constant , then we say @xmath280 is a * submanifold in a sphere with constant jordan angles ( cja ) * relative to @xmath8 . by ( [ sphere ] )
, @xmath280 has cja if and only if the cone @xmath283 generated by @xmath280 is a submanifold in @xmath10 with cja .
thereby , problem [ problem2 ] can be restated as follows : let @xmath291 and @xmath292 be sets consisting of some real numbers taking values in @xmath293 .
@xmath294 if and only if there exists an @xmath279-dimensional compact minimal submanifold @xmath280 in @xmath281 , that is non - totally geodesic , such that its @xmath295-function always equals @xmath296 , i.e. the inner product of each normal @xmath9-plane and a fixed @xmath9-plane @xmath8 is @xmath296 .
similarly , @xmath297 if and only if there exists an @xmath279-dimensional compact minimal and non - totally geodesic submanifold @xmath280 in @xmath281 , which has cja relative to a fixed @xmath9-plane , such that its @xmath295-function always equals @xmath296 . are @xmath291 and @xmath292 discrete ? * remark . * due to ( [ w ] ) , @xmath298 there is a long way to resolving these problems . in this paper
, we only consider cja submanifolds with a small number of distinct jordan angles ( i.e. @xmath299 and @xmath300 ) .
this paper will be organized as follows . in section [ s4 ] ,
the second fundamental form @xmath79 of submanifolds with cja in euclidean space shall be studied . at first , differentiating the jordan angle functions not only gives some nullity properties of @xmath79 , but also reveals the relationship between the induced tangent ( normal ) connection and the second fundamental form .
taking the covariant derivative of the formulas obtained in the previous step , one can compute some components of @xmath301 in terms of @xmath79 . with the aid of the codazzi equations
, we can derive a constraint equation for the second fundamental form ( see lemma [ sec30 ] ) , which is nontrivial when the multiplicity of a tangent jordan angle function @xmath302 , i.e. @xmath118 , is strictly larger than 1 .
this conclusion will play an important part in section [ co - cja ] . based on these formulas , it is easy to get some vanishing theorems for the second fundamental form @xmath79 of submanfolds with cja , including the following one .
[ cja1 ] let @xmath166 be an @xmath264-valued function on an open domain @xmath303 .
if @xmath271 is a minimal submanifold with cja relative to @xmath264 , and @xmath304 , then @xmath166 has to be affine linear , i.e. @xmath6 has to be an affine @xmath92-plane .
note that the example of lawson - osserman s cone implies that the condition @xmath304 in theorem [ cja1 ] can not be omitted .
in @xcite , harvey - lawson introduced a new concept of coassociative submanifolds , as an important example of calibrated geometries , and showed that lawson - osserman s cone is a coassociative submanifold . observing that coassociative submanifolds constitute an important class of 4-dimensional minimal submanifolds in @xmath7 , it is natural to study the structure of coassociative submanifolds with cja , which is the main topic of section [ co - cja ] . with the aid of the algebraic properties of octonions
, one can obtain several interesting conclusions on the jordan angles and the second fundamental form of coassociative submanifolds . in conjunction with lemma [ sec30 ] ,
a structure theorem for coassociative submanifolds with cja is deduced as follows .
let @xmath166 be a smooth function from an open domain @xmath305 into @xmath211 .
if @xmath271 is a coassociative submanifold with cja relative to @xmath211 , and @xmath306 , then @xmath166 is either an affine linear function or @xmath307 , where @xmath308 , @xmath309 and @xmath310 with @xmath311 an arbitrary unit element in @xmath211 . in other words
, @xmath6 is an affine 4-plane or a translate of an open subset of the lawson - osserman s cone .
[ s4 ] let @xmath6 be an @xmath92-dimensional submanifold in @xmath10 with cja relative to a fixed @xmath9-plane @xmath8 .
we use the notations @xmath12 , @xmath44 , @xmath312 , @xmath313 , @xmath120 , @xmath121 , @xmath154 , @xmath117 , @xmath118 , @xmath299 , @xmath300 established in section [ cja ] . for @xmath100 in @xmath6
, we put @xmath314 the second fundamental form @xmath79 is a pointwise symmetric bilinear form on @xmath99 ( @xmath95 ) with values in @xmath98 defined by @xmath315 with @xmath316 the levi - civita connection on @xmath10 .
the induced connections on @xmath93 and @xmath94 are @xmath317 here @xmath318 are smooth sections of @xmath93 and @xmath170 denotes a smooth section of @xmath94 .
the second fundamental form , the curvature tensor of the submanifold , the curvature tensor of the normal bundle and the curvature tensor of the ambient manifold satisfy the gauss , codazzi and ricci equations ( see @xcite for details ) .
let @xmath78 be the shape operator defined by @xmath319 @xmath320 is a symmetric operator on @xmath99 and satisfies the weingarten equations @xmath321 the trace of the second fundamental form gives a normal vector field @xmath322 on @xmath6 , which is called the mean curvature vector field .
if @xmath323 , then we say that @xmath6 has parallel mean curvature . moreover if @xmath324 , @xmath6 is called a minimal submanifold
let @xmath325(@xmath326 ) and @xmath151 denote the anti - involution associated to @xmath3 , then ( [ phi2 ] ) gives @xmath327 for any @xmath328 and @xmath329 for any @xmath330 .
in other words , @xmath331 based on the above formulas , one can easily deduce the following nullity lemmas for the second fundamental form of @xmath6 .
[ sec2 ] for each @xmath325 which takes values in @xmath332 , @xmath333 holds pointwisely for any @xmath334 and @xmath335 .
in particular , @xmath336 for every @xmath337 . by linearity , it suffices to prove ( [ sec1 ] ) for any unit vector @xmath337 .
let @xmath338 be a smooth local section of @xmath121 , such that @xmath339 and @xmath340 , then @xmath341 differentiating both sides with respect to @xmath4 yields @xmath342 ( where we have used ( [ ang1 ] ) ) and then we arrive at ( [ sec1 ] ) .
[ sec6 ] for each @xmath325 taking values in @xmath332 , @xmath343 for any @xmath344 and @xmath345 .
let @xmath346 , then @xmath347 and @xmath348 .
applying lemma [ sec2 ] gives @xmath349 and ( [ sec3 ] ) immediately follows from the above equation .
[ sec25 ] if @xmath350 and @xmath351 or @xmath352 , then @xmath353 for any @xmath334 , @xmath337 and @xmath345 .
if @xmath351 , let @xmath338 be a smooth local section of @xmath121 such that @xmath339 , then @xmath354 for any @xmath355 .
thus @xmath356 . on the other hand
, @xmath345 implies @xmath357 , hence @xmath358 the proof for @xmath359 is similar .
let @xmath360 , @xmath55 , @xmath338 a local section of @xmath93 , @xmath361 and @xmath362 local sections of @xmath121 and @xmath363 , respectively .
define @xmath364 then for any smooth function @xmath166 defined on @xmath6 , @xmath365 , @xmath366 and @xmath367 this means @xmath368 is a smooth tensor field on @xmath6 of type @xmath369 .
more precisely , @xmath368 is a smooth section of the tensor bundle @xmath370 . since @xmath371 is a levi - civita connection on @xmath6 , @xmath372
now we additionally define @xmath373 then ( [ ang1 ] ) still holds when @xmath374 or @xmath352 .
let @xmath375 be a constant depending only on @xmath3 and @xmath376 .
the following result reveals the relationship between @xmath368 and the second fundamental form .
[ sec5 ] let @xmath360 , @xmath55 , then for any @xmath334 , @xmath337 and @xmath377 , @xmath378 let @xmath379 be smooth local sections of @xmath121 and @xmath363 , respectively , such that @xmath380 , @xmath381 , then @xmath382 , @xmath383
. hence @xmath384 differentiating both sides of the above equation with respect to @xmath334 yields @xmath385 ( we have used ( [ ang1 ] ) and ( [ s ] ) ) , which is equivalent to ( [ sec4 ] ) .
similarly , given @xmath334 , @xmath386 , @xmath387 with @xmath388 and @xmath55 , one can define @xmath389 then @xmath390 is a smooth section of @xmath391 , and @xmath392 let @xmath393 be local section of @xmath120 and @xmath394 respectively , then @xmath395 differentiating both sides of the above equality with respect to @xmath334 , one can use ( [ ang1 ] ) to get the following result , as in the proof of lemma [ sec5 ] .
[ sec7 ] given @xmath388 , @xmath55 , @xmath396 for any @xmath397 and @xmath398 .
let @xmath399 , @xmath400 , and @xmath401 be the orthogonal projection of @xmath98 onto @xmath402 .
define @xmath403 for any @xmath404 .
then @xmath405 is a smooth section of the tensor bundle @xmath406 .
obviously @xmath407 , and @xmath408 hence @xmath405 is a curvature type tensor .
note that @xmath409 whenever @xmath410 .
let @xmath360 , and define @xmath411 for any @xmath412 . here
@xmath413 denotes the orthogonal projection of @xmath99 onto @xmath414 . due to lemma [ sec2 ] , @xmath415 for any @xmath335 , hence @xmath416 and @xmath417 whenever @xmath410 .
note , however , that @xmath418 does not satisfy a bianchi type identity .
[ sec30 ] given @xmath399 taking values in @xmath332 , @xmath419 for any @xmath335 , and moreover @xmath420 let @xmath421 for each @xmath422 , then lemma [ sec2 ] tells us @xmath423 and moreover @xmath424 in particular , combining the weingarten equations and lemma [ sec6 ] gives @xmath425 i.e. @xmath426 .
let @xmath379 be local sections of @xmath121 such that @xmath427 , @xmath428 .
by lemma [ sec6 ] , @xmath429 , hence @xmath430 where @xmath431 ( lemma [ sec6 ] , lemma [ sec2 ] , lemma [ sec7 ] and @xmath432 have been used in this calculation ) and @xmath433 ( here we have used the weingarten equations , ( [ sec8 ] ) and lemma [ sec5 ] . ) substituting ( [ sec11 ] ) and ( [ sec12 ] ) into ( [ sec10 ] ) implies @xmath434 again applying lemma [ sec6 ] gives @xmath435 , hence @xmath436 where @xmath437 and @xmath438 if @xmath439 , then @xmath440 is isometric and @xmath441 .
hence @xmath442 on the other hand , @xmath443 whenever @xmath444 or @xmath352 .
therefore @xmath445 substituting ( [ sec14])-([sec16 ] ) into ( [ sec13 ] ) yields @xmath446 the codazzi equations imply @xmath447 . hence by comparing the right hand sides of ( [ sec17 ] ) and ( [ sec18 ] ) we arrive at @xmath448 and then ( [ sec19 ] ) immediately follows from the definition of @xmath405 and @xmath418 .
finally ( [ sec9 ] ) is obtained by substituting ( [ sec24 ] ) into ( [ sec18 ] ) .
[ sec31 ] we consider @xmath399 taking values in @xmath332 and @xmath422 such that @xmath55 . if @xmath449 holds for any @xmath450 , then @xmath451 for any @xmath337 and @xmath377 . in the sequel we make use of the abbreviation @xmath452 for any @xmath453 .
by the definition of @xmath418 , @xmath454 i.e. @xmath455 .
hence @xmath456 i.e. @xmath426
. similarly , one can deduce that @xmath457 .
let @xmath361 be a local smooth section of @xmath121 and @xmath362 be a local smooth section of @xmath363 , such that @xmath427 , @xmath428 .
lemma [ sec2 ] implies @xmath458 , hence @xmath459 where @xmath460 and @xmath461 substituting ( [ sec22 ] ) and ( [ sec23 ] ) into ( [ sec21 ] ) yields ( [ sec20 ] ) .
[ sec32 ] if @xmath462 and @xmath351 or @xmath352 , then for any @xmath337 , @xmath463 and @xmath464 , @xmath465 let @xmath361 be a local section of @xmath121 and @xmath362 be a local section of @xmath93 , such that @xmath427 and @xmath428 , then lemma [ sec25 ] tells us @xmath466
. therefore @xmath467 where @xmath468 and @xmath469 here @xmath470 , and @xmath471 is a direct corollary of lemma [ sec25 ] . substituting ( [ sec27 ] ) and ( [ sec28 ] ) into ( [ sec29 ] ) , we arrive at at ( [ sec26 ] ) . with the above lemmas
, we can now derive vanishing theorems for the second fundamental form of submanifolds with cja .
[ tri ] let @xmath131 be a submanifold of @xmath10 with cja relative to a fixed @xmath9-plane @xmath8 ( @xmath6 need not be complete ) , then \(i ) if @xmath472 , then @xmath6 has to be an affine linear subspace ; \(ii ) if @xmath473 and @xmath6 has parallel mean curvature , then @xmath6 is affine linear ; \(iii ) if @xmath474 , and @xmath6 has parallel mean curvature , then @xmath6 is affine linear ; \(iv ) if @xmath475 , @xmath476 , and @xmath6 is minimal , then @xmath6 is affine linear .
* remarks : * * let @xmath477 be a circle whose tangent vectors are all orthogonal to the @xmath478-axis , then @xmath479 has cja and @xmath480 , @xmath481 .
it is easy to check that @xmath479 has parallel mean curvature .
hence the condition @xmath482 can not be dropped in ( ii ) .
* let @xmath483 be a circular cylinder , whose normal vectors are all orthogonal to the @xmath478-axis , then @xmath125 has cja and @xmath484 , @xmath485 .
its mean curvature vector field is parallel along @xmath125 .
hence the condition @xmath486 can not be dropped in ( iii ) .
* let @xmath125 be a nontrivial minimal surface in @xmath71 , then @xmath487 is a minimal submanifold in @xmath488 .
then @xmath6 has cja relative to @xmath489 , and @xmath490 .
hence the condition @xmath476 can not be dropped in ( iv ) .
\(i ) denote @xmath491 , then lemma [ sec6 ] and [ sec25 ] tell us @xmath492 for any @xmath493 and @xmath494 .
hence @xmath6 is totally geodesic .
\(ii ) as shown in section [ cja ] , there exists @xmath495 , such that @xmath496 , @xmath497 . by lemma [ sec30 ] , @xmath498 for any @xmath499 . in conjunction with @xmath500
, we have @xmath501 . substituting it into ( [ sec9 ] ) implies @xmath502 let @xmath503 be an orthonormal basis of @xmath504 . since @xmath6 has parallel mean curvature , @xmath505 which forces @xmath506 for any @xmath507 .
thus @xmath508 for any @xmath509 .
on the other hand , lemma [ sec6 ] implies @xmath510 .
therefore @xmath511 on @xmath6 .
\(iii ) denote @xmath512 , @xmath513 with @xmath495 . again applying lemma [ sec30 ]
gives @xmath514 i.e. @xmath515 for any @xmath516 .
this means @xmath517 since @xmath518 is an isomorphism , @xmath519 holds for every @xmath520 . on the other hand , @xmath521 is a direct corollary of lemma [ sec6 ] .
thus @xmath522 for every @xmath523 .
let @xmath524 be an orthonormal basis of @xmath525 , and @xmath526 be an orthonormal basis of @xmath527 .
for any @xmath528 , by ( [ sec9 ] ) and ( [ sec20 ] ) , @xmath529 hence @xmath530 and then @xmath531 for any @xmath520 . on the other hand
, @xmath532 holds for any @xmath528 .
therefore @xmath533 for each @xmath534 . in summary , @xmath535
for any smooth section @xmath170 of @xmath94 and then @xmath6 has to be affine linear .
\(iv ) denote @xmath536 .
without loss of generality one can assume @xmath537 .
let @xmath538 be an orthonormal basis of @xmath539 and @xmath540 be an orthonormal basis of @xmath541 . by lemma [ sec30 ] , for any @xmath542 , @xmath543 i.e. @xmath544 on the other hand , lemma [ sec6 ] and lemma [ sec25 ] tell us @xmath545 for every @xmath546 . since @xmath6 is a minimal submanifold , @xmath547 hence @xmath548 for all @xmath542 .
in other words , @xmath549 for any @xmath550 , and @xmath551 for any @xmath552 and @xmath553 , which follows from the weingarten equations .
if @xmath554 , then similarly one can deduce that @xmath555 for any @xmath556 and @xmath557 for any @xmath552 and @xmath553 . in conjunction with @xmath558 for any @xmath550 and @xmath559 for any @xmath556 , we have @xmath511 on @xmath6 and @xmath6 has to be totally geodesic .
if @xmath560 or @xmath352 , then ( [ sec9 ] ) implies @xmath561 for any @xmath552 and each @xmath562 . since @xmath563 for any @xmath550 , ( [ sec20 ] ) tells us @xmath564 for each @xmath565 .
thus @xmath566 which forces @xmath567 for each @xmath565 . in other words , @xmath557 for any @xmath552 and @xmath553 , and @xmath555 for any @xmath568
. therefore @xmath511 on @xmath6 and @xmath6 has to be affine linear .
let @xmath569 be a smooth vector - valued function , then for any @xmath570 , any jordan angle between @xmath98 and the coordinate @xmath9-plane takes values in @xmath571 ( see @xcite ) .
hence theorem [ tri ] implies : let @xmath269 be an open domain of @xmath572 and @xmath573 . if @xmath271 is a minimal submanifold with cja relative to the coordinate @xmath9-plane , and @xmath304 , then @xmath6 has to be an affine @xmath92-plane .
this is the theorem 1.1 mentioned in 1.5 .
[ co - cja ] let @xmath187 denote the octonions , which is an @xmath574-dimensional normed algebra over @xmath575 with multiplicative unit @xmath576 .
more precisely , @xmath187 is equipped with an inner product @xmath577 , whose associated norm @xmath578 satisfies @xmath579 for any @xmath580 .
denote by @xmath581 the 1-dimensional subspace spanned by @xmath576 , and by @xmath210 the orthogonal complement of @xmath582 .
then every @xmath583 has a unique decomposition @xmath584 with @xmath585 .
the conjugation of @xmath205 is defined by @xmath586 for @xmath587 , let @xmath588 ( @xmath589 ) denote the linear operator of right ( left ) multiplication by @xmath295 , respectively . with the aid of ( [ norm ] ) and ( [ conjugation ] ) , one can easily deduce the following fundamental formulas ( see appendix iv.a of @xcite ) : @xmath590 @xmath591 @xmath592 let @xmath1 be a @xmath593-dimensional real subspace of @xmath197 , if @xmath594 is a quarternion subalgebra of @xmath595 ( i.e. @xmath78 is isomorphic to @xmath596 ) , then @xmath1 is said to be _
associative_. [ o4 ]
let @xmath1 be an associative subspace of @xmath210 and @xmath597 be unit elements in @xmath1 that are orthogonal to each other , then @xmath598 is an orthonormal basis of @xmath1 , and @xmath599 conversely , if @xmath600 is an orthonormal basis of an associative subspace @xmath1 , then @xmath601 or @xmath602 . since @xmath603 is a subalgebra of @xmath595 , @xmath604 . by ( [ conjugation ] ) and ( [ o3 ] ) , @xmath605 i.e. @xmath606 . applying ( [ o1 ] ) and ( [ norm ] ) gives @xmath607 hence @xmath598 is an orthonormal basis of @xmath1 .
similarly , one can show @xmath608 is also a unit element in @xmath1 orthogonal to @xmath609 , hence @xmath610 or @xmath611 . if @xmath610 , then @xmath612 on the other hand , since @xmath205 and @xmath613 are linearly independent , @xmath614 and it follows from ( [ norm ] ) that @xmath615 , which contradicts ( [ o6 ] ) .
hence @xmath616 and it follows that @xmath617 similarly one can prove @xmath618 and @xmath619 .
* case i. * @xmath627 and @xmath628 .
this means there exist @xmath11 unit elements @xmath629 that are orthogonal to each other , then it follows from lemma [ o4 ] that @xmath630 is an orthonormal basis of @xmath631 .
hence @xmath632 and @xmath633 .
* case ii . *
@xmath634 and @xmath635
. then there exists @xmath11 unit elements @xmath636 that are orthogonal to each other . by lemma [ o4 ] , @xmath637 is a unit vector in @xmath1 , and @xmath638 .
hence @xmath639 , @xmath640 , @xmath641 , and @xmath1 is spanned by @xmath642 and @xmath643 , which are the angle directions of @xmath1 relative to @xmath211 .
* case iii . *
@xmath644 and @xmath645 .
( note that @xmath646 ( @xmath647 ) means @xmath648 ( @xmath649 ) , respectively . )
firstly , we claim @xmath650 . if not , there exist unit elements @xmath651 and @xmath652 ; by lemma [ o4 ] , @xmath1 is spanned by @xmath653 and @xmath654 ; hence @xmath641 , contradicting @xmath645 .
hence there exist mutually orthogonal elements @xmath655 that are unit angle directions of @xmath1 relative to @xmath211 associated to @xmath656 , respectively .
more precisely , @xmath657 here @xmath12 denotes the orthogonal projection of @xmath197 onto @xmath223 and @xmath13 denotes the orthogonal projection of @xmath197 onto @xmath1 . as in section [ cja ] , we denote by @xmath44 the orthogonal projection of @xmath197 onto @xmath658 , then @xmath659 with @xmath660 and @xmath661 , satisfying @xmath662 for each @xmath663 .
let @xmath311 be the unique element in @xmath595 satisfying @xmath664 , then for every @xmath665 , @xmath666 which implies @xmath667 . and
@xmath201 directly follows from @xmath664 and @xmath668 .
similarly , one can prove that there exists a unique @xmath669 which satisfies @xmath670 , and moreover @xmath671 . by lemma [ o4 ]
, @xmath674 is an orthonormal basis of @xmath1 , thus for each @xmath663 , @xmath675 when @xmath676 , the above equation gives @xmath677 in conjunction with @xmath678 we have @xmath679 , therefore @xmath680 . letting @xmath681 in ( [ o5 ] ) yields @xmath682 and
moreover @xmath683 observing that @xmath684 and @xmath685 form an orthonormal basis of @xmath223 , we have @xmath686 .
otherwise , @xmath695 and ( [ o7 ] ) gives @xmath696 which implies @xmath697 . without loss of generality
, one can assume @xmath698 , then @xmath699 .
now we put @xmath700 and @xmath701 , then @xmath702 which satisfy @xmath703 , @xmath704 and @xmath705 .
[ o8 ] let @xmath1 be an associative subspace of @xmath197 , and @xmath706 be the jordan angles between @xmath1 and @xmath223 , then @xmath694 moreover , there exist an orthonormal basis @xmath235 of @xmath223 satisfying @xmath707 , and a unit element @xmath667 , such that @xmath708 are unit angle directions of @xmath1 relative to @xmath211 , and @xmath709 .
let @xmath6 be a 4-dimensional submanifold in @xmath197 .
if the normal space at every point of @xmath6 is associative , then we call @xmath6 a _ coassociative submanifold _ ( see @xcite ) .
let @xmath100 be an arbitrary point of @xmath6 , denote by @xmath706 the jordan angles between @xmath710 ( an associative subspace ) and @xmath223 , then by proposition [ o8 ] , @xmath711 denote @xmath235 to be the orthonormal basis of @xmath223 satisfying @xmath707 and @xmath311 to be the unit element in @xmath221 , such that @xmath712 are all unit angle directions of @xmath710 relative to @xmath223 , and @xmath713 .
denote @xmath714 then it is easy to check that @xmath715 and @xmath716 for each @xmath717 and @xmath718 .
hence @xmath242 is an orthonormal basis of @xmath719 . whenever @xmath720 , let @xmath721 denote the isometric automorphism of @xmath722 as in [ 1.2 ] , then it follows from ( [ phi ] ) that @xmath723 hence @xmath724 and similarly @xmath725
; in conjunction with ( [ o9 ] ) , @xmath726 in summary we get a proposition as follows .
[ coass ] let @xmath6 be a coassociative submanifold in @xmath197 , @xmath95 and @xmath706 be the jordan angles between @xmath710 and @xmath211 , then there exist an orthonormal basis @xmath727 of @xmath710 and an orthonormal basis @xmath242 of @xmath719 , such that now we extend @xmath727 as an orthonormal normal frame field on @xmath737 , a neighborhood of @xmath100 , such that @xmath738 for every @xmath739 .
lemma [ o4 ] implies @xmath740 or @xmath741 for an arbitrary @xmath742 . due to the continuity , @xmath713 on @xmath737 and differentiating both sides with respect to @xmath743
gives @xmath744 i.e. @xmath745 with the aid of lemma [ o4 ] , lemma [ mult ] and proposition [ coass ] , a straightforward calculation shows @xmath746 and @xmath747 comparing with ( [ o11 ] ) and ( [ o12 ] ) , we arrive at the following conclusion .
[ co_sec11 ] let @xmath6 be a coassociative submanifold in @xmath210 , @xmath95 .
let @xmath748 and @xmath749 be as in proposition [ coass ] .
then for each @xmath750 , @xmath751 here @xmath752 are the coefficients of the second fundamental form at @xmath100 .
now we additionaly assume @xmath6 has cja relative to @xmath211 .
let @xmath753 , @xmath706 be the jordan angles between @xmath754 and @xmath223 , @xmath727 be the orthonormal basis of @xmath754 and @xmath242 be the orthonormal basis of @xmath755 , satisfying the properties in proposition [ coass ]
. then ( [ sec1 ] ) and lemma [ sec25 ] implies @xmath756 and substituting it into ( [ co1])-([co4 ] ) gives @xmath757 furthermore , applying lemma [ sec30 ] and [ sec31 ] yields the following propositions .
* case i. * @xmath763 and @xmath764 .
then @xmath765 and the equality holds if and only if @xmath766 .
it is well - known that coassociative submanfolds are absolutely area minimizing ( see @xcite
iv.2.b ) . by theorem [ tri ]
, @xmath6 has to be an open set of an affine 4-plane .
* case ii .
* @xmath767 and @xmath768 denote @xmath769 , then @xmath770 , @xmath771 ; @xmath772 and @xmath773 with @xmath774 for each @xmath734 ; @xmath775 , @xmath776 and @xmath777 @xmath778 .
lemma [ sec6 ] implies @xmath779 substituting the above equation into ( [ co6 ] ) and ( [ co7 ] ) , we get @xmath780 applying lemma [ sec2 ] gives @xmath781 in conjunction with ( [ co5 ] ) , we have @xmath782 let @xmath405 and @xmath418 be tensors of type @xmath783 , defined in ( [ r ] ) and ( [ u ] ) , respectively .
then @xmath784 @xmath785 and @xmath786 by ( [ sec19 ] ) , @xmath787 where @xmath788 hence @xmath789 and moreover @xmath790 in conjunction with ( [ co_sec1 ] ) , ( [ co_sec2 ] ) and ( [ co_sec3 ] ) , we obtain @xmath791 putting @xmath792 in ( [ sec9 ] ) gives @xmath793 ( where we have used ( [ co_sec4 ] ) , ( [ co6 ] ) and ( [ co7 ] ) ) .
by lemma [ sec31 ] , @xmath794 @xmath795 and similarly @xmath796 combining ( [ co_sec5])-([co_sec6 ] ) gives @xmath797 which forces @xmath798 ( since @xmath799 ) and moreover @xmath800 in conjunction with ( [ co_sec1 ] ) , ( [ co_sec7 ] ) , ( [ co_sec3 ] ) , ( [ co_sec8 ] ) and ( [ co_sec4 ] ) , we have @xmath801 .
the arbitrariness of @xmath228 implies @xmath511 , i.e. @xmath6 is totally geodesic .
* case iv . *
@xmath805 and @xmath806 .
denote @xmath807 , then @xmath808 , @xmath809 ; @xmath810 and @xmath811 with @xmath774 for each @xmath812 ; @xmath813 and
@xmath814 with @xmath815 ; @xmath816 . applying lemma [ sec2 ] and [ sec6 ] gives @xmath817 substituting the above equations into ( [ co5])-([co8 ] ) yields @xmath818 a straightforward calculation shows @xmath819 @xmath820 @xmath821 and then lemma [ sec30 ] implies @xmath822 where @xmath823 since @xmath824 , @xmath825 and then @xmath826 .
hence @xmath827 and moreover @xmath828 in conjunction with ( [ co_sec1 ] ) , ( [ co_sec17 ] ) , ( [ co_sec9 ] ) and ( [ co_sec10 ] ) , we have @xmath829 . with the aid of lemma [ sec31 ]
, one can then proceed as in case ii to deduce that @xmath801 . since @xmath228 is arbitrary
, @xmath6 has to be affine linear .
[ p2 ] let @xmath6 be a coassociative submanifold of @xmath210 .
assume @xmath6 has cja relative to @xmath211 , and @xmath830 , then either @xmath6 is affine linear , or there exists @xmath831 , such that @xmath6 is a translate of an open subset of @xmath832 .
here @xmath832 denotes the lawson - osserman s cone , see ( [ l - o - cone ] ) .
let @xmath833 , @xmath834 and @xmath807 .
for an arbitrary point @xmath753 , let @xmath242 be an orthonormal tangent frame field and @xmath727 be an orthonormal normal frame field on @xmath737 , a neighborhood of @xmath228 , such that for any @xmath95 , @xmath835 and @xmath836 satisfy the properties in proposition [ coass ] . in particular , @xmath837 for each @xmath718 . with the aid of lemma [ sec2 ] , lemma [ sec6 ] and proposition [ co_sec11 ] , one can proceed as above to get some pointwise relations between the coefficients of the second fundamental form , see ( [ co_sec1 ] ) , ( [ co_sec12])-([co_sec10 ] ) .
denote @xmath838 then @xmath839 can be seen as a smooth function on @xmath737 , and @xmath840 applying lemma [ sec31 ] , we have @xmath843 @xmath844 and similarly @xmath845 adding ( [ co_sec13])-([co_sec14 ] ) gives @xmath846 where @xmath847}{(\cos 2\th_1 - 1)(\cos 2\th_1+\cos\th_1)}\\ = & \f{2\sin
2\th_1(3\cos \th_1 - 2)(\cos\th_1 + 1)}{(\cos 2\th_1 - 1)(\cos 2\th_1+\cos\th_1)}=0 .
\endaligned\ ] ] hence @xmath848 and substituting it into ( [ co_sec9])-([co_sec10 ] ) implies @xmath849 . by the frobenius theorem , it suffices to prove that the subbundle @xmath866 of @xmath867 is integrable ; more precisely , given arbitrary smooth sections @xmath318 of @xmath866 , @xmath868 $ ] takes values in @xmath866 as well .
now we write @xmath869 and @xmath870 , then @xmath871=x^i y^j [ e_i , e_j]+x^i ( \n_{e_i}y^j)e_j - y^j(\n_{e_j}x^i)e_i\ ] ] and hence @xmath872,e_1\ran = x^i y^j\lan [ e_i , e_j],e_1\ran.\ ] ] hence it is necessary and sufficient for us to show @xmath873,e_1\ran=0 $ ] for any @xmath874 . since @xmath371 is torsion - free , @xmath875,e_1\ran=&\lan \n_{e_2}e_3,e_1\ran-\lan \n_{e_3}e_2,e_1\ran=\g_{23}^1-\g_{32}^1=-\g_{21}^3+\g_{31}^2\\ = & -(\k_{\th\th_1}h_{3,21}-\k_{\th_1\th}h_{1,23})+(\k_{\th\th_1}h_{2,31}-\k_{\th_1\th}h_{1,32})\\ = & 0 , \endaligned\ ] ]
@xmath876,e_1\ran=&\lan \n_{e_2}e_4,e_1\ran-\lan \n_{e_4}e_2,e_1\ran=\g_{24}^1-\g_{42}^1=-\g_{21}^4+\g_{41}^2\\ = & \k_{\th_1 0}h_{1,24}+(\k_{\th\th_1}h_{2,41}-\k_{\th_1\th}h_{1,42})\\ = & 0 \endaligned\ ] ] and similarly @xmath877,e_1\ran=\k_{\th_1 0}h_{1,34}+(\k_{\th\th_1}h_{3,41}-\k_{\th_1 \th}h_{1,43})=0.\ ] ] then the claim is proved . without loss of generality
, we can assume that the closure of @xmath280 is contained in @xmath737 .
then there exists @xmath878 , such that @xmath879 for every @xmath284 and any @xmath880 , where @xmath881 denotes the position vector of @xmath100 in @xmath210 .
define @xmath882 @xmath883 then @xmath884 is a diffeomorphism between @xmath885 and a neighborhood of @xmath228 in @xmath6 , which is denoted by @xmath886 .
define @xmath894 @xmath895 where @xmath896 is a constant to be chosen
. then @xmath897 \endaligned\ ] ] for each @xmath898 .
more precisely , @xmath899 now we put @xmath900 then combining ( [ co_sec16 ] ) and ( [ k10 ] ) implies @xmath901 for each @xmath898 . hence @xmath902 is a constant map on @xmath280 . without loss of generality , we can assume @xmath903 , i.e. @xmath904 for every @xmath284 .
in other words , @xmath280 lies in the euclidean sphere centered at @xmath140 and of radius @xmath896 , and an arbitrary normal line of @xmath280 , i.e. @xmath905 with @xmath284 , must go through the origin .
therefore @xmath886 is a cone .
let @xmath906 be a submanifold of @xmath197 .
for any @xmath907 , there exist a unit element @xmath680 and a unit element @xmath667 , such that @xmath908 define @xmath909 then @xmath910 is a @xmath593-dimensional subspace of @xmath911 .
furthermore @xmath912 is a @xmath593-dimensional distribution on @xmath125 . for any @xmath284 , @xmath913 is a unit tangent angle direction associated to @xmath253 .
hence there exist @xmath680 and @xmath667 satisfying @xmath914 , such that @xmath915 moreover , @xmath916 therefore @xmath917 .
denote @xmath918 then @xmath919 is a unit angle direction of @xmath98 with respect to @xmath211 .
on the other hand , proposition [ o8 ] implies the existence of an orthonormal basis @xmath920 of @xmath223 satisfying @xmath921 and a unit element @xmath922 , such that @xmath923 are all unit angle directions of @xmath98 relative to @xmath223 . since @xmath924 , @xmath925 , and
then one can assume @xmath926 , @xmath927 without loss of generality , which implies @xmath928 noting that @xmath929 and @xmath930 , it is easy to deduce that @xmath931 , i.e. @xmath280 is an integral manifold of @xmath932 . for any @xmath933 , @xmath208 is a coassociative cone , which has cja with @xmath251 , and each ray is an angle line with respect to @xmath253 .
as above , one can show that @xmath934 and that it is also an integral manifold of @xmath932 .
now we write @xmath935\ ] ] with @xmath936 satisfying @xmath937 . then choosing @xmath938 gives @xmath939\in m(a_0).\ ] ]
therefore @xmath280 and @xmath940 are both integral manifolds of @xmath932 .
since @xmath940 is complete , applying the frobenius theorem implies @xmath941 , and hence @xmath942 . finally ,
because minimal submanfolds in euclidean space are analytic manifolds , @xmath6 has to be an open subset of @xmath832 .
firstly , @xmath943 implies @xmath511 on @xmath280 .
denote by @xmath944 the second fundamental form of @xmath280 in @xmath197 , then @xmath945 for any @xmath946 and @xmath718 , @xmath947 for any @xmath948 and @xmath949 thus @xmath950 , i.e. @xmath280 is totally geodesic .
since @xmath951 for each @xmath898 , @xmath246 is parallel along @xmath280 .
therefore @xmath886 is an open subset of an affine linear subspace of @xmath210 . due to the analyticity of minimal submanifolds
, @xmath6 has to be affine linear . and
the proof is completed .
let @xmath6 be a coassociative submanifold in @xmath210 .
assume @xmath6 has cja relative to @xmath211 .
if @xmath758 and @xmath759 , then either @xmath6 is affine linear , or there exists @xmath831 and @xmath952 , such that @xmath6 is an open subset of @xmath953+w_0:q\in \text{sp}_1,r\in \bbb{r}^+\}.\ ] ] in other words , @xmath6 is a translate of a portion of the lawson - osserman s cone .
let @xmath269 be an open domain of @xmath596 and @xmath954 .
if @xmath271 is a coassociative submanifold with cja relative to @xmath211 , and @xmath758 , then @xmath166 is either an affine linear function or @xmath307 , where @xmath308 , @xmath309 and @xmath310 with @xmath311 an arbitrary unit element in @xmath211 .
s. s. chern , m. do carmo and s. kobayashi : minimal submanifolds of a sphere with second fundamental form of constant length .
in functional analysis and related fields , springer , new york , ( 1970 ) , 59 - 75 .
z. drma : on principal angles between subspaces of euclidean space .
siam j. matrix anal .
* 22 * ( 2000 ) , 173 - 194 .
j. fastenakels , m. l. munteanu and j. van der veken : constant angle surfaces in the heisenberg group .
acta math . sin .
( engl . ser . )
* 27 * ( 2011 ) , 747 - 756 . | to study the lawson - osserman s counterexample @xcite to the bernstein problem for minimal submanifolds of higher codimension , a new geometric concept , submanifolds in euclidean space with constant jordan angles(cja ) , is introduced . by exploring the second fundamental form of submanifolds with cja
, we can characterize the lawson - osserman s cone from the viewpoint of jordan angles
. 0.1 mm |
the study of avoiding repetition in combinatorial structures stems from a 1906 paper by axel thue @xcite , which showed that a there is an infinite word in @xmath10 that has no subword of the form @xmath11 for some finite word @xmath12 .
such a word is called squarefree . a paper by alon , grytczuk , haluszczak , and riordan @xcite introduced the generalization to edge colorings of graphs .
an edge coloring of a graph is _ nonrepetitive _ if for any open path through the graph , the pattern of colors along the path is squarefree .
the thue number @xmath4 is the smallest number @xmath13 for which @xmath5 can be nonrepetitively colored with @xmath13 colors .
alon et al .
also mentioned the possibility of studying _ vertex nonrepetitive colorings _
, i.e. , vertex colorings with the same property .
this notion has been further studied in papers by czerwiski and grytczuk @xcite and by bart and varj @xcite .
other similar notions involving walks rather than paths have been developed by brear and klavv zar @xcite and again by bart and varj @xcite . for a more detailed survey of results regarding vertex nonrepetitive colorings ,
we refer the reader to grytczuk @xcite .
the many papers in this area employ various notations and terminologies ; we will stick to the original terminology of alon et al . whenever possible .
therefore , we refer to edge nonrepetitive colorings simply as nonrepetitive colorings and denote the edge coloring thue number as @xmath4 .
we call a path _ square _ if it is of the form @xmath11 for some color - string @xmath12 .
our goal is to study nonrepetitive colorings from the point of view of computational complexity .
whatever version of @xmath4 is used , deciding whether @xmath14 is an @xmath15 problem : does there exist a @xmath13-coloring of the edges or vertices of @xmath5 such that the pattern of colors along any open path , or one of the kinds of walks , is squarefree ?
such questions usually belong in the class @xmath16 , a class in the second level of the polynomial hierarchy . this class and problems known to be complete for it are described in a survey by schaefer and umans @xcite . many of the results of papers on nonrepetitive colorings which bound the thue number of various classes of graphs for example , alon et al.@xcite show that @xmath17 for constant @xmath18 for any graph of degree @xmath19 . on the other hand ,
alon et al . also show that for any @xmath19 there is a graph @xmath5 of maximum degree @xmath19 such that any nonrepetitive vertex coloring of @xmath5 requires at least @xmath20 colors .
both these results , and many others like them , are probabilistic and thus give us little information about specific graphs .
the complexity of vertex nonrepetitive colorings has recently been studied by marx and schaefer @xcite , who showed that deciding whether such a coloring is nonrepetitive is * conp*-complete . in this paper
we show that the corresponding problem for edge colorings is * conp*-complete .
we then show that deciding whether @xmath14 is @xmath21-complete . finally , we turn to colorings that are nonrepetitive for bounded - length paths .
let @xmath22 be the least number of colors required to color the edges of @xmath5 so that no open path contains the square of a pattern of colors of length @xmath23 , and let @xmath24 be the least number of colors required to color the vertices of @xmath5 under analogous conditions .
since the number of paths of length @xmath25 is bounded by @xmath26 , for fixed @xmath27 each such problem is contained in * np*. then @xmath28 is simply graph coloring and @xmath29 is edge coloring , shown to be * np*-complete by holyer in 1981 @xcite .
@xmath30 is known as the symmetric chromatic number ; the corresponding decision problem is shown to be * np*-complete by coleman and mor @xcite .
we show that the decision problems corresponding to @xmath31 and @xmath32 for any @xmath33 are * np*-complete .
the nonrepetitive colorability of a graph with @xmath13 colors is a rather slippery property . in the construction of a nonrepetitive coloring ,
changing the color of an edge may produce a square path of arbitrary length and thus affected by colors of edges arbitrarily far away .
thus the local properties of a graph can not guarantee that it is nonrepetitively colorable . on the other hand ,
if a graph does not have a nonrepetitive coloring , then we must examine all of the colorings and find a square path in each , which also feels like a hard problem .
thus neither direction of the reduction is easy .
there are two immediately visible approaches to resolving this dilemma .
perhaps we could focus on fairly sparse graphs , thus minimizing the number of paths we must make sure are nonrepetitive .
but this would greatly increase the number of colorings we must consider .
we thus use this approach only in determining whether a preassigned coloring is nonrepetitive . on the other hand
, we could focus on graphs in which the degree of every vertex is close to the maximal degree of the graph .
although this allows us to easily discard most possible colorings , the number of distinct paths grows exponentially .
additionally , it becomes harder to create graphs which differ subtly in global structure in such a way that some of them are nonrepetitively colorable and others are not . in our proof
, we combine these two approaches , producing graphs that are locally dense , in the sense that the degree of most vertices is large and close to the maximal degree , but globally sparse , in the sense that long paths must traverse a number of bottlenecks with only a few connections . while a pattern of local structures , or gadgets , is common to many if not all graph - theoretic complexity proofs , in this case in particular it allows us to partially isolate each portion of the problem : once we have established that the local structure is nonrepetitively colorable in a very constrained set of ways , we can us this to determine whether the global structure is nonrepetitively colorable .
we will first construct the global structure , which is similar to the well - known reduction which shows that hampath is * np*-complete .
the local structure , added later , will be composed of several types of dense subgraphs .
the simplest of these are the cliques of size @xmath34 , which alon et al .
show to be nonrepetitively colorable with @xmath34 colors .
we also use the @xmath6-dimensional hypercube graphs , which have a number of useful properties which are discussed in the next section .
finally , in a number of cases we will _ saturate _ a vertex @xmath35 by adding extra vertices only joined to @xmath35 , called the _ plume _ of @xmath35 , to artificially inflate the degree .
this will allow us to more effectively use the saturation lemma below .
we will now prove some results about nonrepetitive colorings that will aid in the complexity proofs .
let a vertex of a graph be _ saturated _ if it is of maximal degree , and let a _ diamond _ be a cycle of four vertices .
suppose two opposing vertices of a diamond , @xmath36 and @xmath37 , in a graph of degree @xmath13 , are saturated and the saturating edges are not directly connected
. then the graph may only be @xmath13-nonrepetitively colored if the diamond is colored with exactly two colors .
let @xmath38 and @xmath39 be the colors of one path from @xmath36 to @xmath37 through the diamond .
then , since @xmath36 and @xmath37 are saturated , there must be an edge of color @xmath39 at @xmath36 and an edge of color @xmath38 at @xmath37 , forming a path of colors @xmath40 . if this path is not a cycle , then the coloring is not nonrepetitive .
thus , these colors must form the other half of the diamond .
the nonrepetitive @xmath13-edge - coloring of a @xmath13-hypercube exists , is unique up to permutation , and has the following properties : 1 .
the shortest path between any two points consists of distinct colors . 2 .
any permutation of the colors of a path between two points consisting of distinct colors also colors a path between those two points , and no other sequence of distinct colors does .
3 . any path consisting of distinct colors is a shortest path between its endpoints .
4 . there are @xmath41 vertices at distance @xmath27 from a given vertex @xmath42 .
the uniqueness of coloring also holds for portions of a @xmath13-hypercube whose distance from a chosen base vertex @xmath42 is at most @xmath43 , which we call the _
first @xmath43 layers _ of the hypercube .
we will identify the vertices of the hypercube with @xmath44 , where there is an edge between @xmath45 and @xmath46 for every @xmath27 . without loss of generality ,
let us choose the vertex @xmath42 identified with @xmath47 to be a base vertex .
@xmath42 is saturated , so the coloring of the edges adjacent to it is unique up to permutation .
thus the coloring of the first layer is unique .
now suppose that we have colored the first @xmath43 layers . every vertex is saturated and any two adjacent edges form a diamond ( a face . )
hence by the previous lemma the coloring of the @xmath43th layer determines the coloring of the @xmath48st layer . by induction ,
the coloring of the first @xmath43 layers is unique , and so is the coloring of the entire hypercube . to show that a coloring exists ,
let us use the color @xmath27 for all edges for which @xmath49 changes between the two vertices .
for every @xmath27 , the hypercube consists of two @xmath50-hypercubes , one in which the @xmath27th coordinate is 0 for all vertices and one in which it is 1 .
therefore , if we have a path containing two edges of color @xmath27 , then we can create an alternate path with the same start and end vertices by removing these two edges and reversing @xmath49 for the vertices in between them .
now suppose we have a square path .
then for every dimension , there is an even number of edges in that dimension on the path , and we can remove all these edges by pairs from the path . thus the path is equivalent to the empty path , is therefore a cycle , and all open paths are squarefree .
we will now use this construction to prove the rest of the properties . for ( 1 ) , suppose that we have a shortest path between two points that contains two edges of the same color .
then we can remove these two edges and get a shorter path between the same two vertices .
therefore the shortest path must consist of distinct colors .
for ( 2 ) , clearly a permutation of the changes to variables leads to the same result .
on the other hand , for any two shortest paths whose sets of colors differ , there is an @xmath27 such that @xmath49 is different between the two end vertices for one of the paths , but not for the other .
thus the end vertices are different for the two paths .
\(3 ) follows directly from ( 1 ) and ( 2 ) .
a path of length @xmath27 represents a change in @xmath27 members of the @xmath13-tuple .
there are @xmath41 possible such changes , proving ( 4 ) .
in this section , we consider several related problems in which the coloring is entirely or almost entirely predetermined , and the main difficulty of the problem is in finding whether the coloring is nonrepetitive .
the proofs in this section are all related and provide a framework for the proof of the @xmath21-completeness of thue number .
we first consider colorings of directed graphs . given a directed graph @xmath5 and a coloring thereof , it is * conp*-complete to determine whether the coloring is nonrepetitive .
we call this problem directed nonrepetitive coloring .
we reduce from co-3sat .
the reduction is similar to the standard reduction from 3sat to hampath .
the edges are given a large number of colors , with 4 or fewer edges per color .
in effect , we restrict possible square paths to those that go through two edges of every color , mimicking the requirement that a hamilton path must pass through every vertex .
let @xmath51 be an instance of 3sat .
the reduction will use two types of gadgets : a variable gadget and a clause gadget .
the variable gadget consists of two sets of @xmath52 vertices , where @xmath52 is the largest number of number of instances of a literal , together with a beginning vertex @xmath53 .
a path of edges goes from @xmath53 through each of the vertices in a set to @xmath54 , or a vertex @xmath18 if @xmath55 . the sets of vertices and paths going through them correspond to a true assignment and a false assignment ; thus each set signifies a literal .
suppose @xmath56 has clauses numbered @xmath57 .
a clause gadget for clause @xmath58 has one vertex @xmath59 , and we add an extra vertex @xmath60 and an edge from @xmath18 to @xmath61 . for each literal in @xmath58 ,
we add an edge from @xmath59 to a vertex on the path corresponding to this literal and an edge from this vertex to @xmath62 or @xmath60 if @xmath63 , making sure that this is the only clause gadget that uses this vertex . to complete the construction of the graph , we add a series of vertices @xmath64 and edges connecting @xmath65 to @xmath66 and @xmath67 to @xmath68 .
we call this structure the _ snout_. we must now assign a coloring to this graph
. we will use a set @xmath69 of colors .
starting from the tip , we color the @xmath27th edge of the snout with @xmath70 .
we color the @xmath13th edge of a path through the @xmath71-gadget by @xmath72 ; an edge going into @xmath59 by @xmath73 ; an edge leaving @xmath59 by @xmath74 ; and an edge going into @xmath60 by @xmath75 .
given a satisfying assignment to @xmath56 , we can construct a square path by first traversing the snout , then taking the path through each variable gadget corresponding to the opposite of its assignment and the path through each clause gadget corresponding to a true literal , as illustrated for an example .
conversely , if we have such a path , then taking the opposite of a traversed literal - path to be true gives us a satisfying assignment .
, with a square path highlighted .
] now suppose that @xmath56 is unsatisfiable , and that there is a square path @xmath76 , which then can not have the form above .
suppose first that the edges of @xmath76 are colored only by the colors of the variable gadgets .
since the colors of the variable gadgets form the tip , but not the base of the snout , @xmath76 must consist only of the variable gadgets themselves ; but this is impossible since the variable gadgets can only be traversed in the forward direction and thus a path through them has no two edges of the same color .
( * ) thus @xmath76 must contain edges from clause gadgets .
since any two edges in a clause gadget that share a color also share a vertex , an open path can contain only one of such a set .
thus , for every color @xmath38 that colors an edge of @xmath76 contained in a clause gadget , @xmath76 contains an edge in the snout of color @xmath38 . since @xmath68 is only the source of the two branches of the first variable gadget , one of these two branches must also be contained in the path .
furthermore , since these branches share a vertex , only one of them may be contained in @xmath76 .
the only other edge colored by @xmath77 is the edge at the very tip of the snout ; thus the entire snout must be contained in @xmath76 , so @xmath76 must contain a path belonging to each variable and clause gadget . as above , this defines a satisfying assignment to @xmath56 , a contradiction .
clearly , this reduction runs in polynomial , in fact quadratic , time .
the graph resulting from this reduction has maximum in- and out - degree 3 .
each color is repeated at most four times .
from this we can fairly easily show the same result for undirected graphs .
given an undirected graph @xmath5 and an edge - coloring thereof , it is * conp*-complete to determine whether the coloring is nonrepetitive .
we call this problem nonrepetitive coloring . we will now modify the previous reduction to work with undirected graphs .
let us add the colors @xmath78 , for @xmath79 and @xmath80 .
we call these the _ direction - determining colors _ and the rest the _ original colors_. we replace every edge inside the snout with the graph in figure 2 and every edge outside the snout with the path @xmath81 for some @xmath82 .
this can be done in such a way that there are no trivial repetitions , that is , no two edges of the same color are incident to the same vertex , since the maximum in- and out - degree of the directed graph is 3 . clearly , this construction retains the property that each color colors no more than four edges , since none of the direction - determining colors need color more than two edges .
we will show that this modification leaves the reduction valid . .
] if @xmath56 has a satisfying assignment , we can traverse the graph as before , choosing the right path through the snout to produce the same @xmath65s , @xmath53s , @xmath83s , and @xmath84s as in the other half of the path .
now suppose that @xmath56 is unsatisfiable .
suppose that the graph has a square path .
clearly , this path must contain direction - determining colors .
each such color is represented only once in the non - snout edges , so we can not make a square path without any snout edges .
this means that the paragraph marked ( * ) in the previous proof still applies , and we are done . we now modify the statement of the problem slightly in order to get a @xmath21-complete problem .
given an undirected graph @xmath5 and a set @xmath85 of pairs @xmath86 for each edge @xmath87 , with @xmath88 a set of colors , it is @xmath89-complete to determine whether there exists a nonrepetitive coloring of @xmath5 that uses a color @xmath90 for each @xmath87 .
we call this problem restricted thue number . in our case , @xmath91 will contain at most two colors for any edge , and only one color for most edges .
clearly , the problem is in @xmath89 . to prove that it is @xmath89-hard
, we shall slightly modify the previous reduction , reducing from co-@xmath923sat .
let @xmath93 be an instance of @xmath923sat .
our graph @xmath5 will be the graph constructed for @xmath94 in theorem 2 . for each universally quantified variable , the two original edges that initially separate the variable gadget into a `` true '' and a `` false '' branch are given the sets @xmath95 and @xmath96 respectively , where @xmath70 is the color given these edges in the previous reduction .
the corresponding edge in the snout is given the set @xmath97 .
the rest of the edges are given singletons containing the color they were colored in the previous reduction .
thus a coloring of the graph satisfying the restriction corresponds to an assignment of @xmath12 , since it forces any potentially square path to go through the specified branch of the corresponding variable gadget .
this means , by the previous construction , that there is such a nonrepetitive coloring of @xmath5 iff there is an assignment to @xmath12 for which @xmath98 is unsatisfiable .
this proof provides the final incarnation of the global structure which we will later harness to show that thue number is @xmath99-complete .
in this section , we explore two problems in which we consider a polynomial number of possibly square paths , and therefore only the existential quantifier in the statement of the problems is over an exponentially large domain . together , the problems form a generalization of edge coloring , a problem which was shown to be * np*-complete by holyer in 1980 @xcite . in these proofs
, we employ several tricks to find graphs whose small - diameter subgraphs are restricted in the ways they can be colored nonrepetitively .
these tricks will later find applications in the proof of the @xmath21-completeness of thue number .
thus although this material is not directly related to the main result , it builds some machinery for its proof .
still , a reader interested only in the main proof may go on to the next section .
it is * np*-complete to determine whether @xmath100 .
we reduce from edge coloring for cubic graphs ( graphs with 3 edges incident to every vertex ) and 3 colors , which is shown to be * np*-complete in @xcite .
the reduction consists of replacing each edge of graph @xmath5 with the graph depicted in fig . 3 , which we shall call a _
clam_. the vertices incident to the edge are mapped to @xmath36 and @xmath37 .
we shall argue that the resulting graph @xmath101 can be 6-colored with nonrepetitive 4-paths iff @xmath5 can be 3-colored .
suppose that we have such a coloring , and call the six colors @xmath102 .
note first that vertices @xmath36 and @xmath37 are both saturated , since they are images of vertices of @xmath5 of degree 3 , so the outer diamond of the clam only has two colors without loss of generality , @xmath38 and @xmath39 .
now suppose that one of the inner four edges ( let us call them the _ gills _ ) is colored @xmath38 .
then if the color of an edge of the inner diamond connected to this edge is @xmath18 , there is a path @xmath103 through gill , inner diamond , outer diamond , and an edge of another clam .
thus the gills must be colored differently from the outer diamond , as must the inner diamond since it is adjacent to the outer one .
this means that the vertex labeled @xmath42 and the one symmetric to it are saturated for the four colors other than @xmath38 and @xmath39 , and thus the inner diamond must be colored with exactly two colors ( @xmath18 and @xmath60 . )
now let us add two more clams , incident to vertices @xmath104 and @xmath105 respectively , at vertex @xmath36 .
suppose that the outer colors of the clams are @xmath106 and @xmath107 respectively .
then to prevent square paths of the type @xmath103 from the inner diamond of clam @xmath108 to the inner diamond of clam @xmath109 , the inner diamonds of clams @xmath108 and @xmath110 must consist of the colors @xmath107 and @xmath106 respectively .
but then we get square paths of the form @xmath111 from the inner diamond of clam @xmath108 to the inner diamond of clam @xmath110 .
thus the outer colors of @xmath108 and @xmath110 must be @xmath112 and @xmath113 , in some order .
furthermore , the inner colors of @xmath108 must be @xmath113 to prevent @xmath103-type repetitions .
this forces the inner colors of @xmath110 to be @xmath114 .
this means that the coloring of one clam at a vertex of @xmath5 determines the sets of colors available to the other clams ; by induction , each clam in the graph must be colored using one of these three patterns .
this means that if we have a 3-coloring of @xmath5 , we can map the three colors onto the three clam - coloring patterns to produce a 2-nonrepetitive 6-coloring of @xmath101 ; conversely , with a nonrepetitive 6-coloring of @xmath101 , we can map the three clam - coloring patterns onto three colors with which we can edge - color @xmath5 . for any natural number @xmath33 and graph @xmath5
, it is * np*-complete to determine whether @xmath115 .
this is a generalization of the previous reduction .
however , one must view the clam not as a unified gadget , but as a combination of elements of three gadgets .
the two edges incident to vertex @xmath36 belong to vertex gadget @xmath36 , and likewise for vertex @xmath37 .
the edge gadget proper , then , consists of the inner diamond and the gills .
we again reduce from edge coloring for cubic graphs in 3 colors .
let @xmath5 be a cubic graph .
we associate to @xmath5 a 3-coloring @xmath116 of the edges . as we construct the gadgets in the reduction
, we will also construct a @xmath117-coloring of the edges based on @xmath116 .
later , we will show that this coloring is @xmath118-nonrepetitive iff @xmath116 is a valid edge coloring , and that no other coloring of the edges is @xmath118-nonrepetitive .
we shall now enumerate the elements of the reduction , which reduces @xmath5 to a graph @xmath101 .
the vertex gadget consists of the first @xmath13 layers of a @xmath117-hypercube with respect to a central vertex @xmath119 . by the hypercube lemma
, this has a unique nonrepetitive coloring .
now let us divide the @xmath117 colors into three groups of @xmath120 , called _ red _ , _ green _ , and _ blue_. we shall use the term _ grue _ to refer to the union of the sets blue and green .
the elements of the sets will be referred to as _
shades_. for each of the groups , there are @xmath121 vertices whose distance from @xmath119 is @xmath13 such that the path to them from @xmath119 consists of edges of colors from that group .
we identify these vertices with an analogous group from another vertex gadget ( with central vertex @xmath122 ) in order to form an edge . in a @xmath118-nonrepetitive coloring ,
the colors of the edges on either side of this connection are in the same set of @xmath120 .
suppose we have a @xmath118-nonrepetitive coloring that does not have this property .
then take an arbitrary path from @xmath119 to @xmath122 of length @xmath120 , say with color pattern @xmath123 . by the properties of the @xmath117-hypercube
, there must be a path starting with @xmath119 of pattern @xmath124 and a path starting with @xmath122 of pattern @xmath125 .
this forms a square path of length @xmath118 , which must then be a cycle .
this means that the sets of colors on either side of the inter - vertex interface are the same .
assume they are both red .
then , since the fully - red paths from @xmath119 to @xmath122 form a @xmath120-hypercube , they can be colored nonrepetitively .
we assume now that we are working with two vertices , @xmath119 and @xmath122 , that are joined by a red edge .
the construction for the other two sets is analogous .
it remains to add a gadget that will force the green and blue sets of one vertex to be the same as the green and blue sets of the other . to this end , we take two vertices , @xmath36 and @xmath37 , on the interface between two vertex gadgets , whose distance from each other is @xmath120 , to be two ends of a consistency gadget .
the consistency gadget is a @xmath118-hypercube modified in two ways .
first , every vertex whose distance from @xmath36 and @xmath37 is 2 or more has a _ plume _ of @xmath120 additional edges leading to disconnected vertices .
a red edge separated by one edge from @xmath36 or @xmath37 generates a square path of length 4 , so this means all the vertices of the hypercube are effectively saturated .
this implies that the hypercube is entirely grue and the plumes are red . on the other hand
, the farther red plume edges do not generate such a path , since it would have to repeat a red edge followed by two or more grue edges as a path from the red interface vertices , and thus required by hypercube property ( 3 ) to stray a distance more than @xmath13 from the central vertex of the vertex gadget .
next , for every path starting at @xmath36 or @xmath37 and consisting of @xmath13 edges of distinct shades of green , we remove the last edge .
we then do the same for blue .
we also remove all red plume edges incident to the vertices that are left unsaturated by this removal .
we call the @xmath13th layer of the hypercube from either @xmath36 or @xmath37 the _ gap layer_. this consistency gadget is uniquely @xmath118-nonrepetitively @xmath117-colorable up to isomorphism .
this coloring is such that the hypercube edges are grue and colored as the restriction of a full hypercube , and the plumes are red .
any diamond containing edges in the layer adjacent to the gap layer has two saturated vertices at equal distance from the vertex gadget .
suppose there is an edge between two vertices that are both unsaturated due to the removal .
if those two vertices both became unsaturated from the removal of edges of the same color , then the edge would have to be of that color , and thus also removed . on the other hand , an edge between two vertices that are unsaturated due to the removal of different colors would not exist by hypercube property ( 2 ) .
thus any diamond that has edges in the gap layer has saturated vertices either laterally or vertically , and so by the saturation lemma the coloring is a restriction of the full hypercube .
we then add a second , identical consistency gadget at vertices @xmath104 and @xmath105 whose distances from @xmath36 and @xmath37 are at least 4 .
( this distance prevents a square path consisting of a red plume edge in the first consistency gadget , then some path through it , then two red edges , the same path through the second consistency gadget , and another red plume edge . )
we refer to the set of colors that form the interface between two vertex gadgets as an _
edge set_. if the original graph is not 3-edge - colorable , the resulting graph is not @xmath118-nonrepetitively @xmath117-edge - colorable . given a red edge set , suppose that another edge set of the same vertex gadget with central vertex @xmath119 , belonging to edge @xmath87 of @xmath5 , has both green and blue parts .
this means that that one of its consistency gadgets will have a path to it that contains both green and blue , since if one of the consistency gadgets has an all - blue side and an all - green side , then the other one does not .
thus we can find a square path consisting of : 1 . a red plume edge that duplicates the last edge of ( 3 ) , 2 .
a path of length @xmath13 in a consistency gadget of the red edge that duplicates ( 4 ) , 3 . a path from the consistency gadget of the red edge set to @xmath119 , 4 . a mixed blue - green path from @xmath119 to one of the blue - green edge set s consistency gadgets , and 5
. a path of length @xmath50 in this consistency gadget that duplicates the first @xmath50 edges of ( 3 ) .
therefore , the edge sets can not be of mixed colors with respect to the red edge set s consistency gadget .
if @xmath5 is 3-edge - colorable , then the coloring of @xmath101 which we have described is @xmath118-nonrepetitive .
suppose that @xmath5 is 3-edge - colorable , and let @xmath126 be a square path in @xmath101 of length at most @xmath118 , where @xmath127 and @xmath128 are the repetitions of the pattern .
the portion of @xmath76 inside a gadget @xmath129 will be called @xmath130 .
clearly , @xmath76 can not be wholly inside a consistency gadget .
we have also already established that @xmath76 can not be wholly outside any consistency gadgets .
thus it include edges from both consistency and vertex gadgets .
since a consistency gadget has distance @xmath118 between its two attachment points , @xmath76 can not have a consistency gadget in the middle . in introducing the red plumes in the consistency gadget
, we also showed that the whose middle vertex of @xmath76 can not be a junction between a consistency gadget and a vertex .
thus at least one half - path of @xmath76 ( without loss of generality , @xmath127 ) contains parts of both a consistency gadget @xmath131 and a vertex gadget @xmath132 with center @xmath119 .
@xmath133 contains at most one of the colors that are found on a shortest path to @xmath119 from the juncture with the consistency gadget .
if such a color is in @xmath133 , it must be a plume , which forces @xmath133 to be at least three edges long .
suppose now that the rest of @xmath76 is contained in @xmath132 .
we will show that the distance from the last vertex @xmath42 of @xmath76 to @xmath119 is greater than @xmath13 , a contradiction .
if a color has both of its repetitions inside @xmath132 , these cancel each other out , so we are left with the edges that repeat @xmath133 .
but since there is at least one more grue than red edge in @xmath133 , this means that our initial distance @xmath13 from @xmath119 increases by at least one .
this means that @xmath76 must access either a consistency gadget or another vertex gadget . without loss of generality , assume that @xmath131 belongs to a red edge set .
_ case 1 _ : @xmath76 accesses a second consistency gadget that belongs the same edge set .
then there are at least four shades of red that are repeated an odd number of times in the vertex gadget portion of the path .
since there is at most one red edge in each consistency gadget part of the path , nothing can repeat these edges .
_ case 2 _ : @xmath76 accesses a different consistency gadget , without loss of generality from a blue edge set .
then @xmath134 must contain at least @xmath120 colors that color an odd number of edges , since the distance between the two consistency gadgets is @xmath120 .
this is the maximal number of distinct colors in the path we are considering , so each of the @xmath120 must be repeated in one of the two consistency gadgets .
this means that one of the consistency gadgets contains a red - blue path of length at least @xmath13 .
however , this must pass through edges that we have removed , so no such path exists .
_ case 3 _ : @xmath76 accesses a vertex gadget @xmath135 which interfaces @xmath132 at a non - red edge set .
this again means that there are @xmath120 distinct colors in @xmath134 , and the @xmath120 edges in @xmath136 and @xmath137 must repeat these . in particular
this means that the edge that repeats the edge of @xmath134 adjacent to @xmath131 must be the edge of @xmath137 adjacent to @xmath132 , and so this edge must be colored a shade of red , which is a contradiction .
_ case 4 _ : @xmath76 accesses a vertex gadget @xmath138 which interfaces @xmath132 at the red edge set .
each time that @xmath134 or @xmath137 traverses a grue edge before the last vertex in @xmath76 that belongs to both @xmath132 and @xmath138 , another edge of the same color must be accessed , by hypercube property ( 3 ) .
since there must be an even number of edges of any color , this means that unrepeated edges in the consistency gadget must be repeated by similar edges in the portion after this last crossover .
also , since all but at most one shade of red has to be traversed twice within the non - consistency - gadget portion , all the red colors together bring the path to within one edge of the border of a vertex gadget .
but in that case there are at least two grue edges in the consistency gadget portion , again pushing the path beyond the radius of the vertex gadget .
thus there is no way for such a @xmath76 to exist .
this means that the resulting graph may be @xmath118-nonrepetitively @xmath117-edge - colored iff the original graph can be 3-edge - colored .
note that there is nothing special about the original graph being of degree 3 .
if the original graph is regular of degree @xmath60 , then we can construct a similar new graph with a vertex gadget being a slice of a @xmath139-hypercube , a consistency gadget being a modified @xmath140-hypercube , and so on .
taking into account the vertex and consistency gadgets , for an original graph with @xmath38 edges , the resulting graph of this reduction has @xmath141 edges .
this means that if @xmath13 depends logarithmically on @xmath38 , then the reduction still runs in polynomial time with respect to @xmath38 .
this raises questions about other versions of the problem where @xmath13 is not constant , but a sub - linear function of @xmath142 .
we now proceed to the main result of the paper , showing that thue number is @xmath21-complete .
we use a reduction based on that of theorem 3 , taking a graph very similar to that produced in theorem 3 and replace the edges with various gadgets depending on their function in that reduction .
we also add constraint gadgets in order to force the edge gadgets that would have been colored the same in theorem 3 to still have similar patterns of colors . after we show that a nonrepetitive coloring of this graph must satisfy a large number of constraints stemming from this local structure , we can use the global structure from theorem 3 to show that an actual nonrepetitive coloring exists if and only if the ancestral instance of co-@xmath923sat is positive .
while this is the idea of the proof , in reality , the local structure is not cleanly separated from the global structure : the format of the latter helps constrain the former .
thue number is @xmath89-complete .
given an instance @xmath143 of co-@xmath923sat , we will refer to the instance @xmath144 of restricted thue number constructed in theorem 3 .
let @xmath145 be the number of colors coloring @xmath5 , @xmath146 , and let @xmath147 be the smallest integer such that @xmath148 and @xmath149 .
let @xmath150 be the two colors corresponding to the universally quantified variable @xmath71
. then we modify @xmath5 by introducing new colors @xmath151 and @xmath152 and replacing each edge colored by @xmath153 by three edges in series colored @xmath154 . clearly , this graph @xmath155 restricted thue number iff @xmath144 is .
we construct from this graph an instance @xmath156 of thue number .
let @xmath157 be the set of colors that are not @xmath158 for @xmath159 .
for every edge @xmath87 of @xmath160 whose color is from @xmath157 , we add to @xmath101 a 7-hypercube @xmath161 with two opposing vertices ( @xmath162 for the one closer to the tip of the snout along a potentially square path and @xmath163 for the other one ) acting as the vertices of the edge gadget . for each color @xmath164 and for each pair
@xmath165 , we also add a consistency gadget @xmath166 ( respectively @xmath167 ) .
this consists of a @xmath168-clique with every vertex identified either with the central vertex of the first two layers of a 7-hypercube , called the _ plume _ emanating from this vertex , or with a vertex of an edge gadget . for each edge
@xmath87 whose color is @xmath169 , two vertices of @xmath161 are identified with vertices of @xmath166 : @xmath170 for one that is distance 3 from @xmath162 and @xmath171 for one that is distance 4 from @xmath162 .
it may be thought of as a sequence of strung - together hypercubes of dimensions 2 , 1 , 1 , and 3 , together with plumes that inflate the degree of some vertices . ]
it may be thought of as two hypercubes of dimensions 3 and 4 together with plumes that inflate the degree of some vertices . ] for the edges colored by subsets of @xmath172 , the choice edge in the snout becomes the gadget depicted in fig .
4 , which we call a _ @xmath104-gadget _ or @xmath173 . the negative edge in the variable gadget becomes the gadget depicted in fig .
5 , which we call an _ @xmath174-gadget _ or @xmath175 . in both cases ,
the asterisked vertex @xmath176 is identified with a vertex of @xmath177 and @xmath162 and @xmath163 are connected to other edge gadgets .
finally , the other edge becomes a _
this is a 7-hypercube attached to the consistency gadget in the same way as @xmath161 , but is modified in several ways . in order to explain this , let us first label the dimensions of the hypercube @xmath180 .
( this labeling will be explained later . )
let @xmath181 span the distance from @xmath170 to @xmath162 , and , correspondingly , let @xmath182 span the distance from @xmath170 to @xmath163 .
let us eliminate the edges @xmath182 adjacent to @xmath163 .
we then nearly saturate the vertices that are separated from @xmath170 by the paths @xmath183 and @xmath184 ( call them @xmath185 and @xmath186 respectively ) with @xmath187 additional edges , but then remove the original edges corresponding to @xmath39 from both and @xmath18 and @xmath60 from @xmath185 and @xmath186 respectively .
now that we have constructed a graph , we will attempt to color it nonrepetitively with @xmath13 colors , and show that it has a nonrepetitive coloring iff the instance of @xmath923sat is negative .
let us call the set of colors @xmath188 and call the color of an edge @xmath87 under a hypothetical coloring @xmath116 .
conversely , we denote by @xmath189 $ ] an edge of color @xmath18 if its more precise identity is irrelevant .
let @xmath190 be the set of edges of color @xmath191 . then in a nonrepetitive coloring
, @xmath192 must all be colored with the same 7 colors .
similarly , for every @xmath27 , @xmath173 , @xmath175 , and the hypercube portion of @xmath179 must all be colored with the same 7 colors .
suppose we have a nonrepetitive coloring @xmath116 .
inside each consistency gadget @xmath193 , each vertex of the @xmath168-clique is saturated .
let @xmath176 , @xmath35 , and @xmath0 be such vertices , and suppose that @xmath194 is the same as the color of the edge of a plume adjacent to @xmath0 .
then there is an edge of color @xmath195 adjacent to @xmath35 .
together with @xmath196 , these four edges then form a square path . to avoid this
, the set @xmath197 of 7 colors that colors the first layer of each plume must be disjoint from the set @xmath198 of @xmath199 colors that colors the clique . by a similar argument
, the second layer of every plume must also have a color in @xmath197 . according to @xcite
, the coloring can be constructed in such a way that @xmath198 consists of the nonzero elements of the group @xmath200 .
( from now on we will refer to just @xmath201 and @xmath202 when @xmath193 is clear from the context . ) by the same token , the edges in the edge gadgets of distance at most two from @xmath203 must be colored with colors from @xmath201 .
now suppose we have an edge @xmath87 that has distance 3 from @xmath203 . if @xmath204 , then the path consisting of @xmath87 and the two edges that connect it to @xmath203 can be duplicated inside the consistency gadget , so @xmath205 .
the maximum distance of an edge in any edge gadget from the closer @xmath203 is 4 .
let us first consider a gadget @xmath161 , and let @xmath206 .
we know that edges in the first three layers from @xmath170 have colors in @xmath201 , so they must be arranged according to the hypercube lemma .
let @xmath207 be separated from @xmath170 by edges @xmath208[c_j][c_k]$ ] .
let @xmath209 be the edge incident to @xmath210 that would be colored @xmath211 if @xmath161 were nonrepetitively colored by @xmath201 , and let @xmath212 be the edge incident to @xmath213 that would be colored @xmath214 .
these edges are adjacent .
now we have four cases , since both @xmath215 and @xmath216 can be either in @xmath201 or in @xmath202 .
suppose that @xmath217 . then if it is not true that @xmath218 , and we do nt have a trivial repetition , then we have a square path of the form @xmath219[c_2]e_4e_3[c_1][c_2 ] [ c(e_4)][c(e_3)]$ ] .
so suppose that @xmath220 .
but we have a square path @xmath221[c(e_4)][c_2][c_4][c_1]e_3e_4[c_2][c_4][c_1]$ ] that starts with two edges of the consistency gadget .
finally , if @xmath222 but @xmath223 , we have the analogous path @xmath224[c(e_3 ) ] [ c_2][c_3][c_1]e_4e_3[c_2][c_3][c_1]$ ] . hence in a nonrepetitive coloring the fourth layer is also colored by @xmath201 according to the hypercube lemma .
since we can prove the same thing starting from @xmath171 , this means the colorings from the two @xmath203s match , and so the entire hypercube is colored in this way .
this is also valid for the @xmath178-gadget .
let us start from @xmath171 .
none of the removed edges has distance less than 4 from this vertex , and six have distance 4 .
since none of them are joined at a vertex whose distance from @xmath171 is 4 , we can apply the previous argument to the ones that are not removed .
this leaves only the edges whose hypercube distance from @xmath170 is at most 3 ; but since no removed edges have distance less than 3 from @xmath170 , this is the same as their actual distance , so we are done .
if the color of the edge in @xmath225 between @xmath170 and @xmath171 is represented in the plume of @xmath226 , there is a square path of length 8 going through @xmath226 , @xmath170 , and @xmath171 .
hence this must be the one color in @xmath202 that is missing from the plume . in the other two gadgets , all the edges except the clump closest to @xmath163
have distance at most 3 from the vertex @xmath176 attaching them to @xmath225 , and thus are known to have colors from @xmath201 .
now , in @xmath227 we have paths @xmath228[b][ab],\tau_1=[c][d][ac][ad ] , \tau_2=[b][c][ab][ac]$ ] , @xmath229[d][ab][ad]$ ] each between @xmath230 and @xmath162 , and paths @xmath231[b][ab],\sigma_1=[c][ac][a],\sigma_2= [ d][ad][a]$ ] each between @xmath230 and @xmath163 .
now , vertex @xmath162 is also vertex @xmath163 of an @xmath232-gadget @xmath233 , and the color of @xmath234 in @xmath5 is @xmath235 .
some path @xmath236[k_1][k_2]$ ] leads us to the vertex @xmath171 of this gadget .
now , this means that we have a path @xmath237\tau_i[k_0][k_1 ] [ k_2][c(\tau_i)][k_0][k_1]$ ] through the two gadgets and @xmath238 , which a nonrepetitive coloring would require to have a loop .
hence in the group @xmath239 we must have @xmath240 .
furthermore , since a possible @xmath241 is @xmath242 , having a loop of size 3 means that @xmath243 or @xmath244 , which contradicts the previous statement .
hence we have @xmath245 , and by the same argument @xmath246 .
by addition we then have the statements @xmath247 , @xmath248 , and @xmath249 .
it is easy to see that we can create no more such identities using colors from @xmath250 by a similar argument , each @xmath251 must form a loop in @xmath252 , which gives us the same six identities . by construction
, the other two gadgets attached to @xmath225 are also preceded by @xmath235 edge gadgets , so each path from @xmath162 to @xmath176 in these gadgets must have one of the three sets of colors @xmath253 , @xmath254 , @xmath255 . now , in @xmath256 , the edges whose distance from @xmath176 is at most 3 are known to be in @xmath257 .
furthermore , the two paths @xmath258[c][a]$ ] and @xmath259[d ] [ a]$ ] , if they were present starting from @xmath176 , would create a square path where an edge adjacent to some @xmath226 of @xmath227 repeats the edge between @xmath170 and @xmath176 .
now , if @xmath36 is the set of colors coloring the cube in @xmath256 , it must be that @xmath260 in @xmath261 .
thus it must be that @xmath262 and that the vertices of the 4-hypercube that are adjacent to @xmath176 and those that are distance 2 away are effectively saturated . hence by the saturation lemma all of the 4-hypercube , except maybe the edges adjacent to @xmath163 , is colored from @xmath263 . but to prevent a square path through @xmath252 , these last edges must also be colored from that set .
we apply a slightly different argument to @xmath264 . to prevent a repetition going through @xmath226 and @xmath171 of @xmath227 , either the set @xmath265 of colors of the paths from @xmath162 to @xmath176
must be @xmath266 , or the edge labeled @xmath6 must be colored @xmath39 .
furthermore , if @xmath267 , to prevent a repetition going through @xmath226 and @xmath170 , the edge labeled @xmath6 must be colored @xmath18 or @xmath60 , and otherwise the edge labeled @xmath43 must be colored @xmath18 or @xmath60 .
the plumes at @xmath70 and @xmath268 clearly can not contain an edge of the same color as @xmath43 , and the plumes at @xmath269 can not contain an edge of the same color as @xmath6 .
thus these vertices are effectively saturated , and we can apply the saturation lemma to the faces of the cube closer to @xmath176 . furthermore ,
if some @xmath270 is the same color as an edge in the square , then we have a square path of length 6 through the corresponding @xmath70 and @xmath271 .
thus except for the edges adjacent to @xmath163 , we know that the paths from @xmath176 to @xmath163 must be colored from the set @xmath272 .
now , by the argument above , @xmath261 must have @xmath273 , and thus so does @xmath274 .
so @xmath275 in @xmath276 .
then to prevent a square path through @xmath252 , the last edges of paths from @xmath176 to @xmath163 must also be colored from @xmath277 . given a nonrepetitive coloring of @xmath101 , for every existentially quantified variable @xmath12
, there is a path from @xmath162 to @xmath163 of @xmath264 that has the same pattern of colors as a path from @xmath162 to @xmath163 of either @xmath256 or @xmath227 , but not both .
let us assume the coloring of @xmath227 that we have implied : we have already shown that this is unique up to permutation .
we have also already shown that given this coloring , @xmath256 must have colors @xmath181 in the cube and @xmath182 in the hypercube , and that @xmath264 must have one of the zero - sum sets of 3 colors on the shorter side and one of the zero - sum sets of 4 colors on the longer side .
if the shorter set is @xmath253 then clearly there is a path through the gadget that is the same as a path through @xmath256 . on the other hand ,
there is not a path that is the same as a path through @xmath227 , since no path through @xmath227 ends in @xmath278 , or @xmath279 .
otherwise the gadget contains either the path @xmath280 or the path @xmath281 .
the removed edges in the @xmath178-gadget are all in either fourth or last position .
now , the edge @xmath114 adjacent to @xmath163 is present .
furthermore , the removed @xmath39-edges are preceded by the sets @xmath282 and @xmath283 , so they are not the @xmath39-edges in these paths . hence
, these paths are present in @xmath227 .
since the first three colors can not be a permutation of @xmath253 , there is no path that is the same as a path through @xmath256 . if @xmath143 , then every coloring of @xmath101 has a square path .
let us assume that we have a nonrepetitive coloring .
it must then agree with claims 1 and 2 . from claim 1 , the path @xmath284 from @xmath162 to @xmath163 exists for all the @xmath232-gadgets .
on the other hand , claim 2 tells us that a nonrepetitive coloring of a @xmath104-gadget is effectively a choice between the corresponding @xmath174-gadget and @xmath178-gadget .
so by choosing a @xmath268 that satisfies @xmath56 for the choice of @xmath12 provided by the coloring , we can create a long square path analogous to the one for @xmath5 .
now , assuming that @xmath285 , we construct a coloring we claim to be nonrepetitive .
let @xmath286 be an assignment to the universally quantified variables such that @xmath287 is not satisfiable .
we set aside 7 colors @xmath288 for each consistency gadget @xmath289 , all of them distinct .
we color the @xmath232-gadgets so that for one , @xmath290 spans the distance from @xmath170 to @xmath162 and @xmath291 spans the distance from @xmath170 to @xmath163 ; for another , @xmath292 and @xmath293 ; and for the third , @xmath294 and @xmath295 .
if @xmath160 has a fourth edge of that color , then we make sure that the coloring of the gadget corresponding to the edge in the snout is different from all the others , but color the fourth edge gadget in one of these three ways . for a consistency gadget that controls a universally quantified variable @xmath296 , we color @xmath297 and @xmath298 in the way already described .
we then color @xmath299 so that it contains the path @xmath300 if @xmath301 and the path @xmath302 if @xmath303 .
let us set aside a set of 7 colors @xmath304 distinct from the ones enumerated so far .
now we need to color the consistency gadgets , which we do using the group method in @xcite .
first , we assign the elements of the subgroup of @xmath200 generated by the first three generators to the set containing each edge gadget s @xmath170 and @xmath171 , whose cardinality clearly is at most 8 .
furthermore , we set the numbers of the vertices so that the sum of the colors of @xmath170 and @xmath171 within an edge gadget does not depend on the edge gadget .
thus the colors of the edges between these vertices are nonzero members of this subgroup , and the edge between @xmath170 and @xmath171 is the same color for each edge gadget .
we assign to these edges the colors @xmath304 , with @xmath305 being the color between @xmath170 and @xmath171 for every @xmath232- and @xmath178-gadget . for a consistency gadget corresponding to a universally quantified variable @xmath71
, we label the vertex @xmath170 of @xmath306 with 0 and the other three @xmath176-vertices with a generator of the subgroup .
this ensures that all the edges that connect two @xmath176-vertices are colored differently .
now , using the rest of the generators , we can find @xmath307 subgroups @xmath308 of @xmath200 isomorphic to @xmath309 , one for each nonzero element @xmath310 , by taking the generators @xmath311 to have @xmath27th digit @xmath312 if @xmath313 and 0 otherwise .
we next arbitrarily assign group elements to the rest of the vertices , and hence edges .
however , we assign colors to the group elements in such a way that for each set @xmath314 we take a distinct subgroup of the kind described and assign @xmath315 to the generators and @xmath316 to the elements @xmath317 , and @xmath318 , respectively .
since the first three digits are zero for all elements of these sets , we have guaranteed that we can never get @xmath319 by adding together colors of edge gadgets .
this coloring is nonrepetitive when restricted to a consistency gadget @xmath193 together with its edge gadgets .
we have already shown that the coloring of each gadget is nonrepetitive . a path within @xmath193 whose colors
are contained in @xmath197 can not reach a vertex at a distance of more than 2 from the interface with colors in @xmath202 .
on the other hand , by construction , no colors in @xmath198 are present in the edge gadgets within distance 2 of @xmath203 .
thus a portion of a square path within an edge gadget must be repeated at least partially in another edge gadget . for an @xmath232-gadget
, this means that one such portion must not be at the end of a path .
thus , since the path is open , it goes through the edge gadget from @xmath170 to @xmath171 , and contains an odd number of edges of each color in @xmath197 .
if some such portion is repeated by sections at both the beginning and end of the path , then one of these sections must be in an @xmath232-gadget , and therefore we can splice the other section onto that one to create another open square path . thus we can build a shorter square open path by replacing each such portion with the edge between @xmath170 and @xmath171 , which we know has the same color @xmath305 for each gadget .
but this substitution produces a square path entirely inside @xmath193 , which must be a loop .
assume now that @xmath193 corresponds to a universally quantified variable .
we have already constructed our coloring so that a path starting from one of the edges adjacent to @xmath226 and continuing to either @xmath104 or @xmath174 can not repeat .
now , we did not color an edge adjacent to @xmath226 by @xmath305 , so a path from @xmath170 to @xmath171 that is repeated by edges within @xmath178 must be at least two edges long . but a path from @xmath226 in colors from @xmath198 may only be one edge long .
now , the set of colors separating @xmath226 and @xmath203 is different from the set of colors separating @xmath320 and @xmath321 . therefore , we can not find a square path that traverses both .
so any square path must contain at least two sections within @xmath193 .
such sections , if they are each between two @xmath176-vertices , can not repeat each other since by our construction the group element leading between each pair of @xmath176-vertices is different .
but such a path also can not use the plumes because any vertex in @xmath174 and @xmath104 has distance at most 4 from @xmath176 , so this together with the plumes would not be enough to repeat a path through @xmath178 containing 7 distinct colors . in a square path ,
colors in edge gadgets can not be duplicated by the same colors in consistency gadgets , except perhaps in one plume .
suppose we have a square path that serves as a counterexample to this , with some sequence of colors occurring in an edge gadget @xmath322 repeated by a sequence in a consistency gadget @xmath323 .
we give the name @xmath76 to this path and @xmath324 to the portion of it that goes through a gadget @xmath325 .
we orient @xmath76 so that @xmath326 comes before @xmath327 ; we therefore refer to the first and second halves of @xmath76 , and to ends and beginnings of sections .
we start with _ case 1 _ : @xmath322 does not belong to @xmath323 and @xmath326 is not at the beginning of @xmath76 .
then the ends of @xmath326 are vertices from the set @xmath328 . by construction ,
the colors of any such path form a loop in @xmath323 , so the path can not be open . thus if @xmath322 does not belong to @xmath323 ,
then @xmath326 is the beginning of @xmath76 .
then the edges following @xmath326 must come either from a consistency gadget @xmath329 ( _ case 2 _ ) or another edge gadget @xmath330 ( _ case 3_. ) _ case 2 _ : in order to link the two half - paths , the first half must eventually leave @xmath329 . if it leaves early , so that an adjacent edge gadget is repeated in @xmath323 , then we are back in case 1 .
if it leaves late , so that it is repeated by an edge gadget adjacent to @xmath329 , then we are also back in case 1 : if the path ended in that edge gadget , then there would be nothing to repeat the edge gadgets between @xmath329 and @xmath323 .
finally , if the vertex @xmath203 at which the path leaves @xmath329 is repeated by a vertex @xmath331 at which the path leaves @xmath323 , then since we are in different consistency gadgets , the edges we traverse have different sets @xmath201 of colors , a contradiction .
_ case 3 _ : if @xmath330 does not belong to @xmath323 , then it must be duplicated inside @xmath329 , since the other option is an edge gadget belonging to @xmath329 , which does not match colors .
therefore , substituting @xmath330 for @xmath322 , we get case 1 and a contradiction
. so @xmath330 belongs to @xmath323 .
this means that the end of @xmath326 is @xmath332 .
furthermore , since @xmath333 must have odd numbers of at least three colors , and since @xmath326 is the beginning of @xmath76 and therefore all of @xmath333 is in the first half , @xmath333 can not be repeated in a plume of @xmath323 .
thus @xmath323 must end with @xmath331 of an edge gadget @xmath334 which repeats @xmath332 .
if @xmath333 ends with @xmath335 , then @xmath336 ends with @xmath337 and we are again in case 1 . on the other hand ,
if @xmath333 ends at @xmath338 , then @xmath339 ends with @xmath340 , and so either we are back in case 1 or the path ends after @xmath339 . but
that would mean that the colors of @xmath326 and @xmath336 would add up in @xmath341 to @xmath319 for some @xmath27 , which contradicts the construction .
the remaining possibility , _ case 4 _ , is that @xmath322 belongs to @xmath323 .
then for colors to match , the part of @xmath327 that repeats @xmath326 must be contained in a plume of @xmath323 and therefore be the end of the path .
no square path in this coloring goes through a consistency gadget . by claim 5 , colors in edge gadgets
must be duplicated by colors in edge gadgets . now , suppose we have a square path @xmath76 that goes through a consistency gadget @xmath193 .
we use the notation of claim 5 to denote portions of this path . on both sides of @xmath342
, we have paths through edge gadgets @xmath334 and @xmath343 .
we now enumerate cases once again .
_ case 1 _ occurs when @xmath336 is a path between @xmath170 and @xmath171 .
then there must be a subpath repeating it , which , by claim 5 , then also goes from @xmath203 to @xmath321 in an edge gadget @xmath344 of @xmath193 .
we can create another square path by replacing @xmath336 and @xmath345 by single edges within @xmath193 , reducing the problem to any of the other cases .
so we can assume that both @xmath336 and @xmath346 go to a @xmath332 .
then if we orient the path , one subpath goes from an edge gadget to @xmath193 and the other goes from @xmath193 to an edge gadget .
therefore , by claim 5 , one can not be the repetition of the other .
thus @xmath76 must traverse @xmath193 twice and associated edge gadgets four times , at least three of these times passing between a @xmath203 and a @xmath332 .
( the fourth time it does not have to do this since a traversal may be split between the beginning and the end of @xmath76 . )
furthermore , the edge subpaths must be in two pairs , each of which has one set of colors , and at least one of the pairs must have both its subpaths continue in other edge gadgets .
suppose now that @xmath193 is not associated with a universally quantified variable .
_ case 2 _ occurs when the two subpaths are in the same @xmath232-gadget .
since no edge in @xmath160 is adjacent to edges of one color through both incident vertices , the colors of the neighboring edge gadgets must be different and the path can not be square .
in _ case 3 _ , the two subpaths are in different @xmath232-gadgets .
in order for them to be colored the same , we must assume that @xmath193 controls four @xmath232-gadgets and neither of the @xmath232-gadgets traversed by the subpaths corresponds to an edge in the snout of @xmath160 .
but the edges not in the snout all border different colors , so the path once again can not be square . _
case 4 _ : we are left with the possibility of @xmath347 being associated with a universally quantified variable @xmath296 .
we colored @xmath347 in such a way , however , that the sets of paths within it between every pair of @xmath176-vertices are disjoint . therefore ,
if we orient @xmath76 in a certain way , the second @xmath348 must be at the end of the path . by the same token
, @xmath76 must begin with a section in one of the edge gadgets belonging to @xmath347 .
all the @xmath162s of the edge gadgets of @xmath347 are incident to @xmath232-gadgets of colors @xmath349 and all the @xmath163s are incident to @xmath232-gadgets of colors @xmath350 .
hence paths to @xmath163 can be repeated by paths to @xmath163 and paths to @xmath162 can be repeated by paths to @xmath162 . by construction ,
the possible repetitions of paths between @xmath331 and @xmath332 are then 1 . if @xmath303 , a path between @xmath162 and @xmath176 in @xmath297 with a path between @xmath162 and @xmath170 in @xmath298 2 . if @xmath301 : 1 . a path between @xmath162 and @xmath176 in @xmath299 with a path between @xmath162 and @xmath176 in @xmath297 and a path between @xmath162 and @xmath170 in @xmath298 2 . a path between @xmath163 and @xmath176 in @xmath299 with a path between @xmath163 and @xmath176 in @xmath297 .
in each case , therefore , there can be only one such repetition in an open path .
since gadgets belonging to @xmath347 occupy both the beginning and the end of @xmath76 , there may be no traversal of consistency gadgets other than @xmath347 within @xmath76 .
a path through @xmath297 can not repeat a path through @xmath298 since these paths would have to go through edge gadgets corresponding to different direction - determining colors of @xmath160 .
also , a path going through @xmath162 of @xmath299 is going towards the tip of the snout , and therefore can not meet any more gadgets belonging to @xmath296 .
this gives us a contradiction if @xmath162 is part of one of the square sections .
thus we are left with possibility 2(b ) .
this means that there must be a subpath consisting only of edge gadgets from @xmath163 of @xmath299 to some other @xmath332 of an edge gadget belonging to @xmath347 .
if the path is from @xmath299 , then it must traverse an edge gadget belonging to @xmath351 , which must be repeated by an edge gadget at the tip of the snout , which @xmath76 can not reach .
this restricts any potential square paths to going through a chain of edge gadgets . but
this means that a square path maps to a square path in @xmath160 , which does not exist if @xmath352 .
hence this is a reduction .
it is easy to see that the reduction runs in polynomial time , specifically @xmath353 .
we would like to thank chris umans for his mentoring and helpful suggestions at every stage of the research ; and grigori mints and yuri manin for reviewing the proofs .
9 n. alon , j. grytczuk , m. hauszczak , and o. riordan .
`` nonrepetitive colorings of graphs . '' in m. kronski , j. spencer , and a. ruciski , editors , _ proceedings of the 10th international conference on random structures and algorithms ( rs&a-01 ) _ , volume 21 , 34 of _ random structures and algorithms _ , pages 336346 , danvers , ma , aug .
610 2002 .
wiley periodicals .
j. bart and p.p .
`` on square - free vertex colorings of graphs . '' _ studia scientiarum mathematicarum hungarica _
( to appear ) .
j. bart and p.p .
`` on square - free edge colorings of graphs . '' _ ars comb . _
( to appear ) .
b. brear and s. klavar .
`` square - free colorings of graphs . '' _ ars comb . _ 70:313 ,
jan . 2004 .
thomas f. coleman and jorge j. mor .
`` estimation of sparse hessian matrices and graph coloring problems . '' _ math .
programming _ , 28(3):243270 , 1984 . s. czerwiski and j. grytczuk .
`` nonrepetitive colorings of graphs . ''
_ electronic notes in discrete mathematics _ 28:453459 , 2007 . j. grytczuk .
`` nonrepetitive colorings of graphs a survey . '' _ international journal of mathematics and mathematical sciences _ , 2007 .
ian holyer .
`` the np - completeness of edge - coloring . ''
_ siam j. comput .
_ 10:718720 , 1981 .
dniel marx and marcus schaefer .
`` the complexity of nonrepetitive coloring . '' _ depaul university technical report _
, tr 07 - 007 , 2007 .
marcus schaefer and christopher umans .
`` completeness in the polynomial - time hierarchy : a compendium '' ( updated version ) .
_ sigact news _ ,
september 2002 . retrieved from <
http://ovid.cs.depaul.edu / documents / phcom.ps>. marcus schaefer and christopher umans .
`` completeness in the polynomial - time hierarchy : part ii '' ( updated version ) . _ sigact news _ , december 2002 . retrieved from < http://ovid.cs.depaul.edu / documents / phmeth.ps>. axel thue .
`` ber unendliche zeichenreihen . ''
_ norske vid .
christiana 7 _ , 1 - 22 , 1906 . reprinted in nagell , t. ; selberg , a. ; selberg , s. ; and thalberg , k. ( eds . ) .
_ selected mathematical papers of axel thue_. oslo , norway : universitetsforlaget , pp .
139 - 158 , 1977 . | a squarefree word is a sequence @xmath0 of symbols such that there are no strings @xmath1 , and @xmath2 for which @xmath3 .
a nonrepetitive coloring of a graph is an edge coloring in which the sequence of colors along any open path is squarefree .
the thue number @xmath4 of a graph @xmath5 is the least @xmath6 for which the graph can be nonrepetitively colored in @xmath6 colors . a number of recent papers have shown both exact and approximation results for thue numbers of various classes of graphs .
we show that determining whether a graph @xmath5 has @xmath7 is @xmath8-complete .
when we restrict to paths of length at most @xmath6 , the problem becomes * np*-complete for fixed @xmath6 . for @xmath9 ,
this is the edge coloring problem ; thus the bounded - path version can be thought of as a generalization of edge coloring . |
looking for exotic states in transition - metal oxides is a fascinating problem .
spin - orbital models arose from the consideration of orbital degrees of freedom of @xmath1- or @xmath2-electron in transition metals.@xcite when hund s coupling , orbital anisotropy , jahn - teller effect , etc .
are ignored , a @xmath0 spin - orbital model is recurrently proposed and studied in connection with real materials , such as c@xmath3 material,@xcite spin - gap materials na@xmath4ti@xmath4sb@xmath4o and nav@xmath4o@xmath5,psk , fujii and cubic vanadates lavo@xmath6 and yvo@xmath6.@xcite in this paper , we employ the schwinger boson mean - field theory ( sbmft ) to study the low temperature properties of the generalized one - dimensional ( @xmath7 ) @xmath8 spin - orbital model and reveal that the dimer phases in fact consist of three kinds of valence bond phases : orbital valence bond solid ( ovb ) state , spin valence bond solid ( svb ) state and spin - orbital valence bond solid ( sovb ) state .
all of the three phases are gapped , but they can be well characterized by the spin and pseudospin susceptibilities , and thus are distinguishable experimentally . as we will reveal , the @xmath8 type of spin - orbit coupling is responsible for both the spin - peierls and orbital - peierls phenomena occurring in this @xmath7 model .
this paper is organized as follows . in section
ii , we give a detailed schwinger boson mean - field scheme on the spin - orbital model .
then in section iii , we study the phase diagrams for two cases : ( i ) @xmath9 and @xmath10 ; ( ii ) @xmath11 and @xmath10 .
the phase diagram for the former is novel while the one for the latter is a supplementary for previous works.@xcite in section iv , we deduce the spin and pseudospin susceptibilities . at last
, a brief conclusion is given in section v.
the one - dimensional spin - orbital hamiltonian with the @xmath0 symmetry reads@xmath12 where @xmath13 and @xmath14 are two sets of spin operators which satisfy the @xmath15 algebra with eigenvalues @xmath16 and @xmath17 , respectively , and @xmath18 and @xmath19 are two constant parameters
. it can also be viewed as a generalized spin ladder system with four - operator interactions .
there are two special points in this model .
the first is the _ dimer point _
@xmath20 : @xmath21 , whose ground state is perfectly dimerized and can be recognized as a two - fold degenerate dimerized valence bond solid.@xcite another point is the _ fm point _
@xmath22 : @xmath23 , whose ground state possesses the maximal values for both @xmath24 and @xmath25 , and is , we will see in the phase diagrams later , the critical point of three uniform phases .
the model for @xmath26 and @xmath10 was studied both analytically and numerically , and the phase diagrams were proposed by several authors.@xcite except for the phases with conventional spin or orbital ferromagnetic ( fm ) _ _ _ _ or antiferromagnetic ( afm ) orders , the mixed spin - orbital valence bond state , which exhibits no long - range correlation and no gap near the @xmath27 symmetric point @xmath16 : @xmath28,@xcite appears in a large regime in the phase diagram .
the model for @xmath9 and @xmath10 around the point @xmath29 : @xmath30 was revealed to exhibit the ovb state.@xcite in this paper we will employ sbmft to study the phase diagram , and focus on exploring the dimer phases .
sbmft had been introduced to the field of strongly correlated systems for a long time , and shows its merit in describing the spin systems at low temperatures .
it can produce spin liquid phase and spin gap close to the ground state as well as the phases with long - range correlation.@xcite in schwinger boson representation , one can introduce two boson operators @xmath31 and @xmath32 to represent a spin operator @xmath33 : @xmath34@xmath35 with the local constraint @xmath36 in this way one can introduce two types of bond operators , @xmath37 to mimic the ferromagnetic ( fm ) and antiferromagnetic ( afm ) channels of the spin @xmath16 heisenberg interactions .
the low energy physics can be well captured by the mean fields , @xmath38 or @xmath39 , where @xmath40 stands for the thermodynamic average .
likewise , the mean fields for the pseudospin @xmath17 are denoted by @xmath41 and @xmath42 for the fm and afm channels , respectively . as revealed by ceccatto _
et al_.,@xcite retaining both the fm and the afm channels will lead to @xmath43 where : @xmath34 : denotes normal order . in the mean field scheme ,
the bond operators are decomposed as@xmath44 where the higher - order terms , @xmath45 and @xmath46 , are ignored .
we shall avoid applying the identity , @xmath47 which holds due to the constraint eq .
( [ constraint ] ) and implies @xmath48 the identity , eq .
( [ identity ] ) , is largely violated when the constraint is imposed only on average . and this is also true even when the gaussian - fluctuation corrections are considered.@xcite to show how eq .
( [ cd2 ] ) works , we present a two - site example , @xmath49 , which can be solved analytically . after solving it by sbmft , one can find , in the fm ( @xmath50 ) case , eq .
( cd2 ) gives ground energy @xmath51 with @xmath52 and @xmath53 , and eq .
( [ wd ] ) gives the same ground energy with @xmath54 .
they both are in agreement with the exact solution . in the afm ( @xmath55 )
case , eq .
( [ cd2 ] ) gives @xmath56 with @xmath57 and @xmath58 , while eq .
( [ wd ] ) gives @xmath59 with @xmath57 .
the former is just the exact energy ; the latter approaches the exact value only in the limit @xmath60 .
( [ wd ] ) overestimates the ground energy in afm channel .
we have also found that eq .
( [ cd2 ] ) , instead of eq .
( [ wd ] ) , can produce the exact energy of the dimer point in the spin - orbital model of eq .
( [ h ] ) .
thus for a small @xmath16 , the mean field scheme in eq .
( [ cd2 ] ) is better than that in eq .
( [ wd ] ) . in practice ,
when solving the mean field equations , we found that only one of the fm and afm channels can survive on a bond meanwhile for @xmath7 and @xmath61 unfrustrated systems . to perform the mean field calculation , we decompose eq .
( [ h ] ) into the spin @xmath16 and the pseudospin @xmath17 chains , @xmath62 \notag \\ & + \sum\limits_{m}\left [ yj_{t}(m)+xj_{s}(m)+j_{s}(m)j_{t}(m)\right ] \label{hm}\end{aligned}\ ] ] where the effective couplings [ ij ] @xmath63the two chains are coupled by the effective couplings , which will be determined self - consistently . and
( [ cd2 ] ) ensures a reasonable estimation of the strength of the effective couplings , eq .
( [ ij ] ) , for both of the fm and afm channels .
one should notice , for @xmath64 , the mixed channel of @xmath16 and @xmath17 must be considered around the @xmath27 point @xmath65,@xcite which is not the focus of this paper . in the following
we apply the bond operators in eq .
( [ cd2 ] ) for both spin @xmath16 and pseudospin @xmath17 in eq .
( [ hm ] ) .
there are four combinations of uniform fm and afm phases : ( 1 ) @xmath16-fm @xmath66 , @xmath17-fm @xmath67 ; ( 2 ) @xmath16-fm @xmath68 , @xmath17-afm @xmath69 ; ( 3 ) @xmath16-afm @xmath70 , @xmath17-fm@xmath71 ; ( 4 ) @xmath16-afm @xmath72 , @xmath17-afm @xmath73 .
the inequalities in the parentheses should be fulfilled when the mean fields are solved .
the four phases correspond to four separate regions in @xmath74 plane .
we omit details for these uniform phases .
let us focus on more interesting cases .
it was pointed out by shen _
_ that ovb state should be a preferred ground state around the point @xmath29 : @xmath75 for @xmath9 and @xmath10.@xcite this dimerization effect is also confirmed by the tmrg calculation of orbital - orbital correlation function,@xcite and is also thought to be responsible for the observation on the neutron scattering experiment of yvo@xmath6.@xcite in fact , the generalized model of eq .
( [ h ] ) provides two other valence bond solid states , the spin valence bond solid ( svb ) and spin - orbital valence bond solid ( sovb ) .
the three kinds of dimer phases have not been studied and distinguished within an unified framework in the previous works . to approach the dimer phases , we has two schemes as depicted in fig .
1(a ) and ( b ) .
it turns out that the scheme ( b ) is only appropriate for exploring the sovb phase and does not produce the ovb and svb phases .
so we shall focus on the scheme ( a ) .
one can find , for the scheme ( a ) , the spin @xmath16 chain and the pseudospin @xmath17 chain are the same object mathematically : _ dimerized chain with alternating fm and afm bonds_. for the spin @xmath16 part , we can divide it into two sublattices @xmath76 and @xmath77 by assuming that the translational invariance is broken simultaneously .
in fact , the period length of the chain is doubled due to the spin - peierles phenomenon .
we use two sets of schwinger bosons @xmath78 and @xmath79 to represent spins on the sublattices @xmath76 and @xmath77 respectively and choose the unit cell to be composed of two lattice sites with one for sublattice @xmath76 and another for sublattice @xmath80 .
then the fourier transformations read @xmath81 supposed nonzero real mean fields are @xmath82 the lagrangian multipliers , @xmath83 and @xmath84 , are imposed to realize the constraints on the spin @xmath16 by adding the term in the total hamiltonian , @xmath85 for a mean field approach we take @xmath86 after the fourier transform , we introduce the nambu spinor in the momentum space , @xmath87 one can arrive at the mean - field hamiltonian for the spin @xmath16 chain in compact form , @xmath88 where @xmath89 -2n\lambda _ { s}(2s+1),\ ] ] and the matrix @xmath90 is constructed by the kronecker products of the pauli matrices , @xmath91 with @xmath92and @xmath93
. then we can read out the spectra from the poles of the bosonic matsubara green s function ( @xmath94 matrix ) , @xmath95 where @xmath96 ( @xmath97 is an integer and @xmath98 is the inverse temperature ) .
similarly we repeat the mean field calculation on the pseudospin @xmath17 chain . both the spin and the pseudospin chains have two spectra , which read @xmath99 where @xmath100 , @xmath101 , and @xmath102 .
the lower quasiparticle spectrum @xmath103 is found to be gapped at @xmath104 , which is related to the physical gap by the peak of the imaginary part of the dynamic spin susceptibility.@xcite and the gap is also true for half integer spin ( or pseudospin).@xcite by optimizing the total free energy @xmath105 , \ ] ] where @xmath106 is the bose - eienstein distribution and @xmath107,\end{aligned}\ ] ] a set of the mean field equations are established , @xmath108 where we have defined integrals ( the symbols with wave decoration means dimensionless quantities , @xmath109 , and we have substituted the sum with integral : @xmath110 @xmath111 ) , @xmath112
; ( b ) @xmath113 .
the uniform phases are : i for s - fm and t - fm ; ii for s - afm and t - fm ; iii for s - fm and t - afm ; iv for s - afm and t - afm .
the su(4 ) gapless region is quoted from ref .
@xcite the special points are : @xmath22 - the fm point ; @xmath20 - the dimer point , @xmath16 - the su(4 ) symmetric point , and @xmath29 - the representative point of ovb phase
. please see details in the text.,width=302 ] for the transition - metal oxides , the pseudospin is usually one half , i. e. @xmath114 , reflecting two choices of unfrozen @xmath115 or @xmath116 orbitals.tn,kho spin - gap materials , na@xmath4ti@xmath4sb@xmath4o and nav@xmath4o@xmath5 , are related to the @xmath26 case in the model of eq .
( [ h]),psk , fujii while cubic vanadates , lavo@xmath6 and yvo@xmath6 , are related to @xmath117 case due to large hund s coupling.@xcite the phase diagrams are obtained by solving the mean field equations numerically at zero temperature .
2(a ) shows the case for @xmath9 and @xmath114 .
we found three dimer phases : ( 1 ) the ovb phase with @xmath118 and @xmath119 ; ( 2 ) the svb phase with @xmath120 and @xmath121 ; ( 3 ) the sovb phase with @xmath122 and @xmath58 .
as expected , the point @xmath29 : @xmath75 lies in the ovb phase region . at the mean field level ,
the dimer phases are captured by the staggered non - zero afm mean fields @xmath123 and/or @xmath124 , which provide a perfect dimer picture of spin and/or pseudospin in the whole phase regions .
the reason for this maybe lies in two : the mean field treatment omits part of quantum fluctuations ; the dimerization effect is very strong in such a system at least at zero temperature .
in fact , tmrg has confirmed the orbital dimerization at the point @xmath29 : @xmath125 since the correlation function @xmath126 extrapolates to @xmath127 per bond at zero temperature , although this point is not exactly soluble.@xcite and the ovb phase is quite robust even when the anisotropy , hund s coupling , and atomic spin - orbit interaction are taken into account for cubic vanadates and relevant systems.@xcite we also guess these dimer phases may survive in two dimensions , although no rigorous soluble point can be referred to . a dimerized ovb configuration in two dimensions has been proposed and checked by the spin - wave theory in ref .
@xcite , which indeed exhibits lower energy than the uniform phases .
however it is an interesting problem whether sovb can survive in two dimensions .
2(b ) shows the case for @xmath26 and @xmath10 . in fig .
2(b ) , the @xmath27 gapless region is quoted from ref .
we found the svb and ovb phases still occupy two areas that are sandwiched between the sovb and @xmath27 gapless region .
the three dimer phases with corresponding mean fields are : ( 1 ) the ovb phase with @xmath128 and @xmath119 ; ( 2 ) the svb phase with @xmath129 and @xmath58 ; ( 3 ) the sovb phase with @xmath130 and @xmath58 .
previous works ignored the svb and ovb phases.@xcite and there is a @xmath131-afm and @xmath17-afm phase region from the point of view of sbmft .
the result here in fact suggests that the spin - orbit coupling of @xmath0 type plays the roles of the intrinsic spin - peierls and orbital - peierls mechanisms .
the svb formation in this @xmath7 model is an example that expresses the concept of _ orbitally driven peierls state_.@xcite but notice we did not consider the lattice distortion here . for spin @xmath9 , the svb phase is harder to realize since the region shrinks . and @xmath10 at the ovb point
@xmath29 : @xmath132 ( solid line ) and the dimer point @xmath20 : @xmath133 ( dashed line).,width=283 ]
the spin structures of the dimer phases are detectable by spin susceptibility .
let us consider the spin susceptibility ( the psedospin case is handled in the same way ) .
we can define the spin density waves on sublattices @xmath76 and @xmath77 ( due to the rotational symmetry , we only consider @xmath134 component of the spin ) , @xmath135 the spin susceptibility contains contributions from both the intra - sublattice and inter - sublattice fluctuations , @xmath136 in matsubara formalism , the intra - sublattice and inter - sublattice contributions read @xmath137 , \\ \chi_{ab}\left ( q,\tau\right ) & = \left\langle s_{a}^{z}\left ( q,\tau\right ) s_{b}^{z}(-q,0)\right\rangle \notag \\ & = \frac{1}{4}\int_{-\pi}^{\pi}\frac{dk}{2\pi}[2g_{57}^{s}\left ( -k,\tau\right ) g_{13}^{s}\left ( k+q,\tau\right ) \notag \\ & \qquad\qquad - g_{58}^{s}\left ( -k,\tau\right ) g_{14}^{s}\left ( k+q,\tau\right ) \notag \\ & \qquad\qquad - g_{23}^{s}\left ( -k,\tau\right ) g_{67}^{s}\left ( k+q,\tau\right ) ] , \end{aligned}\ ] ] respectively , where the green s function in imaginary time @xmath138 is related to eq .
( [ green ] ) by @xmath139 the numerical solutions of the static spin susceptibilities , @xmath140 and @xmath141 , for @xmath9 and @xmath10 at the ovb point @xmath29 : @xmath132 and the dimer point @xmath20 : @xmath133 are shown in fig .
3 . at the ovb point
@xmath29 , the static spin susceptibilities exhibit finite sharp peaks at @xmath142 for @xmath143 and @xmath144 for @xmath145 reflecting the strong afm fluctuation among sites in sublattice @xmath76 ( or @xmath77 ) and fm fluctuation between sites in sublattice @xmath76 and sites in sublattice @xmath77 .
the finite peaks also indicate no long - range order exists at zero temperature . while there is no sharp peak at the dimer point @xmath20 , since the formation of spin singlets leads to dispersionless spin spectra , @xmath146 .
the spin susceptibilities at the ovb point @xmath147 and the dimer point @xmath20 characterize the whole ovb region and sovb region respectively .
as we go from ovb phase to svb phase , the spin and pseudospin change their roles . in sovb phase
, the sharp peaks will disappear for both spin and pseudospin .
thus the three dimer phases are well characterized by the spin and pseudospin susceptibilities and distinguishable experimentally .
a recent neutron scattering experiment of yvo@xmath6 has provided a tentative proof of the ovb.@xcite
in conclusion , we have employed sbmft to study the valence bond states as well as uniform fm and afm states in the @xmath7 @xmath0 spin - orbital model .
the phase diagrams are constructed by comparing the ground energies of the proposed possible states . in the valence bond
states the translational invariance is broken in space , fm and afm parameters compete with each other and are determined self - consistently .
specifically , two sublattices are introduced to approach the three dimerized svb , ovb and sovb phases by sbmft .
we clarified how to capture properly the low temperature properties of such a system by sbmft .
the main result of the paper is that the svb and ovb have lower energy than the sovb in some regions .
these consequences reveal that the @xmath0 type spin - orbit coupling can provide both spin - peierls and orbital - peierls mechanisms .
the results still should be compared with other phases including those with translational invariance .
static spin and pseudospin susceptibilities had been calculated in the present theory and are available to distinguish spin - singlet and orbital - singlet formations in real materials . | a generalized one - dimensional @xmath0 spin - orbital model is studied by schwinger boson mean - field theory ( sbmft ) .
we explore mainly the dimer phases and clarify how to capture properly the low temperature properties of such a system by sbmft .
the phase diagrams are exemplified .
the three dimer phases , orbital valence bond solid ( ovb ) state , spin valence bond solid ( svb ) state and spin - orbital valence bond solid ( sovb ) state , are found to be favored in respectively proper parameter regions , and they can be characterized by the static spin and pseudospin susceptibilities calculated in sbmft scheme .
the result reveals that the spin - orbit coupling of @xmath0 type serves as both the spin - peierls and orbital - peierles mechanisms that responsible for the spin - singlet and orbital - singlet formations respectively . |
ngc 1866 is one of the most largely studied stellar clusters in the large magellanic cloud ( lmc ) . since the first photometric study by @xcite , who discovered a large number of member cepheid variables , this cluster has been recognized as an ideal laboratory for testing stellar evolutionary models of intermediate - mass stars . ngc 1866 is one of the most populous lmc clusters , having a mass comparable to that of a typical globular cluster in the galaxy ( @xmath14 * ? ? ?
* ) and is relatively young to study the evolution of intermediate - mass ( @xmath15 ) stars .
these stars develop a convective core during the h - burning phase and represent the ideal target for ana analysis of the real extension of the convective cores , i.e. , to see whether the size of convective cores , as determined by the schwarzschild criterion , to which we will refer as `` standard '' , must be `` artificially '' increased in order to fit the observational data . this is a long - standing problem whose solution would deeply influence many aspects of our present understanding of the universe : in fact it would fix a key parameter in the computation of the stellar evolution which would reflect , e.g. , on the intermediate age clusters , on the determination of the various mass limits , like @xmath16 and @xmath17 ( see * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) and on the initial - final mass relation , on the relation between he - core mass and initial mass for massive stars that is linked to the problem of metal enrichment .
ngc 1866 has often been considered a `` proof '' for the existence of the convective overshooting @xcite .
actually the fit to this ( or another ) cluster can only give us information on the likely size of the convective core but it ascertain the mechanism which determines the size . for reasons sinking in the history of the computation of stellar models in the range of massive stars (
whose analysis goes well beyond the purposes of this paper ) , this problem has always been ( unfortunately ! ) , interpreted in terms of `` overshooting '' , therefore as a proof of the existence or non = existence of a specific phenomenon where the convective cells cross the classical border of the convective core and mix more matter than predicted in the `` classical '' scenario , thus producing a larger amount of fuel .
this direct connection is meaningless : in fact , if the observations showed that the convective core must be larger than currently predicted by the models , various causes could be considered .
just to mention two possibilities other than the``overshooting '' , the simple change from the old cox - tabor opacities @xcite to the los alamos ones @xcite , which occurred at the beginning of the 80 s , led , by itself , to a significant increase of the classical size of the convective core .
a second possibility , proposed for example by a. maeder ( private communication ) is a rotationally induced mixing that leads to a more massive convective core .
@xcite were the first to point out that the `` standard '' models could not fit the observed ratio of main sequence vs he clump stars @xmath18 : they found that their standard models predicted at least 4 times more giant stars than observed with respect to the main sequence ones . in the following years , ( *
* hereafter c89 ) and ( * ? ? ?
* hereafter b89 ) , studied in detail the properties of the stellar content of ngc 1866 , using ccd data . a detailed discussion of the general problem as well as the results of c89 and b89 are given in @xcite .
briefly , b89 concluded that the standard models reproduce both the color - magnitude diagram ( cmd ) morphology and the luminosity function ( lf ) of the cluster , while c89 concluded that models including a certain amount of overshooting were required .
the problem has been further discussed in later papers ( * ? ? ?
* ; * ? ? ?
* hereafter bcp94 ) without reaching an agreement . a larger sample of data led bcp94 to conclude that their previous set of data was substantially correct , whereas the ones adopted by c89 were affected by stochastic fluctuations .
actually , the sample adopted by c89 included 554 ms stars ( down to @xmath19 ) and 41 he burning giants , while the latest set of data presented by bcp94 included ( down to the same magnitude ) 462 ms stars and 65 giants .
all these numbers were already corrected by the authors for completeness , crowding and field stars contamination . in order to contribute to clear up the situation
, we began a project specifically devoted to a complete photometric study of the stellar population in ngc 1866 , including a substantial part of the ms . in this paper , we present a new deep ( down to @xmath0 ) photometry of a wide region @xmath1 around the cluster .
the observations and reduction procedures are described in sect .
2 , while the morphology of the inferred cmd is presented and discussed in sect .
4 is devoted to the discussion of the derived lf , with a special attention to the field decontamination and the completeness estimate and correction , together with the comparison with previous studies . the comparison with various theoretical scenarios
is discussed in sect .
5 . a final section with summary and conclusions follows .
lcc fa0122 * & v & 30 _ s _
+ fa0124 & b & 60 _ s
_ + fa0125 & b & 60 _ s _
+ fa0127 & v & 15 _ s _
+ fa0128 & v & 15 _ s _
+ fa0130 & b & 60 _ s _
+ fa0131 & v & 5 _ s _ + fa0133 & b & 15 _ s _
+ fa0134 & v & 480 _ s _
+ fa0136 & b & 120 _ s _
+ fa0139 & b & 1500 _ s _ + fb0140 & v & 30 _ s _
+ fb0142 & b & 90 _ s _
+ fb0143 & b & 1500 _ s _
+ fb0144 * & v & 480 _ s _
the data have been obtained at the 2.2 m eso - mpi telescope at la silla ( chile ) , on jan .
27th , 1993 .
we used efosc2 equipped with the ccd # 19 thomson of size @xmath20 pixels , with a scale of @xmath21/pix , which yields a total field of view of @xmath1 . during the night
we secured b , v frames in two fields around ngc 1866 : @xmath22 , roughly centered on the cluster center and @xmath23 , located at @xmath24 se from the cluster center and observed for statistical field decontamination purposes .
figure [ fig1 ] shows a computer map of the two fields .
a circle is drawn at @xmath25 from the cluster center and encloses the region dominated by stars belonging to the cluster . according to @xcite , at larger distances the contribution of the cluster stars to the total population of the field
is expected to drop below 1% .
the average observational conditions were good , and the seeing was around @xmath26 .
a journal of the observations is reported in table [ obs ] .
the reductions were performed using daophot @xcite , implemented on iraf . in order to identify stars within each frame
, we followed the standard procedure as described on the daophot manuals , and used the routine allstar to perform the final psf - fit .
each frame was reduced independently , then a frame in each filter has been chosen as _ reference _ frame for coordinate and magnitude transformations ( reference frames are marked with an asterisk in table [ obs ] ) ; then all the frames have been transformed to the reference system in order to get a homogeneous set of instrumental magnitudes and coordinates .
a set of averaged instrumental magnitudes ( weighed according to the photometric quality of the frames ) have then been obtained for each field .
the overlap region between the two adjacent fields has been used to transform the coordinates into a common local system . in this arbitrary system
, the center of the cluster has been estimated at pixel ( 520,520 ) in @xmath22 ( see figure [ fig1 ] ) .
since the observations were performed during slightly non - photometric conditions , we used the well - calibrated photometric sequence of standard stars obtained by @xcite in the surrounding region of ngc 1866 , to link the instrumental magnitudes to the standard johnson system .
the cross - correlation between our data set and walker s own identified 2294 stars in common , which were used to calibrate the final catalog .
the list of the final calibrated b , v magnitudes and position for 12680 ( @xmath22 ) + 3044 ( @xmath23 ) stars identified in our survey is available in electronic form .
table [ clu ] reports a sample of the table of photometry of @xmath22 while table [ fie ] reports the analogous for @xmath23 .
figure [ fig2]a shows the cmd for stars in @xmath22 ( hereafter referred to as cluster sample ) ; in the same way , figure [ fig2]b shows the cmd for stars lying at @xmath27 from the cluster center , which are assumed to be representative of the field population around the cluster ( hereafter referred to as field sample ) .
lcccccc 1 & 170.77 & 0.86 & 19.68 & 0.01 & 0.01 & 0.03 + 2 & 859.86 & 1.49 & 20.31 & 0.02 & 0.09 & 0.03 + 3 & 867.64 & 8.95 & 22.23 & 0.07 & 0.45 & 0.10 + 4 & 528.30 & 1.72 & 18.91 & 0.01 & -0.05 & 0.01 + 5 & 466.28 & 1.87 & 22.48 & 0.07 & 0.61 & 0.09 + 6 & 611.78 & 2.28 & 22.78 & 0.09 & 0.32 & 0.13 + 7 & 57.47 & 2.49 & 21.58 & 0.04 & 0.47 & 0.07 + 8 & 479.96 & 2.62 & 20.72 & 0.01 & 0.10 & 0.02 + 9 & 942.88 & 2.92 & 22.65 & 0.09 & 0.95 & 0.16 + 10 & 851.31 & 3.07 & 21.45 & 0.05 & 0.24 & 0.06 + 11 & 368.49 & 3.28 & 22.90 & 0.10 & 0.11 & 0.13 + 12 & 129.26 & 8.74 & 21.33 & 0.02 & 0.21 & 0.04 + 13 & 25.95 & 3.77 & 22.36 & 0.09 & 0.54 & 0.12 + 14 & 295.52 & 5.51 & 21.49 & 0.03 & 0.35 & 0.04 + 15 & 444.89 & 5.91 & 19.03 & 0.00 & -0.07 & 0.01 + 16 & 104.19 & 5.99 & 22.64 & 0.09 & 0.48 & 0.13 + 17 & 224.52 & 6.15 & 22.27 & 0.04 & 0.47 & 0.08 + 18 & 319.41 & 11.00 & 22.39 & 0.06 & 0.54 & 0.09 + 19 & 180.74 & 6.75 & 23.20 & 0.13 & 0.73 & 0.20 + 20 & 383.21 & 9.18 & 21.94 & 0.03 & 0.32 & 0.05 + lcccccc 1 & 58.44 & 1.00 & 19.71 & 0.02 & 0.66 & 0.07 + 2 & 175.82 & 1.20 & 22.03 & 0.04 & 0.33 & 0.11 + 3 & 572.97 & 1.74 & 21.06 & 0.04 & 0.22 & 0.05 + 4 & 5.74 & 1.65 & 22.75 & 0.16 & 1.59 & 0.50 + 5 & 544.28 & 4.60 & 21.57 & 0.03 & 0.23 & 0.05 + 6 & 444.44 & 2.68 & 21.75 & 0.04 & 0.27 & 0.05 + 7 & 457.46 & 2.71 & 21.20 & 0.01 & 0.34 & 0.03 + 8 & 366.48 & 4.05 & 21.56 & 0.03 & 0.53 & 0.05 + 9 & 747.11 & 5.92 & 21.79 & 0.05 & 0.34 & 0.07 + 10 & 109.36 & 4.62 & 21.38 & 0.02 & 0.19 & 0.04 + 11 & 990.28 & 12.23 & 21.20 & 0.06 & 0.11 & 0.11 + 12 & 333.27 & 5.92 & 21.19 & 0.02 & 0.28 & 0.03 + 13 & 203.43 & 5.92 & 20.47 & 0.02 & 0.35 & 0.03 + 14 & 561.95 & 5.95 & 22.53 & 0.06 & 0.30 & 0.09 + 15 & 401.76 & 6.94 & 21.36 & 0.04 & 0.30 & 0.05 + 16 & 400.46 & 12.52 & 19.40 & 0.01 & 0.08 & 0.02 + 17 & 397.26 & 17.49 & 20.58 & 0.01 & 0.48 & 0.02 + 18 & 78.37 & 6.96 & 20.66 & 0.01 & 0.06 & 0.02 + 19 & 677.36 & 7.24 & 21.40 & 0.02 & 0.32 & 0.03 + 20 & 532.09 & 7.86 & 20.43 & 0.01 & 0.83 & 0.02 +
the analysis of the cmds , presented in figure [ fig2]a and figure [ fig2]b for the cluster and the field sample respectively , shows that the overall morphology of the main stellar populations in the ngc 1866 region is well - defined .
in particular , in figure [ fig2]b , some typical features of the population mix in the field are evidenced ( see also the discussion by * ? ? ?
* ; * ? ? ?
* ) : \(i ) the field multi - population main sequence ( ms ) extending up to @xmath28 merges into the ms of ngc 1866 \(ii ) a pretty well - defined giant branch of a presumably old ( some gyr , see * ? ? ?
* ) stellar population is clearly visible at @xmath29 and @xmath30 .
\(iii ) a red giant clump is located at @xmath31 , indicating the existence of an intermediate - age population having @xmath32 myr , ( see * ? ? ?
* ; * ? ? ?
* ) on the other hand , from the comparison of figure [ fig2]a and figure [ fig2]b the main features of the cluster population are also visible : \(iv ) the bright he - burning clump between @xmath33 and @xmath34 @xmath35 and @xmath36
\(v ) a fairly well - populated agb which extends up to @xmath37 and @xmath38 .
in addition , some stars lying in the portion of cmd , that we assume populated by the field , belong to the cluster and are late agb stars detected and studied in the near - ir @xcite
. we shall discuss this point later on .
these features have no correspondence in the field sample . on the basis of these considerations
, we can define three main regions in the cmd : \(1 ) the cluster red giants ( cluster rgs ) for @xmath39 and @xmath40 \(2 ) the field red giants ( field rgs ) for @xmath41 and @xmath30 \(3 ) the cluster main sequence ( cluster ms ) for @xmath42 .
all these regions are separated by dashed lines in figure [ fig3 ] .
however , in order to better enucleate the main features of the cluster population with respect to the field ones , we divided the cluster sample in 5 concentric annuli : * annulus 1 : @xmath43 ; * annulus 2 : @xmath44 ; * annulus 3 : @xmath45 ; * annulus 4 : @xmath46 ; * annulus 5 : @xmath47 ; figure [ fig4](a - f ) shows the radial cmds for the five annuli defined above ( panels a to e ) compared with the field sample ( panel f ) . as we can see , although conspicuous errors affect the photometry in this crowded region , the innermost annulus ( annulus 1 ) is clearly dominated by the cluster population : no features of the field populations are present in this cmd , while both the cluster rgs clump and the agb are clearly visible .
on the other hand , the field rgs regions become , as expected , more and more populated at growing distances from the cluster center ( annuli 2 , 3 , 4 , 5 ) , and the main features of this population ( the red clump and the old red giant branch ) progressively become more and more evident ( compare panel ( c),(d),(e ) , with panel ( f ) in figure [ fig4 ] ) .
another feature , which can be easily evaluated from figure [ fig4 ] ( panels ( a ) , ( b ) , ( c ) ) is the termination point of the cluster ms which is fairly well defined and turns to be located at @xmath48 . in order to estimate photometric errors , we considered the internal errors from daophot combined with the uncertainty of the calibration procedure and , when possible , frame to frame scatter ( i.e. for the shallow frames ) .
we report in figure [ fig5](a - l ) the mean photometric errors in v and ( b - v ) for each of the five annuli defined above , averaged over 0.5 mag - wide bins . as expected , larger errors occur in the most internal crowded regions ( annulus 1 ) and at fainter magnitudes .
since one of the main goals of this paper is to present meaningful lf and population ratios , the treatment of the completeness of the observed sample is a crucial problem : for this reason , we shall discuss this point in detail .
the problem of estimating the completeness factor for a stellar population comes out dramatically when increasing the crowding conditions of the observed field . on the other hand ,
the crowding conditions depend , to a zero - th order , on the distance from the cluster center : in the second place , on the seeing because in bad seeing conditions , it ismore likely to blend close stellar images .
the final effect of this phenomenon is the loss of fainter stars and , eventually , the appearances of spurious bright objects , that could remarkably alter the resulting lf . in order to ( quantitatively )
estimate the fraction of objects lost in each magnitude bin , @xmath49 , we carried out extensive _ artificial star tests_. this procedure has been described in detail in many papers ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the basic idea is to add a number of `` artificial stars '' ( i.e. objects having the same characteristics of the _ real _ stars in the frame ) to the original frame , and to reduce once again the new _ enriched _ frame . in the specific case of ngc 1866 ( as in the case of any lmc cluster ) , the fact that the cluster is almost completely contained within a single frame generates a large density gradient within the frame and requires that the completeness parameter be evaluated also as a function of the distance from the cluster center .
for this reason , we adopted the sub - division in concentric annuli defined in section 3.1 to quantify the dependence of the incompleteness on the distance from the cluster center ( i.e. on the crowding conditions ) .
for each annulus we generated a set of _ artificial _
stars for each bin of magnitude ( 0.5 mag wide ) over the entire magnitude range covered by our observations ( @xmath50 ) . in order to avoid spurious crowding enhancement due to the enrichement procedure
, only 20 - 30 stars were added to the original frame in each iteration , and the procedure was then repeated many times .
a virtually identical reduction procedure was performed on the _ enriched _ frames in order to ensure a homogeneous treatment of both artificial and original frames .
the output list of stars ( reporting magnitude and positions ) was then cross - checked with the list of artificial stars added to the frame .
each artificial star was considered _
`` detected '' _ if the following criteria were satisfied : @xmath51 , @xmath52 , @xmath53 , where @xmath54 , @xmath55 and @xmath56 are the differences in position and magnitude of the recovered star with respect to the simulated one .
the completeness factor ( @xmath57 ) in each annulus ( where the crowding density can be considered constant ) and in each bin of magnitude was finally derived by the ratio between the number of artificial stars recovered ( @xmath58 ) and the number of stars originally simulated ( @xmath59 ) : @xmath60 .
the procedure was applied independently to the deepest images in both filters and repeated many times : a total of more than @xmath61 artificial stars have been simulated . according to @xcite , the final completeness factor ( @xmath62 ) to each point @xmath63 of the cmd is @xmath64 . following this procedure
, the completeness level at the cluster rgs region is @xmath65 in both the filters , while the correction factor in the five considered annuli for the cluster ms are plotted in figure [ fig6 ] .
the completeness of our sample is high ( @xmath66 ) down to @xmath67 in the most external regions ( annuli 3,4,5 ) , then start to significantly decrease in annulus 2 ( being @xmath68 down to @xmath69 ) , and drops rapidly down to @xmath70 in the central field ( annulus 1 ) , even at a relatively bright magnitude level ( @xmath71 ) . for this reason ,
we decide to exclude the contribution of this very central region ( annulus 1 ) in deriving the lf .
the traditional zapping technique consists in gridding the two cmds , `` cluster+field '' and `` field '' , and count the stars in each cell in the two diagrams after normalizing the areas and correcting for completeness .
then , an equivalent number of stars is removed from the single cells of the diagram of `` cluster+field '' , on the basis of the number of field stars found in the diagram of the `` field '' alone .
this technique has two major problems : * the _ a priori _ gridding leads to the possibility that populous sequences are arbitrary cut by the cell boundaries , and one cell is more affected than others by the subtraction . in this way
spurious sequences could be generated , as for example in the case of an almost vertical ms cut in half by the cell boundaries , with a slightly redder field population ms ; * if , within a single cell , the stars to be dragged out are chosen at random from the sample , clearly the probability of picking stars where the density is higher , i.e. in the vicinity of the sequence , is large .
the result would be an artificial depletion of populous sequences .
a possible solution is to trace a mean ridge line , and attribute weights to the points so that it is more likely to drag out stars lying far from the sequence rather than close to it . in order to avoid these problems , in the following we adopted the method described by @xcite .
this method considers , for each star in the `` cluster+field '' cmd , its own cell , then counts the stars which fall in the cell , in the two diagrams ( `` cluster+field '' and `` field '' , respectively ) .
the size of the cells are determined on the basis of the error - box associated to each star , i.e. [ @xmath72 . a probability is then assigned to each star : @xmath73 where @xmath2 is the ratio of the areas of the field and of the cluster , @xmath74 and @xmath75 are the number of stars corrected for completeness in each cell of the `` field '' and in the `` cluster+field '' cmds , respectively .
then , a random number @xmath76 is extracted , between 0 and 1 . if @xmath77 , the star is taken , otherwise is discarded .
as already quoted in section 2.1 , in this paper we assumed as field sample all the observed stars lying at @xmath78 from the cluster center .
this area measures @xmath79 . as we said beforehand , in the literature there are at least three previous photometric surveys specifically aimed at studying the stellar population in ngc 1866 : c89 , b89 and the revision by bcp94 . in these papers ,
the authors presented lfs based on independent photometric studies of the cluster : the region sampled in each of the three papers are all included in the field covered by our observations .
moreover , the cmd plotted in figure [ fig2 ] and figure [ fig3 ] is by far the deepest , most accurate and populous cmd ever published for ngc 1866 ( compare our figure [ fig3 ] with figure 5 in b89 and figure 7 in c89 or even figure 2 in bcp94 ) .
for this reason we will use our sample in order to establish the existence of any systematic difference , bias or incompleteness which may affect the previously published samples . magnitudes and positions of the stars for all the published samples have been kindly provided by the authors in electronic form .
the field observed by c89 covers an area of @xmath80 roughly centered on the cluster center ( see figure 1 by c89 ) .
their pixel size is comparable to ours ( @xmath81pix ) .
first of all , the sample observed by c89 has been reported in our coordinate system . the cluster center assumed by c89
is located at ( @xmath82 ) , slightly different from the one assumed in our sample .
however , in order to be fully consistent with the assumption of c89 we adopted , in the comparison , the above coordinates for the cluster center .
then , we searched for any systematic difference in magnitudes and colors between the two samples over the whole region in common in our field and c89 .
the stars in common between the two catalogs have been used to derive the equations relating c89 magnitudes to our system .
the equations are reported below : @xmath83 @xmath84 once the magnitudes and the position of the original c89 list were reported to our system , we performed an accurate search for the stars in common in the specific area selected by c89 to construct the _ normalized _ lf .
the field specifically considered by c89 in their analysis is located between @xmath85 and @xmath86 .
this area ( @xmath87 ) is represented as a horizontally striped region labeled c89 in figure [ fig7 ] .
stars have been considered identified under the following assumptions @xmath88 figure [ fig8 ] summarizes the results of the comparison : in particular , in figure [ fig8 ] we plotted the cmd for the 652 stars found in common in the selected area ( panel a ) , stars which have been found only in our sample ( panel b ) , and stars which have been measured only in the c89 sample ( panel c ) . according to c89 , dashed lines delimit different regions of the cmd mainly dominated by different populations ( see their figure 7 ) . in particular , _ ( a ) _ the region dominated by the cluster main sequence ( ms ) at @xmath89 , _ ( b ) _ the field rgs at @xmath90 and @xmath91 , and _ ( c ) _ the cluster rgs stars at @xmath90 and @xmath92 . moreover , the cluster rgs region has been divided into three sub - groups ( labeled a , b , c ) as suggested by c89 .
it is worth noting that while c89 , at first , suggested to use only giants counted in group a and b ( since _ `` group c likely do not belong to the same population as those of group a and b''_),they later adopted ( in normalizing the ms lf ) always the whole cluster rgs population ( group a+b+c ) ( see their table 1 and the further discussion in their sect .
the most relevant results of the comparison are : \(i ) only 4 stars ( over 656 stars ) have no counterpart in our catalog ( two of them are in the region of the cmd dominated by the field population see figure 8c ) \(ii ) the c89 list turns out to be seriously incomplete with respect to our sample , even at relatively bright magnitudes ( @xmath93 ) ; \(iii ) all the 39 bright rgs ( brg ) stars found by c89 have been identified in our sample .
in addition , we found another brg ( namely star 3023 : @xmath94 and @xmath95 ) which has no counterpart in the c89 catalog ( see figure [ fig8]b ) .
\(iv ) a brg ( namely 1018 in the c89 catalog ) looks like a blend of two stars : one of the two blended components still lies , though at a fainter magnitude , in the brgs region ( sector b - see figure [ fig8]b ) ; therefore this fact does not affect the total number of brgs .
b89 sampled 4 fields ( @xmath96 each ) at different distances from the cluster center .
one of the fields was centered on the cluster and was used by the authors to construct the lf .
we checked the relative photometric calibration with respect to b89 , following the same procedure described above ( with the same assumptions ) .
after applying the cross - correlation procedure , 1355 stars in common have been found and have been used to obtain the following equations which relate the b89 magnitudes to our system : @xmath97 @xmath98 in this case , we adopted a second order polynomial to derive the correction .
this gives a smoother relation that takes into account both the small corrections at bright magnitudes and bluer colors and the larger correction at faint magnitudes and redder colors .
however , the overall corrections are in general very small ( @xmath99 ) .
the region selected by b89 for constructing the lf is an annular region ranging @xmath100 from the center .
this area ( @xmath101 ) is indicated as a tilted , diagonal shaded region , labeled as b89 , in figure [ fig7 ] .
similarly to figure [ fig8 ] , figure [ fig9 ] reports the cmds resulting from the comparison between our catalog and b89 , in the region selected by b89 for constructing the lf . in this case too , the main results of the comparison closely resemble the discussion reported in the previous section . in summary : \(i ) all the 52 brgs found by b89 have been identified in our sample , and as previously found in the comparison with c89 an additional brg has been identified in our sample ( namely star i d .
1018 ) ; \(ii ) also in this case the b89 sample appears to be seriously incomplete for @xmath102 .
interestingly enough , b89 did not perform any statistical decontamination from field stars in the region of the cluster rgs and in normalizing the ms lf they used all the 52 rgs identified .
bcp94 completed the b89 survey presenting a ( v , v - r)-cmd for an extended region around the cluster ( see figure 1 in bcp94 ) .
as in the previous comparisons , we used their sample in order to find systematic difference in magnitudes and/or colors . in this specific case ,
only the v magnitudes could be compared to our photometry.the bcp94 v magnitudes match our system by applying only a small offset : @xmath103 a detailed analysis of the stars in common has been carried out only in the region that the authors chose for constructing the lf .
this area is the annulus limited by the largest dashed circle which complete the area surveyed by b89 and labeled as bcp94 in figure [ fig7 ] .
the extension of the region discussed by bcp94 is @xmath104 .
figure [ fig10 ] presents the cmds obtained from this comparison .
as can be seen , the limiting magnitude of the bcp94 sample is located at @xmath105 .
it is interesting to note , however , that in this case , for @xmath106 , only few stars in the cluster ms region are present in our list and are missing from the bcp94 list . in particular , all the 65 brgs found by bcp94 have been identified .
figure [ fig11 ] shows the comparison of the integrated stars counts between the sample presented here and the three previous samples ( c89 , b89 , bcp94 , respectively ) .
all stars lying in the sector of the cmd defined as ms in figure [ fig3 ] , in each of the regions selected by the various authors have been used for this comparison .
no corrections ( neither for completeness nor for field contamination ) have been applied to the star counts .
the ratio of the stars counted in each magnitude bin can be used to derive the _
`` relative '' _ completeness ratio of each sample with respect to ours .
the comparison shows that the completeness ratio decreases to less than @xmath107 at @xmath108 for c89 and bcp94 and at @xmath109 for b89 , respectively .
the decontamination procedure described in section 3.4 has been applied to each of the 5 radial annuli defined in section 3.1 .
the procedure has been applied only for magnitudes brighter than @xmath110 , because the lf will be studied only down to the magnitude bin where the overall completeness drops below 50% , to prevent problems due to statistical fluctuations close to the cutoff line .
moreover , field decontamination at faint magnitudes requires too high completeness corrections , which could give rise to unreliable star counts .
figure [ fig12 ] reports the cmds in the radial annuli 1 , 2 , 3 , 4 and 5 before and after the field decontamination , respectively . as discussed in section 3.3 ,
the innermost region ( annulus 1 ) will not be used in the construction of the lf and for this reason it will not be considered in the following discussion .
thus , the final adopted sample covers the region in the @xmath22 field over a radial range @xmath111 ( annuli 2 + 3 + 4 + 5 ) .
this field has approximately an extension of @xmath112 .
a particular attention has been given to the statistical decontamination of the region dominated by the cluster rgs , since this population is used to normalize the ms lf in the comparison with the previous works , and -this is even more important- it is the test population that will be considered in the following model comparison .
seven stars in the field sample have been found to lie in the region of the cmd labeled as cluster rgs , thus we expected @xmath113 field stars lying in that area of the cmd .
this leads to a final number of 107 rgs after the statistical decontamination from the field stars that will be considered in the following discussion .
figure [ fig13 ] reports the final decontaminated cmd for the global adopted sample ( obtained co - adding the cmds of annuli 2 , 3 , 4 , and 5 ) .
the dashed lines reported in figure [ fig13 ] identify the regions defined in figure [ fig3 ] .
there are few stars beyond the red limit of the cluster rgs box , which are indicated as late - agb stars .
these have been identified from the near - ir sample of @xcite as stars in the thermally pulsing agb phase .
since our models do not include the latest stages of agb evolution , and this evolutionary phase is very short within the lifetime of a star , we will not consider them in the giant sample in the following analysis and discussion . in order to obtain the luminosity function of the cluster ms stars , we followed a standard procedure already used in other papers ( see , e.g. , * ? ? ? * ) .
first , we determined the mean ridge line ( mrl ) of the ms was from the statistically decontaminated sample plotted in figure [ fig13 ] , by adopting an iterative procedure : the sample was divided into 0.2 mag - wide bins , then , for each bin , the star distribution in @xmath114 color was computed , the mode of the distribution was assumed as first - guess equivalent color , and all the objects lying more than @xmath115 away from the mode were rejected . secondly , the mode of distribution was re - computed until convergence . when constructing the lf , only stars lying , in each 0.2 mag - width bin , within @xmath116 from the mrl were counted .
however , since the sequence s width depends on the errors , we performed the star selection for the lf in each annulus defined in section 3.1 .
the global ilf obtained by co - adding each ilf computed in each annulus ( 2 to 5 ) is presented in figure [ fig14 ] .
the two lines represent the observed ilf before ( dashed line ) and after ( solid line ) the correction for completeness , respectively .
errors have been computed accordingly to the following relation : @xmath117\ ] ] where @xmath118 is the number of stars effectively observed in each bin , @xmath119 the completeness factor , and @xmath120 the associated error which was determined from the r.m.s . of the repeated completeness trials ( tipically less than 0.05 ) . in the same way ,
figure [ fig15 ] reports the ilf normalized to the total number ( 107 ) of rgs in the statistically decontaminated sample ( nilf ) .
the global uncertainty of the nilf in each bin depends also on the statistics of the number of rgs used to normalize the ilf , accordingly to the relation : @xmath121\ ] ] where @xmath122 is the number of rgs adopted for normalization , @xmath123 is the error in each bin of the ilf obtained from eq . ( 1 ) and @xmath124 is the number of ms stars in each bin of the ilf . in figure [ fig16 ]
we compare the ms nilf , computed adopting our sample , with the ones published in previous papers .
we considered each of the region covered by previous papers on its own ( see figure [ fig7 ] ) . in order to emphasize the number of rgs adopted in each region to normalize the ms ilf ,
we briefly summarize the above results : the actual number of rgs found by our photometry in each area surveyed by previous photometries are , respectively : 40,53 and 65 in the c89 , b89 , and bcp94 area . accordingly to the adopted field population and the relative area extensions , we expected 1,4 , and 6 field stars , respectively , in the cluster rgs sector of the cmd .
this yields a final number of 39 , 49 and 59 rgs , which were used to normalize the ms ilf .
it is particularly interesting to note that , apparently , both b89 and bcp94 did not apply any correction in order to take into account the field star contamination to the number of rgs .
this correction is particularly large ( 6 rgs ) in the area covered by bcp94 .
the original nilf for the three published samples have been reported ( as dashed lines ) for comparison in figure [ fig16 ] .
they have been measured directly on the original published figures since none of the various authors reported the adopted lf in form of a list , and we could not reconstruct the completeness correction and the decontamination procedure applied by the authors to the observed samples
. however , beside differences in the limiting magnitude , techniques and procedure , the nilf obtained in this paper agrees , within the uncertainties , with those published by previous studies at the bright end of the nilf ( @xmath106 ) .
also , the fact that the decontamination procedure has not been applied to the b89 and bcp94 original sample does not seem to have affected the nilf , probably because , in both studies , the authors systematically over - estimate the completeness correction for ms stars .
finally , we report the original nilf from c89 and bcp94 in figure [ fig17 ] .
the nilf obtained by b89 is not reported because it is fully consistent with the one from bcp94 by which it is superseded .
we also draw , for comparison , the global nilf obtained for the whole adopted sample ( plotted in figure [ fig15 ] ) .
the nilfs agree with each other , thus suggesting that the procedure adopted for the field decontamination and completeness correction does not introduce any substantial bias in the derivation of the nilfs .
there is a weak indication that the nilf obtained from the c89 area is slightly more populated , but this should be regarded as an effect of statistical fluctuations . in fact , given the error bars , the three distribution are fully consistent within 1 @xmath4 , and if the normalization factor of c89 is enhanced from 39 to 41 , the relative lf matches almost perfectly the others .
on the basis of figure [ fig16 ] and figure [ fig17 ] , we can conclude that the statement by bcp94 that the differences in the observed nilfs were due to statistical fluctuations in the choice of the samples seems confirmed .
in order to compare the present data with theoretical models , some issues such as metallicity , distance modulus , and metallicity must be discussed . in the next paragraph
we report some of the most recent determinations , as well as our final choices .
_ metallicity _ : most recent determinations of metallicity for ngc 1866 @xcite attribute to the cluster a value @xmath125\sim -0.55\pm0 .30 $ ] .
it should be noted that @xcite performed a relative calibration of ngc 1866 on ngc 330 , so that , depending on the value estimated for the latter cluster , the metallicity of ngc 1866 can be enhanced to @xmath125\sim -0.35 $ ] .
_ reddening _ : values of e(b - v ) from the literature span from 0.05 @xcite to 0.10 @xcite .
most authors , anyway , adopt e(b - v ) = 0.06 ( * ? ? ?
* ; * ? ? ?
. _ distance modulus _ : the distance of the lmc is subject of controversy within the astronomical community . even since the hipparcos results , the problem is far from being solved once for all .
the current values for the distance modulus span from 18.06 @xcite -which is , however , an outlier- to 18.70 @xcite .
useful reviews of recent results on the distance modulus of the lms can be found in @xcite , @xcite and @xcite . in the following discussion
, we adopt @xmath126\simeq -0.4)$ ] , @xmath127 , and two visual distance moduli , namely @xmath128 and @xmath129 , while the reddening will be assumed to vary between 0.06 and 0.10 ( which means that the true distance modulus will range at most between 18.3 and 18.6 , having adopted @xmath130 ) .
we will focus our discussion on the lf because it provides a more robust tool than the distribution of the stars in the hr diagram ( since it does not use the radii or , alternatively , the effective temperatures of the models , i.e. colors ) . since we want to minimize the influence of the corrections for crowding on the observed lf , we limit our analysis to the region brighter than @xmath131 .
in fact , down to this magnitude , the global corrections ( which have been included ! ) to each bin are always smaller than 10% , with only annulus 2 being corrected for a factor 40% in the faintest bin . the global correction to the total
is constrained below 8% ( see sect . 3.3 and 3.4 ) .
the theoretical models have been computed with the evolutionary code franec , whose earliest and latest versions are described , respectively , in @xcite and @xcite , while the latest input physics are described in @xcite .
most models used here have been published by @xcite , while others have been especially computed for this work .
the models computed with an artificially enlarged convective core ( hereinafter aecc models ) have been obtained by increasing the size of the convective core by an amount equal to @xmath132 .
we refer to such models as aecc and not as models computed with overshooting , since the fit to this ( or another ) cluster can only give us information about the probable size of the convective core , not about the mechanism which determines this size , as mentioned above .
all the theoretical luminosity functions have been obtained by means of a code which produces the synthetic clusters ( o. straniero et al .
, in preparation ) : this is a sophisticated tool which can manage a ) the photometric errors , b ) the effect of the crowding , c ) the presence of binary systems having an arbitrary mass ratio , d ) an age spread , e ) an arbitrary initial mass function and f ) a large number of photometric bands .
synthetic cmds can be produced in the age range from few myr up to several gyr .
all the models of which our evolutionary database range from the main sequence up to the advanced stages of interest for this work , so that no artificial match between different evolutionary phases is necessary .
the various studies of the lf of ngc 1866 available in the literature make use of the nilf , as discussed previously .
such approach was certainly valid in the previous studies of this clusters since a solid and complete ms extending over several magnitudes was not available .
we have used the nilf in the previous sections in order to have a common ground with the previous papers on the subject . however , since we now have a well - defined ( and essentially complete ) ms down to at least 3 mag below the to , we have a chance to adopt the ms itself as `` normalization factor '' instead of the he burning stars .
our synthetic cmds are populated until the observed total number of stars brighter than @xmath131 is reached and not until the total number of giants observed is reached .
we believe this choice is , whenever possible , preferable to the other because : i)the total number of stars brighter than the quoted magnitude is largely independent on the assumed size of the convective core , since it is dominated by ms stars whose lf is only marginally affected by the size of the convective core ; ii ) the number of ms stars is much larger than that of the giants and is therefore much less affected by stochastic fluctuations . it must be noted , however , that the total number of stars brighter than the quoted magnitude depends on the adopted distance modulus ( since it influences the limiting magnitude where we stop the counting ) , on the adopted present mass function ( hereinafter pmf ) , and on the fraction of binary systems present in the cluster .
we will discuss this below .
the observed total number of stars brighter than @xmath131 in our sample , corrected for completeness , is 4004 for the visual distance modulus @xmath128 and 4362 if @xmath129 is adopted ; the total number of giant stars is 107 . before discussing the fit to ngc 1866
, we think it important to discuss the dependence of the theoretical lf on various parameters .
panels a ) and b ) in figure [ fig18 ] show , respectively , the differential and the integrated main sequence lfs for three ages , 80 ( solid ) , 120 ( dotted ) and 160 ( dashed ) myr .
clearly , the to luminosity declines as the age increases while the overall shape of the lf around the to remains substantially unaltered within the considered age range . panels c ) to h ) in the same figure present the same lfs shown in the first two panels , but with the inclusion of the giants : from these panels emerges that the clump luminosity also dimmers as the age increases .
it should be noted that the distribution of the stars within the clump shows a double peak which flattens as the age increases .
this behavior is easily comprehensible if we consider that the stars spend most of their he burning lifetime either at the red dip minimum and at the blue nose of the blue loop , not in the middle .
since the older the isochrone the smaller the blue loop , it is also clear that the double peak feature disappears as the age increases .
another clear feature in figure [ fig18 ] is the magnitude gap between the to and the red giant clump .
this gap reduces as the age increases , because the clump luminosity dimmers faster than the to ( see * ? ? ?
figure [ fig19 ] shows , for the same three ages , the effect produced by an increase of a factor of two in metallicity on the differential lf : the left and right panels show , respectively , the ms and the total lf , while the solid and dotted lines refer , respectively , to the larger and the reference metallicity .
the end effect of an increase of 0.3 dex in metallicity is a simulation of an older age ; in fact , the clump luminosity becomes fainter , and both the double peak feature within the clump and the gap between the to and the clump itself are reduced , while the to becomes more luminous ; the latter feature implies that the larger metallicity requires an older age in order to have the same to luminosity .
the influence of the size of the convective core on the lf can be readily seen in figure [ fig20 ] where the solid lines refer to the aecc case while the dotted ones refer , once again , to the reference case .
the increase in mass of the convective core has a deep impact on the lf : the region around the to is modified because the to luminosity increases and the aecc lf is much steeper around the to .
also , the clump is significantly affected by the aecc , since it appears to be brighter and flatter .
it is interesting to note how the synthetic diagrams populate the respective isochrones .
figure [ fig21 ] shows in each panel a different isochrone superimposed onto its synthetic diagram obtained by choosing a number of stars equal to the sample observed in ngc 1866 , i.e. 4004 stars brighter than @xmath131 .
the upper and lower panels refer , respectively , to the two ages t=100 and t=200 myr , while the left and right panels refer , respectively , to the reference ( standard ) and aecc isochrones .
the ms stops being populated before the formal `` theoretical '' to ; this is due to the fact that these clusters ( and ngc 1866 is even one of the most populous ones ! ) are not rich enough to significantly populate the isochrone around the to .
the effect is only marginally present in the standard case , but it is dramatic when the aecc model is considered . as a consequence , the difference between the two @xmath133 ( standard and aecc ) is largely reduced if evaluated on the synthetic diagrams ( i.e. on the basis of the most brilliant ms star ) rather than on the isochrones .
let us explain this point in detail : the region between the small ms hook ( see figure [ fig21 ] ) and the `` theoretical '' to is populated by stars which are in the so - called _ overall contraction _ phase : during this phase , the stars shift from a structure controlled by a central burning to one controlled by a shell burning .
the evolutionary timescale of this phase is largely determined by the size of the he core mass ( which directly depends on the size of the convective core ) , in the sense that it scales inversely with the size of the he core .
thus , the larger the he core mass ( i.e. the larger the previous convective core ) , the faster the evolution between the ms hook and the to ( and hence the lower the expected number of stars in this region ) : this explains why synthetic cmds obtained by assuming a larger convective core have the to region of the isochrone underpopulated with respect to the standard case .
the last effect we want to show is connected to the likely presence of a significant fraction of binaries .
it has been recently found with hst observations @xcite that ngc 1818 , a cluster similar to ngc1866 , but slightly younger , has a conspicuous population of binary systems estimated of the order of 30% . in our analysis ,
the presence of binaries is taken into account as follows : 1 ) the mass of the first star comes , as usual , from a random extraction weighed with a pmf having the same slope @xmath2 used in the extraction of the single stars , 2 ) the mass of the companion is extracted by assuming a gaussian distribution ( of the masses ) with a peak mass equal to that of the first star . in this particular case
, we chose a rather narrow gaussian distribution ( semidispersion equal to 0.1 ) , in order to maximize the effects due to the binary population .
clearly , the effect on the luminosity of the compound system is maximized if the two stars have similar masses because , otherwise , one of the two would dominate and the total luminosity would practically be that of the most luminous of the two . by the way ,
such an occurrence clarifies why the crowding behaves in principle similarly to real binary systems but is , in practice , quantitatively less effective : in fact , once the fraction of `` binaries '' is fixed , if one assumes that they are all due to the crowding , the masses ( i.e. the luminosities ) involved in the formation of the `` binary system '' are not constrained by any rule to be similar , and hence , as a consequence the less luminous of the two is hidden by the other which , in turn , will remain essentially unmodified .
figure [ fig22 ] shows in the left panel a comparison between the integrated lf of the main sequence in the reference case ( thin line , hereafter msilf ) , and the one obtained including a 30% of binaries ( thick line ) for an age of 120 myr and a mass function slope @xmath134 .
the inclusion of a significant fraction of binaries alters the lf by flattening the ms slope and steepening the lf of the to region .
this means that the presence of the binaries masks the real slope of the pmf , in the sense that it appears flatter than it really is .
the right panel shows , for example , how the reference msilf already shown in the left panel ( computed with @xmath134 ) is well reproduced by the one computed with @xmath135 and a 30% of binaries .
such occurrence will have a significant impact on the fit to the cluster ( see below ) . with the aid of the theoretical scenario discussed above
, we can now readily turn to the analysis of ngc 1866 .
in particular , we shall first show the fits with the standard and aecc models , and we shall then discuss how they change by including a 30% of binaries . the comparison between our standard models and ngc 1866 is summarized in figures [ fig23 ] to [ fig26 ] for the assumed distance modulus @xmath128 . in figure [ fig23 ]
we report , for the three ages 100 , 120 and 140 myr and five values of the pmf slope ( @xmath136 ) , the comparison between the theoretical msilf ( histogram ) and the cluster data ( black dots ) .
although this is a standard way of comparing the theoretical and the observed integrated luminosity functions , we think it difficult to judge the goodness of a specific fit .
hence we computed , for each case shown in figure [ fig23 ] , the differences ( observed - theoretical ) , in units of @xmath137 .
these are shown in figure [ fig24 ] .
hereafter , we will refer to this quantity as @xmath3 .
since it will constitute the basis on which we shall determine both the age and the pmf slope , @xmath2 , we discuss it in brief : as the total number of stars brighter than @xmath131 is equal to the observed number , the first point to the right end of each panel always corresponds to @xmath138 ( i.e. theoretical and observed number are identical ) .
moving toward brighter magnitudes , the other values for @xmath3 will not necessarily remain around the zero ( good match ) but their trend with the magnitude will depend on the current adopted values for both @xmath2 and the age : the larger the range in magnitude over which @xmath3 remains close to zero , the better the fit .
now , let us consider the various rows in figure [ fig24 ] : the three panels in each row have been obtained by assuming a constant pmf slope and the three adopted ages . as a general trend , an age increase tends to globally lower the sequence of points . by looking along the columns at figure [ fig24 ] , one can see the effect of a changing @xmath2 at fixed age . in this case , the global trend of @xmath3 is toward lower values as @xmath2 increases .
if we focus on the first column in figure [ fig24 ] , we can see that , for an age of 100 myr , an increase in @xmath2 from 2.0 to 2.2 improves the fit : all the points move down toward zero and a good match is obtained in the magnitude range 1.5 - 2.6 ; a further increase of @xmath2 worsens the fit because the points corresponding to the brightest part of the ms still tend to move down toward zero , whereas those which were previously close to zero move progressively below it .
since , even in the best case ( @xmath139 ) , only a very small part of the ms can be fitted , we conclude that , for this age , models can not fit the data well enough .
the same general behavior is obtained for the other two ages too ; in particular , for t=120 myr the `` best '' fit is obtained for @xmath2 of the order of 2.2 - 2.3 , while for t=140 myr the `` best '' match corresponds to @xmath140 . at variance with the 100 myr case
, however , there is a good fit all along the magnitude range from -0.5 to 2.6 for both these ages . by increasing the age above 140
myr one obtains a situation similar to the one discussed for 100 myr : no good fit is obtained for any value of @xmath2 over a reasonably extended magnitude interval of the ms . by adopting a visual distance modulus of 18.6 ,
we conclude that it is possible to constrain the age of this cluster in the range 120 - 140 myr with @xmath2 in the range 2.1 - 2.3 , depending on the age .
it is worth noting that , by means of this technique , we have been able to constrain the age without uing -so far- the to luminosity : on the contrary , we have shown ( and used ) that the part of the ms down to 3 - 4 magnitudes below the to depends significantly on the age ; an age derived in this way is much more robust than the one obtained by a fit to the most luminous ms stars , because these stars are very few and hence are subject to heavy stochastic fluctuations .
the further fit to the to luminosity remains confined within the age range found above .
the various panels in figure [ fig24 ] show that , in no case , it is possible to obtain a good fit to the ilf near the to region .
standard models predict a smooth decline of the ilf toward the brightest portion of the ms while the observed ilf shows a much steeper profile .
since the theoretical ilfs preserve such a shape around the to over an extended age range , we will not obtain a good fit for any standard model in any -reasonably wide- age range .
we can now extend our analysis to the post main sequence evolution , i.e. the core he burning phase . in figure [ fig25 ] and figure [ fig26 ]
we show , respectively , the differential and integrated lfs of the total sample . by looking at the panels corresponding to the cases showing a good fit to the ms ( see above ) we find that , while the average clump luminosity is reasonably well reproduced , there are a number of discrepancies which can hardly be reconciled : in particular , the models predict a number of clump stars which is almost twice the observed one and also a significant ( @xmath141 0.6 mag ) luminosity gap between the to and the he clump , which is not observed .
we therefore conclude that it is impossible to obtain a reasonable fit to ngc 1866 , having assumed a visual distance modulus of 18.6 , with the use of standard models .
the same analysis has been made for a visual distance modulus of @xmath129 .
the logical steps followed to perform the fit are the same adopted for the shorter distance modulus . in this case , we obtained a good fit to the ms only for @xmath142 , @xmath143 , and @xmath144 , @xmath145 , while there is no good fit for ages outside this range , whichever @xmath2 is used .
we show the best fit for this case in figure [ fig27 ] . the same discrepancies already evidenced for the shorter distance modulus are still present : the shape around the to region is not reproduced , nor are the total number of giants and the gap between the bright end of the ms and the clump .
the adoption of a longer visual distance modulus does not improve the fit to the cluster , which remains rather poor .
we have thus obtained , once again , the classical old problem first raised by @xcite : i.e. standard models predict much more he clump stars than observed , with the difference that the validating sample is now much larger .
the fit with aecc models is shown in figure [ fig28 ] to figure [ fig31 ] for the distance modulus @xmath128 .
now , a good fit to the msilf is obtained in the age range 200 - 220 myr and @xmath146 , but only for @xmath147 myr and @xmath148 there is also a reasonable fit to the to .
an analysis of the post ms evolution ( figure [ fig30 ] and figure [ fig31 ] ) shows that the number of giants predicted for this `` best '' case is in good agreement with the observed one .
however , it must be noted that the clump lf is not well reproduced since it is predicted to be , on the whole , less luminous than observed .
the increase of the visual distance modulus to @xmath149 leads to a largely poorer fit . in this case
, the ms may be fitted in the age range 180 - 200 myr and @xmath2 s in the range 2.0 - 1.8 , but in no case do we obtain a fair fit to the to region .
the best overall fit ( shown in figure [ fig32 ] ) corresponds to @xmath150 myr and @xmath151 .
the number of giants predicted agrees with the observed ones , while the synthetic lf remains systematically below the observed and shows a well - defined gap between the clump and the brightest ms stars , which is not observed ( panels c ) and d ) of figure [ fig32 ] ) .
we conclude that the aecc models may fairly fit the observed lf only if a `` short '' distance modulus is adopted .
we now discuss the influence that a consistent fraction of binary systems ( 30% ) has upon the fit to a cluster like ngc 1866 .
as mentioned above , this value has already been found in ngc 1818 , by means of hst observations , and hence it is a reasonable starting value .
a good fit to the ms is obtained for @xmath152 myr and @xmath2 between 2.4 and 2.6 when a distance modulus @xmath128 is assumed .
although for both ages the shape around the to is relatively well reproduced , an age of 120 myr and @xmath153 give the best fit to the data ( see panels a ) and b ) in figure [ fig33 ] ) .
note that , at variance with the case without binaries , we now have a good fit to the to region too , because the presence of the binaries tends to steepen the ilf toward the bright end of the ms ( see above ) .
a further analysis of the giant stars is partly contradictory : while the predicted number of giants ( 148 ) is @xmath154 larger than observed ( 107 ) , the clump luminosity is reasonably well reproduced ( see panels c ) and d ) in figure [ fig33 ] ) . by increasing the distance modulus to 18.8
, we obtain a really good fit ( shown in figure [ fig34 ] ) for an age @xmath155 and @xmath156 . the whole ms sequence , including its bright end ,
is well reproduced .
the post ms evolution is also fairly well reproduced ( see panels c ) and d ) in figure [ fig34 ] ) : the number of giants predicted is within 2@xmath4 of the observed number and , even more important , the luminosity distribution of the giants closely matches the observed ones . in conclusion , the adoption of models computed with a standard extension of the convective core may lead to a good fit to ngc 1866 if we account for a non negligible population of binary systems @xmath157 and adopt the `` long '' visual distance modulus ( i.e. @xmath158 ) .
the to mass predicted for this `` overall globally best '' case is @xmath159 , while the he clump mass is around @xmath160 . for sake of completeness
, we also consider the case of aecc models with a 30% of binaries . a good fit to the entire ms and the to is now obtained for an age range @xmath161 and @xmath162 , if @xmath163 is assumed ( see panels a ) and b ) in figure [ fig35 ] ) .
the number of predicted giants closely matches the observed one .
nonetheless , this fit must be rejected because the luminosity of the clump stars is largely missed : panels c ) and d ) in figure [ fig35 ] clearly show that the sum of the two effects , i.e. larger convective cores and a 30% of binaries , leads to predicted giants with a luminosity far too weak in comparison with the observed ones .
the adoption of the longer distance modulus , which is not shown , does not improve at all the fit since , here too , the `` theoretical '' he clump is largely underluminous with respect to the observed one .
we conclude that the simultaneous inclusion of a binary component , together with the use of aecc models , does not provide an acceptable fit to the observed lf of this cluster . before closing this section
, we shall comment the so - called nimslf ( i.e. normalized integrated main sequence luminosity function ) for the cases obtained by including a binary component .
panels e ) in figures [ fig33 ] to [ fig35 ] show the comparison between the observed ( black dots ) and synthetic ( histogram ) nimslfs ; the open dots correspond to a 2@xmath4 changing in the observed nimslf . since this kind of figure gives a pivotal role to the number of giants , the fit obtained with the aecc+binary models clearly leads to the best fit .
we have , however , shown that a deeper analysis , i.e. one which addresses all the features available in the observed lf , leads to the conclusion that this case must be rejected . on the contrary , the fit obtained by adopting the standard models plus the binary component , although it gives a worse fit to the observed nimslf ( they agree within 2@xmath4 ) , leads to a reasonably good fit to this cluster .
in this paper , we presented a new set of photometric data which largely increase the samples presented up to now .
we have also shown systematically how the various theoretical models and/or scenarios can fit the observed data . as a first conclusion , we find that the discrepancy first evidenced by @xcite , i.e. that standard models predict an excess of he clump stars , is still present even with the latest standard models . the easiest and most straightforward way to avoid the discrepancy is to artificially increase the size of the convective core .
unfortunately , a side - effect connected to the use of aecc models is a fainter clump luminosity .
if the convective core is larger , an older age is needed in order to fit the bright end of the ms , but in this case , the isochrone displays a fainter he - burning clump .
another method used to reconcile the theoretical predictions with the present set of data is to take into account the possible presence of a fraction of binary stars .
already c89 temptatively tried to check the effect of the presence of a binary population on the fit to their nilf .
they concluded that the inclusion of even 50% of binaries did not significantly alter the fit and , hence , binaries were not relevant for the understanding of ngc 1866
. a look at fig .
18 of c89 shows that the inclusion of binaries increases significantly the nilf , although not enough to match the observations .
nevertheless , from that figure it was possible to show that binaries mimick ( of course only from this point of view ) an enlarged convective core : the observed ratio @xmath164 is increased .
this effect was also noted by b89 , who clearly stated in their final discussion that the presence of the binaries could mimick the effect of the overshooting on the nilf , and lead to a closer agreement between ( their ) data and the standard models .
note , however , that while models computed with aecc produce a fainter clump luminosity , models with binaries do not do so .
binaries alter a synthetic diagram in two ways : they overpopulate the bright end of the ms ( thus reducing the ratio @xmath164 ) and require a steeper `` intrinsic '' pmf slope .
this implies that , with a fixed total number of stars to be extracted , less massive stars are favored with respect to those around the to and beyond : hence the predicted number of giants is reduced . by including a fraction of 30% of binaries ,
we have shown that standard models can nicely fit the data .
of course , one could question the real occurrence of binary systems in ngc 1866 .
we have , in fact , two alternative scenarios : \a ) if there are essentially no binary systems in ngc 1866 , only models computed by significantly increasing the size of the convective core can fairly fit the present sample of data . in this case , however , we are still faced with the fact that it is impossible to fit the luminosity function of the clump stars .
such discrepancy increases together with the distance modulus . in this scenario ,
the age of ngc 1866 would be @xmath165 myr and the pmf slope @xmath166 ; \b ) if there is a conspicuous fraction of binary systems , standard models provide a good fit to the present data . in this case , however , the preferred visual distance modulus is @xmath129 , i.e. @xmath158 , depending on the adopted reddening . in this scenario ,
the age is @xmath167 myr and the intrinsic pmf slope , @xmath2 , of the order of 2.4 .
the choice between these two alternative scenarios can be done either by solving the problem of the distance modulus , because for each given modulus only one of them is favored , and/or by improving our knowledge of the fraction of binary systems existing in ngc 1866 . in our opinion a real advance in the knowledge of this ( and maybe other ) class of objects requires an observational effort for a real determination of the fraction of binaries present in the cluster , more than a further increase in the data sample ( which is , however , always welcome ) . as final remark
, we hope that the incoming cycle 8 hst observations will finally solve the enigma of ngc 1866 .
we gratefully acknowledge enzo brocato , cesare chiosi and antonella vallenari for giving us computer - ready tables of their published photometries and their isochrones , and for helpful discussions .
we warmly thank giuliana giobbi for editorial help , and the anonymous referee for useful comments .
, thackeray a.d . , 1967 , , 149 , 73 becker s.a .
, iben i jr . , 1979 , , 232 , 831 becker s.a .
, mathews g.j . , 1983 , , 270 , 155 bertelli g. , mateo m. , chiosi c. , brressan a. , 1992 , , 388 , 400 bolte m. , 1989 , , 341 , 168 brocato e. , buonanno r. , castellani v. , walker a.r . , 1989 , , 71 , 25 brocato e. , castellani v. , piersimoni a.m. , 1994 , , 250 , 59 cassatella a. , barbero j. , geyer e.h . , 1987 , , 64 , 83 castellani v. , giannone p. , renzini a. , 1971 , , 10 , 355 castellani v. , chieffi a. , pulone l. , tornamb a. , 1985 , , 294 , l31 castellani v. , chieffi a. , straniero o. , 1990 , , 74 , 463 chieffi a. , sraniero o. , 1989 , , 71 , 47 chieffi a. , limongi m. , straniero o. , 1998 , , 502 , 737 chiosi c. , bertelli g. , meylan g. , ortolani s. , 1989 , , 219 , 167 cox a.n . , tabor j.e .
, , 31 , 271 dominguez i. , chieffi a. , limongi m. , straniero o. , 1999 , , accepted ( astro - ph/9906030 ) elson r.a.w .
, 1991 , , 76 , 185 elson r.a.w .
, sigurdsson s. , davies m. , hurley j. , gilmore g. , 1998 , , 300 , 857 feast m.w .
, 1999 , pasp , 111 , 775 feast m.w . ,
catchpole r.m . , 1997 , , 286 , l1 ferraro f.r . ,
carretta e. , corsi c.e . , fusi pecci f. , cacciari c. , buonanno r. , paltrinieri b. , hamilton d. , 1997 , , 320 , 757 fischer p. , welch d.l .
, ct p. , mateo m. , madore b.f .
, 1992 , , 103 , 857 frogel j.a . , mould j.r .
, blanco v.m . , 1990 , , 352 , 96 hilker m. , richtler t. , gieren w. , 1995 , , 294 , 648 huebner w.f .
, merts a.l . , magee n.h . ,
eargo m.f . , 1977 , los alamos sci . lab .
report ( la-6760-m ) lattanzio j.c .
, vallenari a. , bertelli g. , chiosi c. , 1991 , , 250 , 340 madore b.f . , freedman w.l . , 1998 , , 492 , 110 mateo m. , 1998 , , 331 , 261 mighell k.j . , rich r.m . , shara m. ,
fall s.m . , 1996 , , 111 , 2314 oliva e. , origlia l. , 1998 , , 46 , 54 sandquist e. , bolte m. , stetson p.b .
, hesser j.e . , 1996 , , 470 , 910 stanek k.z . , zaritsky d. , harris j. , 1998 , , 500 , l141 stetson p.b . , 1987 , , 99 , 191 straniero o. , chieffi a. , limongi m. , 1997 , , 490 , 425 walker a.r . , 1992 , , 104 , 1351 walker a.r . , 1995 , , 110 , 638 walker a.r . , 1999 , in post - hipparcos cosmic candles , ed . a. heck and f. caputo , kluwer acad . publisher , p. 125 walker m. , 1974 , , 169 , 199 wood p.r .
, in frontiers of stellar evolution , ed .
d. l. lambert , aspcs vol .
20 , p. 213 | we present a new deep ( down to @xmath0 ) photometry of a wide region ( @xmath1 ) around the lmc globular cluster ngc1866 : our sample is much larger ( by more than a factor three ) than any previous photometry and with a main sequence which may be considered complete , down to at least 3 mag below the brightest ms star : such an occurrence allows a meaningful and robust comparison with various theoretical scenarios produced by means of models computed with the evolutionary code franec .
both age and present mass function slope , @xmath2 , are derived by a fit to the available ms and by the use of the parameter @xmath3 , which is simply the difference , in @xmath4 s , between the observed and predicted integrated ms luminosity functions .
our main conclusions are : \a ) the adoption of standard models ( i.e. computed by adopting the schwarzschild criterion to fix the border of the convective core ) allows a fair fit to the ms for an age of the order of 100 - 140 myr and a present mass function having a slope @xmath2 between 2.3 and 1.9 , the exact values depending on the adopted distance modulus .
it is moreover possible to reproduce the average he clump luminosity while the total number of stars predicted in the he clump is twice the observed value : this means that we re - obtain and confirm the first finding of @xcite , according to whom the simple adoption of a `` classical '' scenario leads to a neat discrepancy concerning the prediction of the number of stars in the he clump .
\b ) the adoption of models computed by increasing the size of the convective core by a certain amount , i.e. @xmath5 , leads to a fair fit to the main sequence only for a visual distance modulus @xmath6 , an age @xmath7 and @xmath8 . in this case ,
the total number of he clump stars is well reproduced , although the luminosity function of the he clump itself is predicted to be sistematically less luminous than observed .
the previous conclusions are based on the assumption that there is no appreciable population of binaries in ngc 1866 .
though there are not yet sufficient data on the frequency of binary systems in these clusters , we analyzed how the previous scenarios would change if a consistent @xmath9 population of binary systems were present in the cluster .
this choice is based on the fact that a fraction of binaries of the order of 30% has already been found in ngc 1818 , a cluster similar to ngc 1866 @xcite .
the inclusion of a 30% binary population leads to the following conclusions : \c ) the adoption of the standard models now leads to a good fit to the entire luminosity function , i.e. main sequence , turn off , and he clump stars , for a visual distance modulus @xmath10 , an age @xmath11 myr and a mass function slope @xmath12 , thus largely removing the `` classical '' discrepancy between observed and predicted number of stars in the he burning clump .
the quoted visual distance modulus constrains the unreddened distance modulus @xmath13 within 18.50 and 18.62 , depending on the reddening ( whose most common values available in the literature range from 0.05 to 0.10 ) .
\d ) at variance with point c ) , the fit obtained by using models computed with an enlarged convective core gets worse when a binary component is taken into account .
this is due to the fact that the presence of binary systems increases the existing discrepancy between the observed and predicted clump luminosity , since the he clump is predicted to be even less luminous than in absence of binaries . as a consequence of this analysis
, we think that the next step towards a proper understanding of ngc 1866 , and similar clusters , must include the accurate determination of the frequency of binary systems that will be hopefully performed with the incoming cycle 8 hst observations of ngc 1866 . |
the first gamma - ray burst ( grb ) ever discovered was in the year 1967 ( klebesadel et al .
1973 ) , but it took 30 more years to finally identify these sources as extragalactic and cosmologically distributed ( metzger et al .
1997 ) . as of today ( 2008 september 22 ) , the total number of grbs with known redshiftjcg / grbgen.html , maintained by j. greiner . ]
is 154 , half of which are at @xmath9 , and 69% were discovered by the dedicated space mission swift ( gehrels et al .
2004 ) after 2005 january .
although the grb population with redshift is rather small , studies dedicated to the hosting galaxies are often deep and can cover the wavelength range from the radio , to mid - ir , to optical , and uv in imaging and spectroscopy ( e.g. , bloom et al . 2002 ; le floch et al .
2006 ; priddey et al .
2006 ; prochaska et al . 2006 ; berger et al . 2007b ; ovaldsen et al . 2007 ) .
the typical nature of grb hosts is of a faint star - forming galaxy , dominated by a young stellar population ( christensen et al .
2004 ) , detected at any redshift from 0 to 6.3 ( berger et al .
luminosities are generally low ( chary et al . 2002 ; le floch et al .
2003 ) , indicating low masses and low metallicities ( gorosabel et al .
2005b ; wiersema et al .
2007a ; kewley et al . 2007 ) .
new cosmological simulations suggest that hosts associated with long grbs are representative of the whole galaxy population ( nuza et al .
2007 ) .
many hosts are fainter than the observational limits achieved today by the typical galaxy survey at high redshift ( see for instance cimatti et al .
2002 ; abraham et al . 2004 ; reddy et al . 2006 ; noeske et al .
in fact , the bright optical grb afterglow often facilitates obtaining a spectroscopic redshift and the mere presence of a grb encourages deeper photometric and spectroscopic observing campaigns than in conventional galaxy surveys .
it is not clear yet whether this faint population is stand - alone and characterized by the association with a grb event ( stanek et al .
2006 ; fruchter et al .
2006 ) , or whether we see many faint galaxies simply because these are the most common galaxies in the universe .
wolf & podsiadlowski ( 2007 ) showed that there is no dramatic difference between grb hosts and what is expected for a galaxy population tracing star formation .
this paper is dedicated to the study of the largest possible sample of grb host galaxies . in the past , many different tools applied to individual or a few grb hosts led to very heterogeneous results , not always easy to compare .
we used a compilation of grb hosts to measure a large number of galaxy parameters in a robust and consistent way , and try to establish the role of grb hosts in the cosmological scenario of galaxy formation and evolution .
the sample is selected by requiring that optical and/or near - ir ( nir ) photometry ( available in the literature ) have allowed the host identification . for a better stellar mass estimate ,
the important observable is the galaxy photometry redward of the 4000 break ( glazebrook et al .
2004 ) , which restricts our search to ( mainly ) grbs with @xmath10
. the total number of grb hosts for which the stellar mass is estimated is 46 ( figure [ f1 ] ) . for a subsample of 33 , rest - frame optical emission - line fluxes from spectra
are also available . in the past
, the stellar mass for several grb hosts was derived by chary et al .
( 2002 ) , castro cern et al .
( 2006 ) , and michaowski et al .
( 2008 ) . throughout the paper ,
we distinguish between the subsamples of galaxies originating in short grbs , which are associated with neutron star / black hole mergers possibly in evolved stellar populations , or long grbs , associated with core - collapsed sne and more abundant in young stellar populations ( woosley 1993 ) .
it has been argued by different authors that the hosts of the two classes are substantially different , the former in early- as well as late - type galaxies , the latter mainly in young star - forming galaxies ( e.g. , prochaska et al .
2006 ; berger et al .
long grb hosts are the vast majority ( seven out of eight ) , as the long duration allows an easier identification of the afterglow , whereas the detection of short afterglows is very recent ( gehrels et al .
2005 ) .
our group created a public database , where many grb - host observed parameters , available from the literature , are classified and stored .
the database , the largest of its kind , is called grb host studies ( ghosts ) and is accessible at the web site www.grbhosts.info .
ghosts is an interactive tool offering many features .
it also incorporates virtual observatory services , in particular sloan digital sky survey ( sdss ) and digital sky survey ( dss ) sky viewing .
the paper is organized as follows : in [ sample ] , we describe the sample selection ; in [ sp ] , multiband photometry and emission lines are used to derive galaxy parameters ; results are given in
[ gas ] and [ metallicities ] ; the discussion and summary are presented in [ discussion ] and [ summary ] . throughout the paper
we adopt an @xmath11 , @xmath12 , @xmath13 cosmology ( spergel et al . 2003 ) .
the sample selection is based on the requirement that multiband photometry is available for the grb host , mainly optical and nir photometry . some mid - ir detections from spitzer
are also included in our analysis ( le floch et al .
the sample is listed in table [ tsample ] , for which we report redshift , the type of grb ( short or long ) , and the morphological classification , as given by conselice et al .
( 2005 ) and wainwright et al .
the galactic color excess @xmath14 for each object ( table [ tsample ] ) is estimated using the reddening maps by schlegel et al .
( 1998 ) , and is less than 0.15 for 83% of the cases .
all measurements reported in this work are corrected for the galactic extinction ( cardelli et al .
figure [ f1 ] shows the number of objects in the sample per redshift bin and the comparison with the total sample of grbs with measured redshift .
we compare observed parameters , with the same parameters measured in field galaxies , observed at different redshifts .
in particular , we use results from the @xmath15 gemini deep deep survey ( gdds ; abraham et al .
2004 ) , the lyman break galaxies ( lbgs ; erb et al . 2006 ; reddy et al .
2006 ) , and the local dwarf galaxies ( lee et al . 2006 ) .
the gdds is an ultra - deep nir - selected survey ( @xmath16 , @xmath17 ) targeting galaxies in the `` redshift desert '' ( @xmath18 ) .
gdds is designed to find the most massive galaxies up to @xmath19 .
the limit in @xmath20 gives very high sensitivity to low - mass galaxies at @xmath21 ( e.g. , savaglio et al .
lbgs are uv - selected ( observed @xmath22 ) high-@xmath23 galaxies ( @xmath24 ) .
dwarfs in the local universe ( distance @xmath25 mpc ) have absolute magnitudes at 4.5 @xmath26 m in the interval @xmath27}<-13.3 $ ] .
measurements from gdds and the grb hosts are treated in the same way , in particular in the modeling analysis we use the initial mass function ( imf ) proposed by baldry & glazebrook ( 2003 ) . a correction for the different imfs ,
generally salpeter is assumed , is necessary for the other samples .
the grb - host sample studied here contains 46 objects ( 30% of all grbs with measured redshift ) .
six of these are associated with short grbs , all at redshift @xmath28 , with a mean redshift @xmath29(table [ tsample ] ) .
the mean and median redshift of the long - grb sample is @xmath30 and @xmath31 , respectively , with 88% being at @xmath32 , when the universe was 4 gyr old ( i.e. , 29% of the age today ) .
we note that the grb hosts that are considered have relatively low redshift , in comparison with the mean and median values for the total grb population with known redshift ( @xmath33 and @xmath34 , respectively ) .
this is because grb hosts are faint and hard to detect for @xmath35 , and also because redshifts in the pre - swift era were mainly determined from strong optical emission lines , redshifted in the more difficult nir regime for @xmath36 .
most grbs ( @xmath37% ) in our sample belong to the pre - swift era .
this can potentially bias the results that we discuss in this paper .
for instance , grbs are associated with regions of star formation .
as the star formation rate ( sfr ) density in the universe is a strong function of the galaxy stellar mass ( juneau et al .
2005 ) , it might be that grb hosts at low and high redshifts are associated with less- and more - massive galaxies , respectively . from the multiband photometry , we derive a ` pseudophotometry ' to create a homogeneous sample , as described in [ phot ] ( table [ tcol ] ) .
thirty - three out of 46 hosts have detected optical emission lines , indicating ongoing star formation ( table [ temi ] ) .
this is partly an observational bias , as the redshift of a grb is often determined from the emission lines in the host .
the bluest of these lines is the [ oii ] at @xmath38 .
as many hosts are spectroscopically observed in the optical , the highest redshift of the sample is marked by the [ oii]@xmath39 line , at @xmath40 ( figure [ f1 ] ) .
details are given in
[ spe ] .
the photometry of the sample is mainly covering the observed optical ( including the @xmath41 band ) and nir bands . for a subsample , mid - ir detections or upper limits
are also available ( le floch et al .
2006 ; berger et al .
2007b ) . only grb hosts detected in at least two bands are included in the sample , the reddest of which has to be above the 4000 break .
this means mainly @xmath42 hosts , the only exception being the host of grb 050904 at @xmath43 , for which only optical and mid - ir upper limits are available ( berger et al .
the total number of grbs with @xmath42 is 109 , two - fifths of which are in our sample .
the rest were not considered , because optical - nir photometry is not available , either because observations were never attempted or not deep enough .
the detection of the host above the 4000 break is necessary to determine one of the galaxy key parameters , the total stellar mass .
parameters derived from only uv photometry are very uncertain , because the mass - to - light ratio in the uv can vary by a very large factor in young stellar populations and due to the presence of dust .
the main difficulty of dealing with the photometric sample of grb hosts is that it is very heterogeneous and sparse .
for instance the data taken by different groups are treating galactic dust extinction or aperture corrections in different ways .
moreover , heterogeneous filters are often used . in a few cases the grb host
is observed in the same band with different telescopes , and results differ by more than the observational uncertainties . to reduce the confusion , from the observed multiband photometry we have derived a pseudophotometry " , which is a homogeneous photometry for a reduced set of filters .
magnitudes for the pseudophotometry in the ab system , corrected for galactic extinction , are reported in table [ tcol ] .
the principle of pseudophotometry " is to first define a large set of commonly used filters including sdss , hubble space telescope ( hst ) , bessel , johnson , and irac filters .
then , we reduce this set of filters to the filters which are necessary to represent all the photometric points involved in the available grb - host observations . to do so
, we use a criterion @xmath44 , or maximum wavelength difference of 300 at 5000 : the effective wavelengths @xmath45 of the filters involved in the real photometry must be close enough to one of the effective wavelengths @xmath46 included in the reduced set of filters . we thereby build a small set of 16 filters , ranging from @xmath41 band to irac 8@xmath26 m band , which we use in practice for our spectral energy distribution ( sed ) fitting . all the grb - host seds are therefore described with this set of filters .
the effect of using the reduced set of filters instead of the exact bibliographic filters is small for our study : the offset in effective wavelengths is typically less than 200 . for each grb host , we checked with monte carlo simulations that random shifts of this order on the filters wavelengths produce small errors on the mass estimates .
we note that not all observed magnitudes are in table [ tcol ] . for more details on the observed photometry , see the ghosts database .
the @xmath47 and @xmath48 magnitudes for the sample are shown in figure [ mag ] .
the comparison with the gdds @xmath49 galaxies shows that grb hosts are generally faint galaxies , with about half being fainter than @xmath50 and @xmath51 .
all short - grb hosts are at @xmath28 .
this could be a selection effect , as short afterglows are harder to detect than long afterglows ( gehrels et al .
2005 ) . in figure
[ col ] , we show the apparent colors of the sample .
the comparison with a complete sample of gdds galaxies , and the predicted colors assuming different stellar populations ( e / s0 , sbc , irregulars , and starbursts ) , indicates that grb hosts are generally blue star - forming galaxies ( see also berger et al .
the short - grb hosts are still too few to conclude anything about their stellar population from their colors
. fluxes of optical emission lines from [ oii]@xmath39 to the [ sii]@xmath52 doublet , originating in the star - forming ( or hii ) regions are measured for a subsample of 33 grb hosts and corrected for galactic extinction ( table [ temi ] ) .
errors are reported when available .
the [ oii ] and h@xmath53 lines are detected up to @xmath54 and @xmath55 , in 31 and 11 hosts , respectively .
h@xmath56 and [ oiii]@xmath57
are both detected in 19 hosts . the [ neiii]@xmath58 line is present in 13 hosts ; the [ neiii]-to-[oii ] flux ratio is @xmath59 in six hosts .
such relatively strong values indicate high temperature or high ionization , which is expected in the presence of hot massive stars .
strong [ neiii ] is common in hii regions and blue compact dwarf galaxies ( stasiska 2006 ) .
contamination from active galactic nucleus ( agn ) is unlikely , as seen from emission - line ratios ( section 5.6 ) .
the true emission - line fluxes can be several times higher than those measured .
some corrections are necessary in order to have a better estimate of the total sfr , metallicity , and dust extinction in the hosts .
the slit - aperture flux loss is determined for each host individually .
it depends on the size of the host with respect to the slit aperture used and the seeing conditions during observations .
the aperture correction is reported in the last column of table [ temi ] , but it is not applied for the fluxes given in this table .
this is the factor we have to multiply the observed flux to obtain a more realistic total flux .
it is ` 1 ' ( no correction necessary , almost half of the hosts ) either when it is negligible , or it is already applied by authors of the papers from where fluxes are taken , or the redshift is high enough ( galaxies get smaller ) for the given slit ( generally 1 arcsec ) . in all other cases
it is @xmath60 and derived by us , mainly comparing the observed multiband photometry with the flux - calibrated host spectrum , or it is estimated by considering the galaxy size in comparison with the slit aperture used and the seeing condition during observations .
one arcsec corresponds to a physical size from redshift @xmath54 to @xmath61 ( the highest and lowest redshift in the sample ) of 8.4 to 1.9 kpc , respectively .
the aperture correction is @xmath60 and @xmath62 for eight hosts . for the remaining 5 ,
it is @xmath63 and @xmath64 .
of course our aperture correction is not perfect , but it is the best that can be done , given the inhomogeneity of the sample .
dust extinction in the host ( see [ dust ] for details ) is determined from the balmer decrement for the subsample with simultaneous h@xmath56 and h@xmath53 detection ( 10 grb hosts , table [ tbal ] ) , and is calculated assuming a gas temperature @xmath65 k. the dust extinction is not applied for fluxes in table [ temi ] , but is applied when estimating sfrs .
the stellar absorption correction , generally more important in evolved stellar populations , is estimated only for the balmer lines , as it is typically negligible for the metal lines . in local irregular and spiral galaxies ,
the typical stellar balmer absorption equivalent width ( ew ) is 3 , with 2 uncertainty ( kobulnicky , kennicutt , & pizagno 1999 ) .
it is lower in extragalactic hii regions , ew@xmath66 ( mccall et al . 1985 ) .
to estimate the stellar balmer absorption , we inspected the observed host spectrum . in some cases
it is clearly negligible ( e.g. , the emission - line flux is high with respect to the stellar continuum ) and no correction is applied . in all other cases
it is generally estimated by adding 2 to the measured , rest - frame equivalent width of the lines . in one case ( the host of grb 060505 ) , the stellar correction is the one estimated by thne et al .
( 2008 ) . in table
[ tbal ] , we list the intrinsic - to - observed flux ratios ( after and before stellar absorption correction ) for balmer emission lines . for h@xmath53 ,
the correction is up to 13% of the observed emission flux ( mean value 4% ) . for h@xmath56 and h@xmath67
it is clearly higher , up to 30% and 56% ( mean 10% and 25% ) , respectively ( table [ tbal ] ) . the stellar absorption correction , although generally not very important , was never applied before in grb - host studies . for the remaining of the paper
, we will use emission fluxes corrected for stellar absorption and aperture - slit loss unless otherwise stated .
the stellar properties of grb hosts in the sample have been studied using the observed multiband photometry . following the procedure of glazebrook et al . ( 2004 ) , we model the stellar population using pgase.2 ( fioc & rocca - volmerange 1997 ; 1999 ) and pgase - hr ( le borgne et al . 2004 ) . in order to avoid distorted mass - to - light ratios from young burst components superimposed on older stars ,
we represent the stellar population by two components , with sfr @xmath68 , where @xmath69 is time and @xmath70 is the @xmath71-folding time .
the first component is one of 10 alternate seds , each characterized by a different discrete value of @xmath70 ( from starburst to constant sfr ) , covering a range of star formation histories ( sfhs ) .
the primary stellar component is coupled with a secondary component representing a star - bursting episode , with @xmath72 gyr , and a total stellar mass in the range 1/10,000 up to two times the mass of the primary component .
the age of the stellar population is also constrained by the age of the universe at the observed galaxy redshift .
the attenuation of the stellar emission , due to dust , is a free parameter and can vary in the range @xmath73 ( the calzetti law is used ) .
the same for the metallicity , in the range @xmath74 .
the modeling also includes an emission - line nebulosity component which we include as it can have small effects on the photometry near bright lines .
we use the standard pgase prescription for this and note that this is only used in the sed modeling and not in the interpretation of real galaxy emission - line spectra .
the imf used to estimate stellar masses is that which is derived by baldry & glazebrook ( 2003 ) .
this more realistic imf gives a total stellar mass which is generally 1.8 and 1.3 times lower than that derived by using salpeter ( 1955 ) and kroupa ( 2001 ) imf , respectively .
the full distribution function of allowed masses was calculated by monte carlo resampling the photometric errors .
the monte carlo error approach is to : ( 1 ) apply a scaled poisson error distribution of the fluxes and flux - errors ( this is superior to using a gaussian distribution as fluxes come from photon counting detectors ) ; ( 2 ) create a new sed by flux and poisson realization ( flux - error ) and re - fit the masses ; ( 3 ) iterate @xmath75 times to make a distribution .
the resulting mass distribution then samples the range of sfhs consistent with the error distribution of the photometry .
the final masses and errors are the mean and standard deviation of the monte carlo mass distribution .
our code is more general than most in the literature because we include a secondary burst component .
this is advantageous as it allows for possible mass biases associated with young bursts superimposed on old stellar populations to be explored ( by turning off the burst term ) . in general
we find that masses are robust to turning on / off bursts , and varying metallicities and dust , and are good to the @xmath76 @xmath77 level
. with this code then the mass of grb hosts with poor photometry ( e.g. , few data points or photometry not sampling redder rest frame wavelengths ) is poorly constrained , and this is reflected correctly as large error bars in table [ tsed ] .
the requirement that one or more photometric detections above the 4000 balmer break is known , and that at least two photometric bands are used , provides a reasonably robust total stellar mass determination . at @xmath784000 , the seds are far less affected by younger stellar populations and uncertain dust modeling .
figure [ fsed1 ] shows the best fit to the observed sed .
the stellar mass - to - light ratio in the @xmath20 band @xmath79 and the stellar mass for each grb host are listed in table [ tsed ] , together with the absolute @xmath20 and @xmath80 magnitudes .
the age of the last episode of star formation , the metallicity , the sfr and dust extinction in the stellar component are also calculated .
however , these parameters are more uncertain due to the well - known degeneracy , according to which the colors of metal - rich stellar populations are indistinguishable from those of old or dusty stellar populations ( worthey 1994 ) .
our method is more general than previous work which make assumptions which speed up the calculation ( for example by assuming a single simple sfh or a single metallicity ) .
in contrast with the method adopted by christensen et al .
( 2004 ) , our metallicity is given by the best fit , and not assumed to be solar .
moreover , christensen et al .
( 2004 ) assume a single episode of star formation and do not derive the total stellar mass . for the stellar mass fitting ,
we have utilized a very robust code which has been applied and tested in the literature extensively ( e.g. , glazebrook et al .
2004 ; baldry , glazebrook , driver 2008 ) .
stellar masses and ab absolute @xmath20 and @xmath80 magnitudes are shown in figure [ all ] and table [ tsed ] .
magnitudes are corrected for dust attenuation in the host ( a calzetti law is assumed ) . in figure
[ all ] we show the comparison with field galaxies , from the gdds ( glazebrook et al .
2004 ) and lbgs ( reddy et al .
the grb - host median stellar mass is @xmath3 m@xmath4 , and is much lower than in the gdds , by about one order of magnitude .
for short - grb hosts , the median mass is higher , @xmath81 m@xmath4 .
more objects are necessary to conclude that the typical mass of short - grb hosts is larger than that of long - grb hosts .
the curves in the upper plot of figure [ all ] represent the stellar mass as a function of redshift of a galaxy with ab @xmath20 observed magnitude @xmath82 , in the case of a old simple stellar population or a young stellar population with constant sfr .
in general , the sample is too small and inhomogeneous to study the mass function of the grb - host population .
nevertheless , we inspected the fraction of grb hosts per stellar mass bin . in figure
[ mass_hist ] , we compare this with that of the high-@xmath23 lbg sample ( reddy et al .
2006 ) , after normalizing it to the grb - host sample for galaxy stellar mass @xmath83 m@xmath4 , where the lbg sample is less affected by incompleteness . from the grb - host stellar mass histogram ( which drops for @xmath84 m@xmath4 ) it is clear that galaxies similar to the typical grb host are under - represented in high-@xmath23 spectroscopic galaxy surveys . as already described before , the stellar mass of galaxies is better represented by the observed fluxes redward of the balmer break at 4000 .
the rest - frame @xmath20-band luminosity is generally well correlated with the galaxy stellar mass .
this correlation is also shown by grb hosts ( figure [ mk ] ) , and can be expressed by @xmath85 where @xmath86 is the dust - corrected @xmath20-band absolute magnitude .
this relation can be used to estimate the stellar mass of grb hosts , provided that the sed is known redward of the balmer break .
the dispersion of the sample in figure [ mk ] around eq .
[ emk_m ] is about a factor of 2 , and gives the minimum error on the stellar mass that one obtains by using this procedure .
such an error does not take into account systematic uncertainties on the modeling of the stellar mass and errors on the measured photometry .
we note that on average grb hosts have higher @xmath20-band luminosities than field galaxies ( from gdds ) with the same stellar mass ( figure [ mk ] ) .
this means that @xmath79 is lower , which is expected from younger galaxies ( glazebrook et al .
2004 ) , suggesting an intrinsic difference between the two galaxy populations .
we compared our results with the stellar masses derived for seven hosts by chary et al .
( 2002 ) and another seven hosts by castro cern et al .
we find a weak correlation in the former , and no correlation at all in the latter , although the sample is too small to draw any conclusion .
optical emission lines line has not been considered because it is rarely detected ( kulkarni et al .
1998 ) , and , more importantly , because of its strong attenuation due to resonant scattering by the hydrogen atoms . ]
detected in a subsample of 33 grb hosts ( redshift range @xmath87 ) are used to estimate sfrs , metallicities , and dust extinction in the gas component . these are all ( but one ) long grb hosts .
the exception is the host of the short grb 051221 ( soderberg et al .
stellar absorption and slit - aperture flux loss ( table [ tbal ] ) are taken into account to estimate the total emission flux .
the dust extinction in the hosts is obtained as described below , and is used when measuring sfr and metallicity .
the final adopted sfrs for grb hosts are listed in the eighth column of table [ tsfr ] .
it is the value measured from h@xmath53 when available , or [ oii ] otherwise .
if no emission line is detected , uv luminosities at 2800 are used .
details are in [ sfremi ] and [ sfruv ] . the dust extinction in the gas component is estimated using the expected balmer decrement ( h@xmath53-to - h@xmath56 ratio ) in absence of dust , for case @xmath80 recombination at a gas temperature of @xmath88k ( osterbrock 1989 ) . we adopt the milky way ( mw ) extinction law , equivalent to large magellanic cloud ( lmc ) and small magellanic cloud ( smc ) extinctions in the optical range .
if the real grb - host extinction law is flatter than in the mw , then our dust - extinction correction is underestimated .
the observed balmer decrement is available for 10 grb hosts ( table [ temi ] ) .
extinctions in the visual band @xmath89 for these hosts are derived after correcting for stellar absorption and are reported in table [ tbal ] .
the mean value is @xmath90 ( a dispersion of 0.54 ) . for the other grb hosts with no balmer decrement measurement
, we used the mean value @xmath90 . for six of these hosts , h@xmath67 and h@xmath56 are detected simultaneously , thus the line ratio can in principle be used to estimate @xmath89 .
however h@xmath67 is much more uncertain and errors are too large , thus we ignore this method .
we note that measuring dust reddening in the gas component using emission lines does not correspond to measuring the dust attenuation of the stellar component ( calzetti 2001 ) . from emission lines
one can estimate the dust extinction valid for point sources . in our case ,
the hii regions are considered point sources .
galaxies are not uniformly obscured , because the distribution of dust is clumpy .
we can only apply an approximate ` mean ' extinction correction , or dust attenuation correction .
the dust attenuation applied to the stellar component is generally the calzetti law .
the visual extinction derived from emission lines of the hii regions is generally larger than the visual attenuation derived for the stellar component .
a ratio of about 2 is obtained for two galaxies in the local universe by cid fernandes et al .
similar results were obtained by calzetti et al .
( 2000 ) using eight star - forming galaxies .
the relation between attenuation and extinction can depend on the galaxy mass and/or sfr .
although we found that in our sample the gas extinction derived from balmer lines is higher than the stellar dust attenuation derived from the sed best fit , we found no relation between the two parameters .
we note , however , that the stellar dust attenuation is very uncertain , because it is degenerate with metallicity and age of the stellar population .
traditionally , the best sfr indicator available for the optical - uv is the h@xmath53 luminosity , corrected for dust extinction and stellar absorption .
otherwise , the more uncertain [ oii ] or h@xmath56 can be used instead .
the conversions derived by kennicutt ( 1998 ) are the most commonly used .
moustakas , kennicutt & tremonti ( 2006 ) also provided conversions , valid for local star - forming galaxies .
we used the dust - corrected [ oii ] , h@xmath56 , and h@xmath53 luminosities .
a mean visual extinction @xmath91 corresponds to an sfr correction for [ oii ] , h@xmath56 , and h@xmath53 fluxes of a factor of 2.1 , 1.8 , and 1.5 , respectively ( for mw extinction law ) , is applied when the balmer decrement is not measured .
the balmer stellar absorption correction is on average 4% and 10% of the observed flux , for h@xmath53 and h@xmath56 , respectively ( table [ tbal ] ) .
the slit - aperture flux loss ( up to a factor of 5 , table [ temi ] ) is also taken into account . to derive sfr from h@xmath53 , we used the prescription given by kennicutt ( 1998 ) .
we only correct for the different imf adopted , which gives an sfr a factor of 1.8 lower , going from salpeter to the more realistic imf proposed by baldry & glazebrook ( 2003 ) .
this gives @xmath92 where @xmath93 is the h@xmath53 luminosity corrected for stellar absorption and dust extinction .
equivalently , the sfr from h@xmath56 can be expressed by @xmath94 where @xmath95 is the dust and stellar absorption corrected h@xmath56 luminosity , and the multiplicative constant 12.6 is the one in eq .
[ sfrha ] multiplied by 2.86 ( the recombination factor assuming @xmath65 k ; osterbrock 1989 ) .
the advantage of using h@xmath56 instead of [ oii ] to estimate sfr is that , to first order , it does not depend on metallicity .
however , the stellar balmer absorption is generally uncertain , and important for old stellar populations . on the other hand [ oii ] is generally stronger and easier to detect than h@xmath56 , but it is more affected by an at times undetermined dust extinction .
moreover , the [ o ii ] flux depends on metallicity and ionization level ( kewley et al . 2004 ) .
moustakas et al . ( 2006 ) found that the mean dust - corrected [ oii]-to - h@xmath53 flux ratio in local star - forming galaxies is about 1 , with negligible dependence from the galaxy luminosity .
sfrs based on the observed [ oii ] luminosity are 0.4 dex uncertain .
the metallicity , from asplund et al .
] dependence is weak ( @xmath96/h\alpha)=-0.1 $ ] to 0.2 ) in the metallicity range @xmath97 , and becomes linear ( in log - log scale ) for @xmath98 , down to @xmath99 , where @xmath100/h\alpha)\sim-1.0 $ ] . for galaxies with relatively high metallicity ,
the [ oii ] sfr is rather robust . in nine
grb hosts the dust - corrected ratio is relatively low , in the range @xmath101/h\alpha)<0.0 $ ] , which can be explained by the generally low metallicities ( see [ metallicities ] ) .
metals are an important source of radiative cooling , which gives the relation between metallicity and [ oii]/h@xmath53 ( kewley et al .
we note the different mean redshift for the grb hosts and moustakas et al .
( 2006 ) sample ( @xmath102 and @xmath103 , respectively ) .
however , this is likely too small to sensibly affect the results .
we derived an [ oii]sfr relation valid for grb hosts , by considering the subsample with simultaneous [ oii ] and h@xmath53 detection .
the [ oii]sfr best conversion is empirically derived by determining the best concordance between the sfrs estimated from h@xmath53 and [ oii ] ( figure [ sfr_oii ] ) .
this is given by @xmath104 } = 5.54\times10^{-42 } \frac{l(\rm [ oii])_{corr}}{erg\ s^{-1 } } m_\odot\ yr^{-1},\ ] ] where @xmath105_{corr}$ ] is the dust - corrected [ oii ] luminosity .
the dispersion in figure [ sfr_oii ] is 0.22 dex ( a factor of 1.7 ) .
[ sfroii ] is proposed as a valuable tool to estimate sfrs for grb hosts when h@xmath53 or h@xmath56 are not detected .
[ sfroii ] gives an sfr which is a factor of 1.39 lower than the same relation proposed by kennicutt ( 1998 ) , after converting to the same imf .
to check the validity of our [ oii ] sfr conversion , we studied the sample of 18 grb hosts with simultaneous measurement of [ oii ] sfr and h@xmath56 sfr ( figure [ sfr_hb ] ) .
the correlation is very good , and is basically indistinguishable from the best - fit linear correlation .
the dispersion in this sample is 0.26 dex ( a factor of 1.8 ) .
when h@xmath53 is not measured , we estimated sfr from [ oii ] using eq .
[ sfroii ] ( table [ temi ] ) .
sfr from h@xmath56 is also estimated for a subsample of grb hosts ( table [ temi ] ) .
the slit - aperture correction is not necessary when deriving sfr relations , but it is required when estimating the total sfr .
this correction is uncertain , because it depends on the galaxy size and how the star - forming hii regions are distributed in the galaxy .
all sfrs derived from emission lines ( table [ tsfr ] ) are corrected for aperture - slit loss ( given in table [ temi ] ) . for 13 grb hosts , there is no information on emission - line luminosities from hii regions .
this is because either the redshift of the grb is too high , so significant emission lines are shifted to the nir ( where observations are much harder ) , or lines are intrinsically too weak , or simply the host was never spectroscopically observed
. the uv luminosity can be used instead to estimate sfr , as this is a tracer of almost instantaneous conversion of gas into stars .
however , we should say that residual uv emission can still be detected even in the absence of star formation , for instance from blue horizontal branch stars .
uv luminosities were used several times in the past to study the sfr history of the universe ( madau , pozzetti , & dickinson 1998 ; glazebrook et al .
1999 ; meurer et al .
1999 ; erb et al . 2006 ; moustakas et al .
the conversion often used is the one provided by kennicutt ( 1998 ) at 1500 .
moustakas et al .
( 2006 ) use the @xmath41-band luminosities at 3600 , as this is less affected by dust .
we derived relations between uv luminosities and sfr more suitable for grb hosts , using the same empirical approach as for [ oii ] .
we compare sfrs derived from emission lines , with uv luminosities obtained from the best fit of the observed sed ( figure [ fsed1 ] ) .
we obtain @xmath106 @xmath107 @xmath108 for dust - corrected rest - frame 1500 , 2800 and 3600 luminosities , respectively .
results for 33 grb hosts are shown in figure [ sfr_uv ] and given in table [ tsfr ] ( for @xmath109 only ) .
the dispersions of the points around the three relations are 0.49 , 0.41 and 0.40 dex , respectively .
this is relatively small , given the large interval spanned by the sfrs ( 3.5 orders of magnitude ) and the uncertainties of the uv sfr relations ( glazebrook et al .
the dispersion found by erb et al .
( 2006 ) , who derived sfrs using h@xmath53 and uv at 1500 and the kennicutt ( 1998 ) conversion , is 0.34 dex .
moustakas et al .
( 2006 ) used the uv luminosities at 3600 , on the subsample of galaxies with @xmath80 absolute magnitude @xmath110 . when converted to take into account the different imf and the monochromatic luminosity , the constant in our eq .
[ esfr_uv3 ] is a factor of 2.0 larger than theirs .
this factor is 1.5 if dwarf galaxies only are considered ( @xmath111 ) .
meurer et al .
( 1999 ) estimated an sfr conversion in local starbursts using uv luminosities at 1600 .
after correcting for the same imf , eq .
[ esfr_uv1 ] gives sfrs which are 2.7 times higher than theirs .
these differences can be explained by variations in sfh , dust attenuation , and redshift among the samples ( effectively , different sample selections ) .
[ esfr_uv1]-[esfr_uv3 ] , proposed for grb hosts , can be used to estimate sfr when no emission lines are detected , likely for low sfrs or @xmath112 .
the good agreement between sfrs from emission lines and sfr@xmath113 is surprising as one expects these longer continuum wavelengths to be increasingly affected by older stellar populations ( madau , pozzetti , dickinson 1998 ) .
the fact that the dispersion is still small indicates again that grb galaxies are dominated by young stellar populations .
we note that 3600 is much less affected by dust than 1500 , but likely marginally affected by uv photons from older stars .
the best measure of sfr , on balance , might be that derived at 2800 . in figure [ sfr_uv_z ]
, we show sfr@xmath114 and sfr derived from emission lines , as a function of redshift .
uv luminosity are very useful to estimate sfrs at @xmath115 , and in general are more sensitive to low sfrs .
in the same figure also sfrs from gdds galaxies and lbgs ( erb et al .
2006 ) .
an interesting finding is a possible correlation between the sfr and the stellar mass ( figure [ sfr_m ] ) .
this has not been seen before in high - redshift or local galaxy samples .
if confirmed , it might be peculiar to the grb - host sample , and suggest an approximately constant specific sfr ( ssfr ) . whether or not it has physical significance needs to be determined from more extensive data .
the sfr by itself does not tell how active the galaxy is .
another parameter to consider is the ssfr , that is the sfr normalized to the total stellar mass of the galaxy .
the ssfr has been studied in galaxies at low redshift ( prez - gonzlez et al .
2003 ; brinchmann et al .
2004 ) and high redshift ( juneau et al . 2005 ; bauer et al . 2005 ;
noeske et al . 2007 ) .
in figure [ ssfrm ] we show the ssfr as a function of stellar mass , investigated for the first time for a large sample of grb hosts .
these are compared with ssfrs of star - forming galaxies of the gdds in the redshift interval @xmath116 ( juneau et al .
2005 ; savaglio et al .
2005 ) and lbgs at @xmath117 ( reddy et al .
field galaxies of the aegis at @xmath118 cover a similar region in the ssfr@xmath119 plot as the gdds ( noeske et al .
the ssfrs in grb hosts tend to show properties different from field galaxies in the gdds , with lower masses and higher ssfrs .
the median ssfr for the grb hosts is 0.8 gyr@xmath7 , similar to lbgs , but the mean stellar mass is six times lower , @xmath3 m@xmath4 for the former and @xmath120 m@xmath4 for the latter . the inverse of the ssfr is called the growth timescale @xmath121 .
it defines the time required by the galaxy to form its observed stellar mass , assuming that the measured sfr has been constant over the entire galaxy history . in figure [ ssfr ]
we show @xmath121 ( left @xmath122-axis ) or ssfr ( right @xmath122-axis ) as a function of redshift .
the symbol size is bigger for larger stellar masses . for all but one grb host with detected emission lines
, @xmath121 is smaller than the age of the universe ( hubble time ) at the observed redshift .
this is true for four more grb hosts , with sfr estimated from uv luminosities . for the total sample of 46 hosts ,
the median value is @xmath123 gyr , and for 2/3 @xmath124 gyr .
it is also apparent that the growth time - scale is longer for more - massive galaxies .
this is different than what has been observed in other galaxy samples , which have a large fraction of quiescent galaxies ( @xmath121 larger than the hubble time ) . when @xmath121 is @xmath125 gyr , true for about 2/5 of the grb - host sample , the galaxy is in a
bursty mode " .
for instance , a galaxy with a stellar mass of a few times @xmath126 m@xmath4 ( the stellar mass of the lmc ) and sfr = 5 m@xmath4 yr@xmath7 ( 10 times larger than lmc ) is a starburst , with @xmath127 myr .
metallicity in hii regions is typically measured using detected emission lines . at redshift @xmath128 ,
the spatial resolution of the data is generally too low to derive metallicity in small regions , so what is measured from integrated fluxes is an optical luminosity - weighted mean value in the galaxy .
this is the case for most grb hosts analyzed here , as 85% of the spectroscopy sample is at @xmath128 . for low-@xmath23 hosts , like that of grb 060505 at @xmath129 studied in great detail by thne et al .
( 2008 ) , we use integrated fluxes over the entire galaxy to treat it as the rest of the sample .
metallicities can be derived from different emission - line sets , according to the spectral coverage , redshift , and galaxy properties .
the numerous methods , developed in the last few years , have complicated the effort of having a tool for a realistic estimate . each method
is affected by systematic errors that are not easy to determine ( for a review , see kewley & ellison 2008 ) .
a direct estimate is possible through a measurement of the electron temperature @xmath130 ( izotov et al .
2006 ) , possible when several lines of the same element with very different ionization levels are detected ( bresolin 2007 ) .
for instance , these could be the generally weak auroral line [ oiii]@xmath131 and the nebular lines [ oiii]@xmath57 .
for its definition , the @xmath130 method is sensitive to low - metallicity systems . as this is based on the generally weak [ oiii]@xmath131 line , it is very hard to measure in small and distant galaxies
. other calibrators are easier to measure , but give more systematically uncertain results .
the o3n2 calibrator , which uses a combination of [ oiii]@xmath132 , [ nii]@xmath133 , h@xmath56 , and h@xmath53 lines ( alloin et al . 1979 ) , needs nir spectroscopy when exceeding redshift @xmath134 . for higher redshift and relatively low spectral sensitivity , the very popular @xmath135 calibration ( pagel et al . 1979 ) and
its refinement , the @xmath136-method ( pilyugin 2000 ) , both requiring h@xmath56 , [ oii ] , and [ oiii ] fluxes , can be used up to redshift @xmath137 .
the @xmath135 calibration is notoriously problematic .
first , it gives two solutions ( lower and upper branch ) , which are not easy to disentangle .
moreover , the upper branch solution , when applied to integrated fluxes , is affected by systematic errors which could be as large as a factor of 3 ( kewley & dopita 2002 ; kennicutt et al . 2003 ; stasiska 2005 ;
bresolin 2006 ) . a value of @xmath138/[oii ] } ) < -1.2 $ ] can point to the lower branch solution ( kewley & dopita 2002 ) . despite the difficulties , the @xmath135 calibrator is an important resource because it is easier to measure than other methods .
our choice for the final metallicity is done as follows : the @xmath130 metallicity is preferred , measurable for four grb hosts with detected [ oiii]@xmath131 .
when this is not detected , we give priority to the o3n2 metallicity , available for five grb hosts , and use the prescription given by pettini & pagel ( 2004 ) . for all other grb hosts
, we used the @xmath135 calibrator , available for a total of 18 hosts .
recently , kewley & ellison ( 2008 ; hereafter ke08 ) derived conversion relations between different methods .
following their prescriptions , we used the @xmath135 calibrators given by kewley & dopita ( 2002 ) and kobulnicky & kewley ( 2004 ) , for the lower and upper branch solutions , respectively .
we converted @xmath135 metallicities into o3n2 metallicities using the relations proposed by ke08 .
these are tested for metallicities in the ranges @xmath139 and @xmath140 , for the lower and upper branch solution , respectively . for metallicities outside these ranges
, we use the recently tested @xmath135metallicity relation proposed by nagao et al .
( 2006 , hereafter n06 ) . the hosts for which we could measure metallicity are all at @xmath141 . for @xmath142 , no nir spectra of grb hosts
are good enough to measure metallicity . in the following subsections
, we list measurements for different methods . the temperature sensitive [ oiii]@xmath1434363/[oiii]@xmath132 line ratio has been detected in four hosts ( table [ tte ] ) .
these are the hosts of grb 980425 , grb 020903 , grb 031203 , and grb 060218 ( hammer et al
. 2006 ; prochaska et al .
2004 ; wiersema et al .
though according to wiersema et al .
( 2007 ) the [ oiii]@xmath1434363 line in grb 060218 is 5@xmath5 significant , the line in the spectrum is somehow doubtful , because located in a noisy part of the spectrum .
we will consider this detection with caution .
we recalculated the oxygen abundance using the prescriptions provided by izotov et al .
( 2006 ) , which are not very sensitive to the electron density , provided that this is smaller than @xmath144 @xmath145 .
metallicities derived with this method are low ( table [ tte ] ) .
temperatures in the hii regions are slightly larger than @xmath88 k ( table [ tte ] ) , with the exception of the host of grb 060218 ( wiersema et al .
2007 ) , for which the temperature is twice larger . however , as already said , the detection of o@xmath131 in this host is weak . the o3n2 method , originally proposed by alloin et al .
( 1979 ) , was updated by pettini & pagel ( 2004 ) . in the new formulation ,
the metallicity is given by @xmath146 where @xmath1475007 / h\beta)/([nii]6583 / h\alpha)\}$ ] .
this method is useful only for @xmath148 , and much more uncertain for @xmath149 .
this metallicity can be measured in five grb hosts , and in two additional objects an upper limit is set ( table [ to3n2 ] ) .
derived values are low , with a mean metallicity of 1/3 solar .
the @xmath135 calibrator uses the following combination of [ oii ] , [ oiii ] and h@xmath56 lines : @xmath150\lambda3727+[oiii]\lambda\lambda4959,5007}{\rm h\beta}\right ) , \ ] ] and the @xmath151 parameter : @xmath152\lambda\lambda4959,5007}{\rm [ oii]\lambda3727}\right ) , \ ] ] which takes into account different ionization levels of the gas . in figure [ r23 ]
we show @xmath135 and @xmath151 for grb hosts , and comparison with galaxies of the gdds and canada france redshift survey ( cfrs ; lilly et al .
2003 ) at @xmath153 , before dust - extinction correction ( the dust extinction in gdds and cfrs galaxies is not measured ) .
grb hosts tend to have higher @xmath135 values , which could be due to low metallicity or / and lower dust extinction .
to measure @xmath135 metallicities , we followed the prescriptions given by ke08 .
namely , we used kewley & dopita ( 2002 ) and kobulnicky & kewley ( 2004 ) for the lower and upper branch solutions , respectively .
then we used the relations given by ke08 to convert to o3n2 metallicities .
these relations are valid for @xmath154 and @xmath155 only ( for the input metallicity ) . see [ nagao ] for more @xmath135 metallicities .
the @xmath135 and @xmath151 values used are corrected for dust extinction , and are reported in table [ tr23 ] , together with the derived metallicities . for metallicities outside the intervals @xmath154 and @xmath155 we used the @xmath135metallicity relation provided by n06 .
this is described by a best - fit polynomial , whose coefficients are in table 6 of n06 .
six grb hosts have @xmath135 in the turnover region , for which @xmath156 ( table [ tr23 ] ) . for the other grb hosts , in the case of the lower branch part of the relation , we estimated the correction from n06 to ke08 , by comparing results for those hosts with metallicity calculated with both calibrations . the conversion relation from n06 to ke08
is given by @xmath157 + 1.177 , \ ] ] valid in @xmath158 for the n06 metallicity .
results for the lower branch metallicities , after applying eq .
[ enagao ] , are given in the sixth column of table [ tr23 ] .
the upper branch solution of n06 is not used , as this does not add anything to our analysis . the [ oii ] doublet at @xmath159 is sensitive to the electron density ( osterbrock & ferland 2006 ) .
a low density was measured in the host of grb 060218 ( wiersema et al .
thanks to the use of the high - resolution spectroscopy , [ oii]@xmath160 is resolved in 6 grb hosts . assuming that the typical temperature in the star - forming regions is around @xmath88 k ( table [ tte ] ) , we derive the electron density using the definition of osterbrock & ferland ( 2006 ) .
results are in the range @xmath161 @xmath145 to @xmath162 @xmath145 ( table [ tne ] ) . figure [ agn ] shows the location of grb hosts , in the diagrams that use line ratios to distinguish galaxies with bright hii regions from agn dominated galaxies . for comparison , we also show gdds and cfrs star - forming galaxies at high redshift , and local metal - poor galaxies ( izotov et al .
2006 ) .
the [ oiii]/h@xmath56 versus [ nii]/h@xmath53 relations proposed by kewley et al .
( 2001 ) and kauffmann et al .
( 2003 ) are relatively robust diagnostics and generally used in galaxy surveys .
the [ oiii]/h@xmath56 versus [ oii]/h@xmath56 relation is more uncertain ( lamareille et al .
however , this is easier to measure at @xmath163 , when nir spectra are not available . from figure [ agn ]
, we conclude that the agn contamination in grb hosts is likely not significant .
the [ oiii]/h@xmath56 versus [ nii]/h@xmath53 plot indicates that grb hosts behave like local metal - poor galaxies , i.e. with high values of [ oiii]/h@xmath56 . in fact , the majority of galaxies in the local universe from the sdss have [ oiii]/h@xmath164 ( stasiska et al . 2006 ) .
the main goal of this work is to characterize the galaxy population hosting grbs . in the past ,
several authors conducted studies over a small sample of grb hosts and concluded that this population is not representative of the bulk of all galaxies as a function of cosmic time ( fruchter et al .
1999 ; le floch et al . 2003 ; christensen et al . 2004 ; berger et al .
2003 , tanvir et al . 2004
; fruchter et al . 2006 ; le floch et al . 2006 ) .
in fact , they are on average bluer , younger , and fainter than the field galaxy population . however , we consider that faint , blue , and young galaxies are by far the most common galaxies that existed in the past .
galaxies can only get bigger over time and not smaller , through merging and star formation processes .
nevertheless , they are the most elusive to find in the distant universe . through a grb event , our capabilities of digging deeper in terms of galaxy mass and distance
is considerably increased .
fruchter et al . ( 2006 ) showed that long grbs happen much more in the brightest regions of their host galaxies than type ii sne , suggesting that grbs are associated with different galaxies . this result
is not confirmed by kelly et al .
( 2008 ) , who considered sn ic hosts and found that sn ic occur in an environment similar to that of grbs . indeed , some nearby grbs are associated with sn ic , a subclass of core - collpase supernovae ( galama et al .
1998 ; stanek et al . 2003
; hjorth et al . 2003 ; modjaz et al . 2006 ) .
the connection between grb events and the nature of their hosts is advocated by modjaz et al .
( 2008 ) , who found that five grb hosts are significantly more metal poor than sample of sn ic hosts in a similar luminosity range in the @xmath80 band .
however , sn discoveries are observationally biased toward luminous ( and hence on average more metal rich ) galaxies .
the result of modjaz et al .
( 2008 ) is affected by the small number statistics .
prieto et al . ( 2008 ) found in a statistically significant sample a much larger metallicity spread , and a significant mass metallicity relation which would include grb hosts .
moreover , one should consider that the @xmath80-band absolute luminosity in galaxies is weakly correlated with mass .
the @xmath20-band luminosity would instead better represent the galaxy stellar mass ( figure [ mk ] ) , and this would better support the special nature of grb hosts .
wolf & podsiadlowski ( 2007 ) quantitatively showed that any grb rate dependence on galaxy mass ( or ultimately metallicity ) is at best small .
to better understand the impact of grb hosts on our knowledge of galaxy formation and evolution , we have analyzed the properties of the largest possible sample , 46 grb hosts , using multiband photometry and optical spectroscopy .
relevant parameters are summarized in table [ tsummary ] .
the median redshift , @xmath20-band absolute magnitude , stellar mass , dust extinction , metallicity , sfr , and ssfr are reported at the bottom of the table .
the ultimate goal is to understand whether grb hosts are special galaxies , related to the occurrence of a grb event , or just normal galaxies , but small and metal poor because these are the most common galaxies in the universe . due to the large redshift interval spanned by the sample , @xmath0
, different populations likely coexist .
the one at low redshift is different from that at high redshift because of the intrinsic evolution of galaxies with time , and because of the redshift dependence of the selection effects . from the colors of the sample , we confirm that grb hosts tend to be blue galaxies
. the best - fit sed can constrain the fraction of stars involved in the burst phase , which in our sample is generally less than 10% of the total mass . a reliable estimate of the @xmath20-band absolute magnitudes
is provided . due to observational limits , at high redshift
only bright galaxies are detected .
grb hosts below @xmath165 have @xmath166 .
most galaxies ( 83% ) have stellar masses in the range @xmath167 m@xmath4 to @xmath168 m@xmath4 .
the stellar mass completeness limit of the typical present - day high - redshift galaxy survey is higher than @xmath168 m@xmath4 . for @xmath84 m@xmath4 ,
the grb - host sample is highly incomplete .
for the six short - grb hosts the median stellar mass is @xmath169 m@xmath4 , indicating that the association of short grbs with elliptical more - massive galaxies needs a larger sample to be proven .
we provide a relation between stellar mass and absolute magnitude @xmath20 ( eq .
[ emk_m ] ) , which is based on the finding of a nearly constant stellar mass - to - light ratio in the @xmath20 band , of the order of @xmath170 @xmath171 .
the sfrs derived from emission lines , after dust - extinction , aperture - slit loss , and stellar balmer absorption correction , are the best approximation to the true total sfr .
when emission lines are not measured , we propose an sfr relation which uses the uv luminosity at 2800 .
sfrs in the total sample span three and a half orders of magnitudes ( sfr @xmath172 m@xmath4 yr@xmath7 ) .
two out of six short - grb hosts have high sfrs , more than 10 m@xmath4 yr@xmath7 .
the sfr normalized by the stellar mass ssfr , or its inverse , the growth timescale @xmath121 indicate that grb hosts are active and young systems , which is expected given the generally low mass of the hosts , the high redshifts , and the fact that redshifts are often found trough the detection of emission lines .
berger et al . ( 2003 ) and le floch et al .
( 2006 ) have claimed much higher sfrs , based on radio and mid - ir detections .
the host of grb 980613 looks involved into a merger of several subcomponents ( hjorth et al .
2002 ; djorgovski et al .
our sfr estimate refers to component a that hosted the grb .
the high sfr derived by le floch et al .
( 2006 ) refers to component d , which is more than 30 kpc away from component a. regarding grb 980703 and grb 000418 , le floch et al .
( 2006 ) commented that agn contamination would explain the high radio emission and no spitzer detection .
however , the existence of two grb hosts with agn on a relatively small sample observed in the radio is hard to explain ( michaowski et al .
if the radio is associated with sf , then the nondetection in the mid - ir would imply a strong absorption by the silicates around 10 @xmath26 m ( e. le floch , 2008 , private communication ) .
the blue optical colors for some of the grb hosts can be explained if large regions of these galaxies are totally obscured by dust .
optical light might be dominated by relatively less obscured stars outside star - forming regions , leading to blue colors ( michaowski et al .
2008 ) .
we further investigate the relation between grb hosts and other galaxies , by analyzing the mass - metallicity ( @xmath173 ) relation .
this is the correlation between the stellar mass of galaxies and their metallicity ( lequeux et al .
1979 ; tremonti et al .
2004 ; savaglio et al .
2005 ) . in the @xmath173 plot of figure [ mz ]
, we compare grb hosts with local dwarf galaxies ( lee et al .
2006 ) , and gdds and cfrs galaxies at @xmath153 ( savaglio et al .
the local sample metallicities are estimated using the @xmath130 method . since the @xmath130 or o3n2 metallicities are not measured for gdds galaxies and grb hosts , proper conversions are applied , as explained in [ metallicities ] .
in the left and right panels of figure [ mz ] , we show the lower and the upper branch solution for metallicity in a subsample of 9 grb hosts .
the picture is not very clear , and we do not detect a @xmath173 relation in the grb hosts of any kind , especially because for 9 hosts the metallicity is poorly constrained .
however , we can say that there is no indication that grb hosts have metallicity lower than expected , given their stellar mass ( stanek et al .
grb hosts are not a special class of galaxies . given the redshift and the stellar mass , the lower branch metallicity is favored for the 9 grb hosts with double solution ( right hand side of figure [ mz ] ) ,
as regards the redshift evolution , we compare grb host metallicities , with the metallicities measured in 11 damped lyman-@xmath53 systems through afterglow spectroscopy ( grb - dlas ; savaglio 2006 ) at redshift @xmath35 ( figure [ zz ] ) .
again , the situation is not very clear .
grb - dlas indicate that metallicities of the order of half solar are expected at @xmath141 .
this is actually observed in the subsample of 6 grb hosts whose metallicity ( determined with the @xmath130 or o3n2 calibration ) is better constrained .
however , for the rest of the sample , the lower branch solution suggests no redshift evolution in @xmath0 , an interval of more than 13 gyr .
any conclusion needs more and deeper observations , to break the degeneracy in those objects already observed , or to study new objects .
for instance , the o3n2 calibrator can be detected using nir spectroscopy for the brightest objects at @xmath174 .
the typical grb host is a small star - forming galaxy with likely sub - solar metallicity , but a non - negligible dust extinction ( table [ tsummary ] ) .
it is in some regards similar to a young lmc , observed at redshift @xmath175 , when it was more active than now , with an sfr which is 5 times higher than today .
we still do not know whether hosts associated with short grbs are considerably different from those associated with long grbs .
we have presented a complete study of the largest sample of galaxies hosting grbs , 46 objects , distributed along the redshift interval @xmath0 .
grb hosts can be used as important probes of the cosmic history of galaxy formation and evolution .
most grbs are associated with the death of young massive stars , which are more common in star - forming galaxies .
therefore , grbs are an effective tool to detect star - forming galaxies . as shown by recent studies ( glazebrook et al . 2004 ; juneau et al . 2005 ; borch et al .
2006 ) , the star - formation density in the @xmath141 universe is carried out by small , faint , low - mass star - forming galaxies , similar to the typical grb host . moreover , in the @xmath176 universe , grb hosts observed with spitzer are @xmath177 times fainter than the typical spectroscopically confirmed galaxy in the great observatories origins deep survey ( goods ; yan et al .
2006 ) , suggesting that not all star - forming galaxies at these redshifts are detected in deep surveys ( chary , berger & cowie ( 2007 ; yksel et al .
we consider however , that our view of grb hosts is still partial , as we mainly detect those at @xmath178 .
hosts at higher redshift are harder to observe .
it is possible that high-@xmath23 grb hosts are more massive than those at low redshift , because star formation could be carried by more - massive galaxies in the remote universe .
future deeper multiband observations of high-@xmath23 hosts are mandatory to help solve this issue .
* grb hosts are generally small star - forming galaxies .
the mean stellar mass is similar to the stellar mass of the lmc , @xmath179 m@xmath4 .
about 83% of the sample has stellar mass in the interval @xmath180 m@xmath4 ) .
the median sfr @xmath181 m@xmath4 yr@xmath7 is 5 times higher than in the lmc ; * to estimate sfr , we derived new relations , suitable for grb hosts .
these give the total sfr when h@xmath53 is not detected , but [ oii ] or uv are detected .
our sfrs span an interval of more than three orders of magnitudes , from 0.01 m@xmath4 yr@xmath7 to 36 m@xmath4 yr@xmath7 ; * the dust extinction in the visual band is on average @xmath90 .
the balmer stellar absorption is generally small , but not negligible .
dust extinction , balmer absorption and slit - aperture flux loss are considered when measuring sfr ; * the median star - formation rate per unit stellar mass ssfr is @xmath182 gyr@xmath7 , with a small scatter , such that sfr @xmath183 , a somewhat surprising result .
the median ssfr is about 5 times higher than in the lmc .
a large fraction of grb hosts are the equivalent of local starbursts ; * metallicities derived from emission lines in the host galaxies at @xmath141 are relatively low , likely in the range 1/10 solar to solar ; * metallicities measured from uv absorption lines in the cold medium of grb hosts at @xmath35 ( grb - dlas ) are in a similar range . combining this with the results for @xmath141 grb hosts ,
we see no significant evolution of metallicity in grb hosts in the interval @xmath184 ; * the subsample of 6 short - grb hosts have stellar masses @xmath185 m@xmath186 m@xmath4 and ssfrs in the range @xmath187 gyr@xmath7 .
the suggestion that short - grb hosts are large quiescent galaxies requires a larger sample to be confirmed ; * there is no clear indication that grb host galaxies belong to a special population .
their properties are those expected for normal star - forming galaxies , from the local to the most distant universe .
our investigation will continue , and results will be made public in the ghosts public database . in the final sample ,
we plan to include all grb hosts detected with swift , for a total of a few hundred objects .
this will give a more complete picture on the nature of grb hosts and their relation with all other galaxies .
we thank avishay gal - yam , emeric le floch , lisa kewley , maryam modjaz , john moustakas , daniele pierini , gra@xmath188yna stasiska and christian wolf for stimulating discussions .
we are indebted to the anonymous referees for the many valuable comments .
we also acknowledge the inspiring collaboration with tams budavri , the programmer behind the ghosts database .
gerhardt meurer is acknowledged for constant scientific support . ,
r. g. et al .
2004 , , 127 , 2455 alloin , d. , collin - souffrin , s. , joly , m. , & vigroux , l. 1979 , , 78 , 200 aoki , k. , furusawa , h. , ohta , k. , yamada , t. , & kawai , n. 2007 , arxiv astrophysics e - prints , arxiv : astro - ph/0612367 asplund , m. , grevesse , n. , sauval , a. j. 2005 , asp conf . ser . 336 : cosmic abundances as records of stellar evolution and nucleosynthesis , 25 baldry , i. k. & glazebrook , k. 2003 , , 593 , 258 baldry , i. k. , glazebrook , k. , & driver , s. p. 2008
, , 388 , 945 barth , a. j. , et al .
2003 , , 584 , l47 bauer , a. e. , drory , n. , hill , g. j. , & feulner , g. 2005 , , 621 , l89 berger , e. , cowie , l. l. , kulkarni , s. r. , frail , d. a. , aussel , h. , & barger , a. j. 2003 , , 588 , 99 berger , e. , fox , d. b. , kulkarni , s. r. , frail , d. a. , & djorgovski , s. g. 2007a , , 660 , 504 berger , e. , et al .
2005 , , 438 , 988 berger , e. , et al . 2007b , , 665 , 102 berger , e. , et al .
2007c , , 664 , 1000 bersier , d. , et al .
2006 , , 643 , 284 bloom , j. s. , berger , e. , kulkarni , s. r. , djorgovski , s. g. , & frail , d. a. 2003 , , 125 , 999 bloom , j. s. , djorgovski , s. g. , & kulkarni , s. r. 2001 , , 554 , 678 bloom , j. s. , djorgovski , s. g. , kulkarni , s. r. , & frail , d. a. 1998 , , 507 , l25 bloom , j. s. , kulkarni , s. r. , & djorgovski , s. g. 2002 , , 123 , 1111 borch , a. , et al . 2006 , , 453 , 869 bresolin , f. 2006 , arxiv astrophysics e - prints , arxiv : astro - ph/0608410 bresolin , f. 2007 , , 656 , 186 brinchmann , j. , charlot , s. , white , s. d. m. , tremonti , c. , kauffmann , g. , heckman , t. , & brinkmann , j. 2004 , , 351 , 1151 calzetti , d. 2001 , pasp , 113 , 1449 calzetti , d. , armus , l. , bohlin , r. c. , kinney , a. l. , koornneef , j. , & storchi - bergmann , t. 2000 , , 533 , 682 cardelli , j. a. , clayton , g. c. , & mathis , j. s. 1989 , , 345 , 245 castro cern , j. m. , michaowski , m. j. , hjorth , j. , watson , d. , fynbo , j. p. u. , & gorosabel , j. 2006 , , 653 , l85 castro - tirado , a. j. , et al .
2001 , , 370 , 398 castro - tirado , a. j. , et al . 2005 , , 439 , l15 castro - tirado , a. j. , et al .
2007 , , 475 , 101 chary , r. , becklin , e. e. , & armus , l. 2002 , , 566 , 229 chary , r. , berger , e. , & cowie , l. 2007 , , 671 , 272 cid fernandes , r. , mateus , a. , sods , l. , stasiska , g. , gomes , j. m. 2005 , mnras , 358 , 363 cimatti , a. , et al .
2002 , , 392 , 395 christensen , l. , hjorth , j. , & gorosabel , j. 2004 , , 425 , 913 cobb , b. e. , bailyn , c. d. , van dokkum , p. g. , buxton , m. m. , & bloom , j. s. 2004 , , 608 , l93 cobb , b. e. , bailyn , c. d. , van dokkum , p. g. , & natarajan , p. 2006a , , 645 , l113 cobb , b. e. , bailyn , c. d. , van dokkum , p. g. , & natarajan , p. 2006b , , 651 , l85 conselice , c. j. , et al .
2005 , , 633 , 29 covino , s. , et al .
2006 , , 447 , l5 della valle , m. , et al .
2006 , , 444 , 1050 de ugarte postigo , a. , et al .
2005 , , 443 , 841 djorgovski , s. g. , bloom , j. s. , & kulkarni , s. r. 2003 , , 591 , l13 djorgovski , s. g. , frail , d. a. , kulkarni , s. r. , bloom , j. s. , odewahn , s. c. , & diercks , a. 2001 , , 562 , 654 djorgovski , s. g. , kulkarni , s. r. , bloom , j. s. , goodrich , r. , frail , d. a. , piro , l. , & palazzi , e. 1998 , , 508 , l17 erb , d. k. , steidel , c. c. , shapley , a. e. , pettini , m. , reddy , n. a. , & adelberger , k. l. 2006 , , 647 , 128 fioc , m. , & rocca - volmerange , b. 1997 , a&a , 326 , 950 fioc , m. , & rocca - volmerange , b. 1999 , arxiv : astro - ph/9912179 frail , d. a. , et al .
2002 , , 565 , 829 fruchter , a. , burud , i. , rhoads , j. , & levan , a. 2001 , grb coordinates netw . , 1087 fruchter , a. s. , et al . 1999 , , 519 , l13 fruchter , a. s. , et al .
2006 , , 441 , 463 fynbo , j. p. u. , et al .
2003 , , 406 , l63 fynbo , j. p. u. , et al . 2006 , , 444 , 1047 galama , t. j. , et al .
1998 , , 395 , 670 galama , t. j. , et al .
2003 , , 587 , 135 gal - yam , a. , et al .
2006 , , 444 , 1053 garnavich , p. m. , et al .
2003 , , 582 , 924 gehrels , n. , et al .
2004 , , 611 , 1005 gehrels , n. , et al .
2005 , , 437 , 851 glazebrook , k. , blake , c. , economou , f. , lilly , s. , & colless , m. 1999 , , 306 , 843 glazebrook , k. , et al .
2004 , , 430 , 181 gorosabel , j. , jelnek , m. , de ugarte postigo , a. , guziy , s. , & castro - tirado , a. j. 2005a , nuovo cimento c , 28 , 677 gorosabel , j. , et al . 2005b , , 444 , 711 gorosabel , j. , et al .
2006 , , 450 , 87 hammer , f. , flores , h. , schaerer , d. , dessauges - zavadsky , m. , le floch , e. , & puech , m. 2006 , , 454 , 103 hjorth , j. , et al .
2002 , , 576 , 113 hjorth , j. , et al .
2003 , , 423 , 847 hjorth , j. , et al .
2005 , , 437 , 859 holland , s. , et al .
2000 , grb coordinates netw , 753 izotov , y. i. , stasiska , g. , meynet , g. , guseva , n. g. , & thuan , t. x. 2006 , , 448 , 955 jakobsson , p. , et al .
2005 , , 629 , 45 jansen , r. a. , fabricant , d. , franx , m. , & caldwell , n. 2000 , , 126 , 331 juneau , s. et al .
2005 , , 619 , l135 kauffmann , g. , et al .
2003 , , 346 , 1055 kelly , p. l. , kirshner , r. p. , & pahre , m. 2008 , , 687 , 1201 kennicutt , r. c. , 1998 , , 36 , 189 kennicutt , r. c. , bresolin , f. , & garnett , d. r. 2003 , , 591 , 801 kewley , l. j. , brown , w. r. , geller , m. j. , kenyon , s. j. , & kurtz , m. j. 2007 , , 133 , 882 kewley , l. j. , & dopita , m. a. 2002 , , 142 , 35 kewley , l. j. , dopita , m. a. , sutherland , r. s. , heisler , c. a. , & trevena , j. 2001 , , 556 , 121 kewley , l. j. , & ellison , s. l. 2008 , , 681 , 1183 kewley , l. j. , geller , m. j. , & jansen , r. a. 2004 , , 127 , 2002 klebesadel , r. w. , strong , i. b. , & olson , r. a. 1973 , , 182 , l85 kobulnicky , h. a. , kennicutt , r. c. , & pizagno , j. l. 1999 , , 514 , 544 kobulnicky , h. a. , & kewley l. j. 2004 , apj , 617 , 240 kroupa , p. 2001 , mnras , 322 , 231 kulkarni , s. r. , et al .
1998 , , 393 , 35 kpc yolda , a. , greiner , j. , & perna , r. 2006 , , 457 , 115 kpc yolda , a. , salvato , m. , greiner , j. , pierini , d. , pian , e. , & rau , a. 2007 , , 463 , 893 lamareille , f. , mouhcine , m. , contini , t. , lewis , i. , & maddox , s. 2004 , , 350 , 396 le borgne , d. , & rocca - volmerange , b. 2002 , , 386 , 446 le borgne , d. , rocca - volmerange , b. , prugniel , p. , lanon , a. , fioc , m. , & soubiran , c. 2004 , a&a , 425 , 881 lee , h. , skillman , e. d. , cannon , j. m. , jackson , d. c. , gehrz , r. d. , polomski , e. f. , & woodward , c. e. 2006 , , 647 , 970 le floch , e. , charmandaris , v. , forrest , w. j. , mirabel , i. f. , armus , l. , & devost , d. 2006 , , 642 , 636 le floch , e. , et al .
2002 , , 581 , l81 le floch , e. , et al .
2003 , , 400 , 499 lequeux , j. , peimbert , m. , rayo , j. f. , serrano , a. , & torres - peimbert , s. 1979 , , 80 , 155 lilly , s. j. , carollo , c. m. , & stockton , a. n. 2003 , apj , 597 , 730 madau , p. , pozzetti , l. , & dickinson , m. 1998 , , 498 , 106 mangano , v. , et al .
2007 , , 470 , 105 masetti , n. , et al .
2005 , , 438 , 841 meurer , g. r. , heckman , t. m. , & calzetti , d. 1999 , , 521 , 64 mccall , m. l. , rybski , p. m. , & shields , g. a. 1985 , apjs , 57 , 1 metzger , m. r. , djorgovski , s. g. , kulkarni , s. r. , steidel , c. c. , adelberger , k. l. , frail , d. a. , costa , e. , & frontera , f. 1997 , , 387 , 878 michaowski , m. j. , hjorth , j. , castro cern , j. m. , & watson , d. 2008 , , 672 , 817 mirabal , n. , halpern , j. p. , & obrien , p. t. 2007 , , 661 , l127 mirabal , n. , et al .
2002 , , 578 , 818 modjaz , m. , et al .
2006 , , 645 , l21 modjaz , m. , et al .
2008 , , 135 , 1136 mller , p. , et al .
2002 , , 396 , l21 moustakas , j. , kennicutt , r. c. , jr . , & tremonti , c. a. 2006 , , 642 , 775 nagao , t. , maiolino , r. , & marconi , a. 2006 , , 459 , 85 noeske , k. g. , et al .
2007 , , 660 , l47 nuza , s. e. , tissera , p. b. , pellizza , l. j. , lambas , d. g. , scannapieco , c. , & de rossi , m. e. 2007 , , 375 , 665 osterbrock , d. e. 1989 , astrophysics of gaseous nebulae and active galactic nuclei ( mill valley , ca : univ .
science books ) osterbrock , d. e. , & ferland , g. j. 2006 , astrophysics of gaseous nebulae and active galactic nuclei ( 2nd . ed .
: sausalito , ca : univ .
science books ) ovaldsen , j .- e . , et al .
2007 , , 662 , 294 pagel , b. e. j. , edmunds , m. g. , blackwell , d. e. , chun , m. s. , & smith , g. 1979 , mnras , 189 , 95 pellizza , l. j. , et al .
2006 , , 459 , l5 perley , d. a. , et al .
2008 , arxiv astrophysics e - prints , arxiv : astro - ph/0703538 pettini , m. , & pagel , b. e. j. 2004 , , 348 , l59 prez - gonzlez , p. g. , gil de paz , a. , zamorano , j. , gallego , j. , alonso - herrero , a. , & aragn - salamanca , a. 2003 , , 338 , 525 pian , e. , et al .
2006 , , 442 , 1011 pilyugin , l. s. 2000 , , 362 , 325 piro , l. , et al .
2002 , , 577 , 680 price , p. a. , et al . 2002a , , 571 , l121 price , p. a. , et al .
2002b , , 573 , 85 price , p. a. , et al .
2003 , , 589 , 838 priddey , r. s. , tanvir , n. r. , levan , a. j. , fruchter , a. s. , kouveliotou , c. , smith , i. a. , & wijers , r. a. m. j. 2006 , , 369 , 1189 prieto , j. l. , stanek , k. z. , & beacom , j. f. 2008 , , 673 , 999 prochaska , j. x. , et al .
2004 , , 611 , 200 prochaska , j. x. , et al .
2006 , , 642 , 989 rau , a. , salvato , m. , & greiner , j. 2005 , , 444 , 425 rau , a. , et al .
2004 , , 427 , 815 reddy , n. a. , steidel , c. c. , erb , d. k. , shapley , a. e. , & pettini , m. 2006 , , 653 , 1004 rol , e. , et al .
2007 , , 669 , 1098 salpeter , e. e. 1955 , , 121 , 161 savaglio , s. , 2006 , new j. phys . , 8
, 195 savaglio , s. , et al .
2005 , , 635 , 260 schlegel , d. j. , finkbeiner , d. p. , & davis , m. 1998 , , 500 , 525 shapley , a. e. , steidel , c. c. , erb , d. k. , reddy , n. a. , adelberger , k. l. , pettini , m. , barmby , p. , & huang , j. 2005 , , 626 , 698 soderberg , a. m. , et al .
2004 , , 606 , 994 soderberg , a. m. , et al . 2006a , , 636 , 391 soderberg , a. m. , et al .
2006b , , 650 , 261 soderberg , a. m. , et al .
2007 , , 661 , 982 sollerman , j. , stlin , g. , fynbo , j. p. u. , hjorth , j. , fruchter , a. , & pedersen , k. 2005 , new astron . , 11 , 103 sollerman , j. , et al .
2006 , , 454 , 503 spergel , d. n. , et al .
2003 , , 148 , 175 stanek , k. z. , et al .
2003 , , 591 , l17 stanek , k. z. , et al .
2006 , acta astron . , 56
, 333 stasiska , g. 2005 , , 434 , 507 stasiska , g. 2006 , , 454 , l127 stasiska , g. , cid fernandes , r. , mateus , a. , sodr , l. , & asari , n. v. 2006 , , 371 , 972 tanvir , n. r. , et al .
2004 , , 352 , 1073 thne , c. c. , greiner , j. , savaglio , s. , & jehin , e. 2007 , , 671 , 628 thne , c. c. , et al .
2008 , , 676 , 1151 totani , t. , kawai , n. , kosugi , g. , aoki , k. , yamada , t. , iye , m. , ohta , k. , & hattori , t. 2006 , , 58 , 485 tremonti ,
c. a. , et al .
2004 , , 613 , 898 wainwright , c. , berger , e. , & penprase , b. e. 2007 , , 657 , 367 wiersema , k. , et al .
2007 , , 464 , 529 wiersema , k. , et al .
2008 , , 481 , 319 wolf , c. , & podsiadlowski , p. 2007
, , 375 , 1049 woosley , s. e. 1993 , , 405 , 273 worthey g. , 1994 , apjs , 95 , 107 yan , h. , dickinson , m. , giavalisco , m. , stern , d. , eisenhardt , p. r. m. , & ferguson , h. c. 2006 , , 651 , 24 yksel , h. , kistler , m. d. , beacom , j. f. , & hopkins , a. m. 2008 , , 683 , l5 lcccccc grb 970228 & 0.695 & long & [ 1,2,3 ] & late & disk & 0.234 + grb 970508 & 0.835 & long & [ 3,4,5 ] & early & spheroid & 0.026 + grb 970828 & 0.960 & long & [ 5,6,7 ] & . . .
& asymmetric , merger & 0.036 + grb 971214 & 3.420 & long & [ 2,8,5 ] & late & asymmetric & 0.016 + grb 980425 & 0.009 & long & [ 9,10,5,11 ] & . . .
& 0.065 + grb 980613 & 1.097 & long & [ 2,12 ] & merger & asymmetric , merger & 0.087 + grb 980703 & 0.966 & long & [ 3,13,5 ] & late & spheroid & 0.061 + grb
990123 & 1.600 & long & [ 3,5 ] & late & disk , merger & 0.016 + grb 990506 & 1.310 & long & [ 14,5 ] & early & spheroid & 0.068 + grb 990705 & 0.842 & long & [ 9,15,16,5 ] & late & disk & 0.120 + grb 990712 & 0.433 & long & [ 3,17 ] & late & disk , merger & 0.033 + grb 991208 & 0.706 & long & [ 3,18 ] & early & spheroid , merger ? & 0.016 + grb 000210 & 0.846 & long & [ 3,19 ] & . . .
& 0.019 + grb 000418 & 1.118 & long & [ 3,14 ] & early & spheroid & 0.032 + grb 000911 & 1.058 & long & [ 20,21 ] & . . .
& 0.107 + grb 000926 & 2.036 & long & [ 3 ] & late & merger & 0.023 + grb 010222 & 1.480 & long & [ 22,23,24,5,25 ] & merger & spheroid & 0.023 + grb 010921 & 0.451 & long & [ 3,25,26 ] & late & disk & 0.148 + grb 011121 & 0.362 & long & [ 27,28 ] & late & disk & 0.370 + grb 011211 & 2.141 & long & [ 29 ] & late & merger & 0.045 + grb 020405 & 0.691 & long & [ 30,25 ] & merger & asymmetric , merger & 0.055 + grb 020813 & 1.255 & long & [ 31,7,25 ] & early & . . .
& 0.111 + grb 020819b & 0.410 & long & [ 32 ] & . . .
& 0.070 + xrf 020903 & 0.251 & long & [ 33,34,25 ] & merger & merger & 0.033 + grb
021004 & 2.330 & long & [ 35,36,37 ] & late & spheroid & 0.060 + grb 021211 & 1.006 & long & [ 25 ] & late & spheroid & 0.028 + grb 030328 & 1.520 & long & [ 38 ] & . .
& 0.047 + grb 030329 & 0.168 & long & [ 39,40 ] & late & . .
& 0.025 + xrf 030528 & 0.782 & long & [ 41,42 ] & . .
& 0.620 + grb 031203 & 0.105 & long & [ 43,44 ] & . . .
& 1.040 + grb 040924 & 0.859 & long & [ 25,45 ] & . .
& 0.058 + grb 041006 & 0.712 & long & [ 25,46 ] & . .
& asymmetric & 0.070 + grb 050223 & 0.584 & long & [ 47 ] & . .
& 0.090 + xrf 050416 & 0.653 & short & [ 48 ] & . .
& 0.059 + grb 050509b & 0.225 & short & [ 49 ] & . .
& 0.019 + grb 050709 & 0.161 & short & [ 50,51 ] & . .
& 0.012 + grb 050724 & 0.257 & short & [ 52,53 ] & . .
& 0.613 + grb 050826 & 0.296 & long & [ 54,55 ] & . .
& 0.590 + grb 050904 & 6.295 & long & [ 56,57,58 ] & . . .
& 0.060 + grb 051022 & 0.807 & long & [ 59,60 ] & . .
& 0.040 + grb 051221 & 0.546 & short & [ 61 ] & . . .
& 0.069 + xrf 060218 & 0.034 & long & [ 62,63,64,65 ] & . .
& 0.140 + grb 060505 & 0.089 & long ? & [ 66 ] & . .
& 0.021 + grb 060614 & 0.125 & long & [ 67,68,69,70 ] & . . .
& 0.070 + grb 061006 & 0.438 & short & [ 71 ] & . .
& 0.320 + grb 061126 & 1.159 & long & [ 72 ] & . . .
& 0.182 + lccccccccc 970228 & 0.695 & . .
& @xmath189 & @xmath190 & @xmath191 & @xmath192 & . .
& @xmath193 & @xmath194 + 970508 & 0.835 & . .
& @xmath195 & @xmath196 & @xmath197 & @xmath198 & . .
& @xmath199 + 970828 & 0.960 & . .
& @xmath200 & . .
& @xmath201 + 971214 & 3.420 & . .
& @xmath202 & @xmath203 & . .
& @xmath204 + 980425 & 0.009 & . .
& @xmath205 & @xmath206 & @xmath207 & @xmath208 & . .
+ 980613 & 1.097 & . .
& @xmath209 & @xmath210 & @xmath211 & @xmath212 & . .
& @xmath213 + 980703 & 0.966 & . .
& @xmath214 & @xmath215 & @xmath216 & @xmath217 & @xmath218 & @xmath219 & @xmath220 + 990123 & 1.600 & @xmath221 & @xmath222 & @xmath223 & @xmath224 & @xmath225 & . .
& @xmath226 + 990506 & 1.310 & . .
& @xmath227 & . .
& @xmath228 + 990705 & 0.842 & . .
& @xmath229 & @xmath230 & . .
+ 990712 & 0.433 & @xmath231 & @xmath232 & @xmath233 & @xmath234 & @xmath235 & @xmath236 & @xmath237 & @xmath238 + 991208 & 0.706 & . .
& @xmath239 & @xmath240 & @xmath241 & @xmath242 & . .
& @xmath243 + 000210 & 0.846 & @xmath244 & @xmath245 & @xmath246 & @xmath247 & @xmath248 & @xmath249 & @xmath250 & @xmath251 + 000418 & 1.118 & . .
& @xmath252 & @xmath253 & @xmath254 & @xmath255 & @xmath256 & . .
& @xmath257 + 000911 & 1.058 & . .
& @xmath258 & . .
& @xmath259 & . .
& @xmath260 & . .
+ 000926 & 2.036 & . .
& @xmath261 & @xmath262 & @xmath263 & @xmath264 & @xmath265 & . .
+ 010222 & 1.480 & . .
& @xmath266 & @xmath267 & @xmath268 & @xmath269 & . .
& @xmath270 + 010921 & 0.451 & . .
& @xmath271 & @xmath272 & @xmath273 & @xmath274 & @xmath275 & @xmath276 & @xmath277 + 011121 & 0.362 & . .
& @xmath278 & @xmath279 & . .
& @xmath280 & @xmath281 & . .
+ 011211 & 2.141 & . . .
& @xmath282 & . .
& @xmath283 & . .
+ 020405 & 0.691 & . .
& @xmath284 & . .
& @xmath285 & . .
+ 020813 & 1.255 & . .
& @xmath286 & . .
& @xmath287 & @xmath288 & . .
+ 020819b & 0.410 & . .
& @xmath289 & . .
& @xmath290 & . .
& @xmath291 + 020903 & 0.251 & . .
& @xmath292 & @xmath293 & @xmath294 & @xmath295 & . .
+ 021004 & 2.327 & . .
& @xmath296 & . .
& @xmath297 & @xmath298 & . .
& @xmath299 & . .
+ 021211 & 1.006 & . .
& @xmath300 & . .
& @xmath301 & . .
+ 030328 & 1.520 & @xmath302 & @xmath303 & @xmath304 & @xmath305 & @xmath306 & . .
& @xmath307 & @xmath308 + 030329 & 0.168 & @xmath309 & @xmath310 & @xmath311 & @xmath312 & @xmath313 & @xmath314 & @xmath315 & . .
+ 030528 & 0.782 & . .
& @xmath316 & @xmath317 & @xmath318 & @xmath319 & . .
& @xmath320 + 031203 & 0.105 & . .
& @xmath321 & . .
& @xmath322 & @xmath323 + 040924 & 0.859 & . .
& @xmath324 & . .
+ 041006 & 0.712 & . .
& @xmath325 & . .
+ 050223 & 0.584 & . .
& @xmath326 & . .
& @xmath327 + 050416 & 0.653 & . .
& @xmath328 & @xmath329 & . .
+ 050509b & 0.225 & @xmath330 & . .
& @xmath331 & @xmath332 & @xmath333 & @xmath334 & @xmath335 + 050709 & 0.161 & . .
& @xmath336 & @xmath337 & @xmath338 & @xmath339 & . .
+ 050724 & 0.257 & . .
& @xmath340 & @xmath341 & . .
& @xmath342 & @xmath343 & @xmath344 & @xmath345 + 050826 & 0.296 & . .
& @xmath346 & @xmath347 & @xmath348 & @xmath349 & . .
+ 050904 & 6.295 & . .
& @xmath350 & . .
+ 051022 & 0.807 & . .
& @xmath351 & @xmath352 & @xmath353 & @xmath354 & . .
& @xmath355 & @xmath356 + 051221 & 0.546 & . .
& @xmath357 & @xmath358 & . .
+ 060218 & 0.034 & @xmath359 & . .
& @xmath360 & @xmath361 & . .
+ 060505 & 0.089 & @xmath362 & @xmath363 & @xmath364 & @xmath365 & @xmath366 & . .
& @xmath367 + 060614 & 0.125 & @xmath368 & @xmath369 & @xmath370 & @xmath371 & @xmath372 & . .
+ 061006 & 0.438 & . .
& @xmath246 & @xmath373 & . .
+ 061126 & 1.159 & . .
& @xmath374 & . .
+ lcccccccccccc 970228 & 0.695 & @xmath375 & @xmath376 & . . .
& @xmath377 & . .
& @xmath378 & . .
& 1 + 970508 & 0.835 & @xmath379 & @xmath380 & . .
& 1 + 970828 & 0.958 & @xmath381 & @xmath382 & . .
& 1 + 980425 & 0.0085 & @xmath383 & @xmath384 & . .
& @xmath385 & @xmath386 & @xmath387 & @xmath388 & @xmath384 & @xmath389 & @xmath390 & 4.6 + 980613 & 1.097 & @xmath391 & @xmath382 & . .
& 1.2 + 980703 & 0.966 & @xmath392 & . .
& @xmath393 & @xmath394 & @xmath395 & . .
& 1 + 990506 & 1.31 & @xmath396 & . .
& 1 + 990705 & 0.8424 & @xmath397 & . .
& 1 + 990712 & 0.434 & @xmath398 & @xmath399 & @xmath400 & @xmath401 & @xmath402 & @xmath403 & @xmath404 & @xmath405 & . .
& 1.3 + 991208 & 0.706 & @xmath406 & . .
& @xmath407 & @xmath408 & @xmath409 & . .
& 1 + 000210 & 0.846 & @xmath410 & . .
& 1 + 000418 & 1.118 & @xmath411 & @xmath412 & @xmath413 & . .
& 1.9 + 000911 & 1.0585 & @xmath414 & . .
& 1 + 010921 & 0.451 & @xmath415 & @xmath416 & . .
& @xmath417 & @xmath418 & @xmath419 & @xmath420 & . .
& 1 + 011121 & 0.362 & @xmath421 & @xmath422 & . .
& @xmath423 & . .
& @xmath424 & @xmath425 & . .
& 1 + 020405 & 0.691 & @xmath426 & @xmath427 & @xmath428 & @xmath429 & @xmath430 & @xmath431 & . .
& 1.3 + 020813 & 1.255 & @xmath432 & . .
& 1 + 020903 & 0.251 & @xmath433 & @xmath434 & @xmath435 & @xmath436 & @xmath437 & @xmath438 & @xmath439 & @xmath440 & @xmath441 & @xmath442 & 2.7 + 030329 & 0.168 & @xmath443 & @xmath444 & @xmath445 & @xmath446 & @xmath447 & @xmath448 & @xmath449 & @xmath450 & . .
& @xmath451 & 1 + 030528 & 0.782 & @xmath452 & @xmath125 & @xmath453 & @xmath454 & @xmath455 & @xmath456 & . .
& 3.1 + 031203 & 0.1055 & @xmath457 & @xmath458 & @xmath459 & @xmath460 & @xmath461 & @xmath462 & @xmath463 & @xmath464 & @xmath465 & @xmath466 & 2 + 040924 & 0.858 & @xmath467 & @xmath468 & . .
& @xmath469 & @xmath470 & . .
& 2 + 041006 & 0.712 & @xmath471 & . .
& 1 + 050223 & 0.584 & @xmath472 & . .
& @xmath473 & . .
& @xmath474 & . .
& 1.2 + 050416 & 0.6528 & @xmath475 & . .
& @xmath476 & @xmath477 & @xmath478 & @xmath479 & . .
& 1 + 050709 & 0.1606 & . .
& @xmath480 & . .
& @xmath481 & . .
& 1.15 + 050826 & 0.296 & @xmath482 & . .
& 2.6 + 051022 & 0.8070 & @xmath483 & . .
& @xmath484 & . .
& @xmath485 & . .
& 1 + 051221 & 0.5459 & @xmath486 & . .
& @xmath487 & @xmath488 & @xmath489 & @xmath490 & . .
& 1.4 + 060218 & 0.03345 & @xmath491 & @xmath492 & @xmath493 & @xmath494 & @xmath495 & @xmath496 & @xmath497 & @xmath498 & @xmath499 & @xmath499 & 1 + 060505 & 0.0889 & @xmath500 & . .
& @xmath501 & @xmath502 & @xmath503 & @xmath504 & @xmath505 & @xmath506 & @xmath507 & @xmath508 & 5 + 060614 & 0.1254 & . .
& @xmath509 & . .
& 1 + 061126 & 1.1588 & @xmath510 & . .
& 1.3 + lcccccccc 970228 & long & 0.695 & @xmath511 & @xmath512 & . .
& 0.53 & 0.082 & . .
+ 970508 & long & 0.835 & @xmath513 & @xmath514 & . .
& 1.14 & 0.534 & . .
+ 970828 & long & 0.960 & @xmath515 & @xmath516 & . .
& 0.87 & -0.246 & . .
+ 971214 & long & 3.420 & @xmath517 & @xmath518 & . .
& 11.40 & 0.467 & . .
+ 980425 & long & 0.009 & @xmath519 & @xmath520 & 1.73 & 0.21 & -0.883 & 8.16 + 980613 & long & 1.097 & @xmath521 & @xmath522 & . .
& 4.70 & 1.184 & . .
+ 980703 & long & 0.966 & @xmath523 & @xmath524 & . .
& 16.57 & 0.885 & 7.6/8.14 + 990123 & long & 1.600 & @xmath525 & @xmath526 & . .
& 5.72 & 0.340 & . .
+ 990506 & long & 1.310 & @xmath527 & @xmath528 & . .
& 2.50 & -0.081 & . .
+ 990705 & long & 0.842 & @xmath529 & @xmath530 & . .
& 6.96 & -0.357 & . .
+ 990712 & long & 0.433 & @xmath531 & @xmath532 & @xmath533 & 2.39 & 0.093 & @xmath534 + 991208 & long & 0.706 & @xmath535 & @xmath536 & . .
& 4.52 & 1.121 & @xmath537/8.73 + 000210 & long & 0.846 & @xmath538 & @xmath539 & . .
& 2.28 & 0.049 & . .
+ 000418 & long & 1.118 & @xmath540 & @xmath541 & . .
& 10.35 & 0.757 & . .
+ 000911 & long & 1.058 & @xmath542 & @xmath543 & . .
& 1.57 & -0.124 & . .
+ 000926 & long & 2.036 & @xmath544 & @xmath545 & . .
& 2.28 & -0.165 & . .
+ 010222 & long & 1.480 & @xmath546 & @xmath547 & . .
& 0.34 & -0.290 & . .
+ 010921 & long & 0.451 & @xmath548 & @xmath549 & @xmath550 & 2.50 & -0.289 & 8.0/8.15 + 011121 & long & 0.362 & @xmath551 & @xmath552 & 0.38 & 2.24 & -0.464 & 7.50/8.64 + 011211 & long & 2.141 & @xmath553 & @xmath554 & . .
& 4.90 & -0.084 & . .
+ 020405 & long & 0.691 & @xmath555 & @xmath556 & . .
& 3.74 & -0.174 & 7.78/8.44 + 020813 & long & 1.255 & @xmath557 & @xmath558 & . .
& 6.76 & 1.167 & . .
+ 020819b & long & 0.410 & @xmath559 & @xmath560 & . .
& 6.86 & -0.664 & . .
+ 020903 & long & 0.251 & @xmath561 & @xmath562 & 0.59 & 2.65 & 0.555 & 8.22 + 021004 & long & 2.327 & @xmath563 & @xmath564 & . .
& 3.12 & -0.705 & . .
+ 021211 & long & 1.006 & @xmath565 & @xmath566 & . .
& 3.01 & -0.841 & . .
+ 030328 & long & 1.520 & @xmath567 & @xmath568 & . .
& 3.20 & 0.680 & . .
+ 030329 & long & 0.168 & @xmath569 & @xmath570 & @xmath571 & 0.11 & 0.304 & 7.97 + 030528 & long & 0.782 & @xmath572 & @xmath573 & . .
& 15.07 & 1.355 & @xmath534 + 031203 & long & 0.105 & @xmath574 & @xmath575 & @xmath576 & 12.68 & 1.287 & 8.02 + 040924 & long & 0.859 & @xmath577 & @xmath578 & . .
& 1.88 & 0.071 & . .
+ 041006 & long & 0.712 & @xmath579 & @xmath580 & . .
& 0.34 & -0.131 & . .
+ 050223 & long & 0.584 & @xmath581 & @xmath582 & . .
& 1.44 & -0.568 & 7.5/8.66 + 050416 & short & 0.653 & @xmath583 & @xmath584 & . .
& 2.32 & -0.476 & 7.97/8.44 + 050509b & short & 0.225 & @xmath585 & @xmath586 & . .
& 16.87 & -0.853 & . .
+ 050709 & short & 0.161 & @xmath587 & @xmath588 & @xmath589 & 0.14 & -0.512 & . .
+ 050724 & short & 0.257 & @xmath590 & @xmath591 & . .
& 18.76 & -0.367 & . .
+ 050826 & long & 0.296 & @xmath592 & @xmath593 & . .
& 9.13 & 0.172 & . .
+ 050904 & long & 6.295 & @xmath594 & @xmath595 & . .
+ 051022 & long & 0.807 & @xmath596 & @xmath597 & . .
& 36.46 & 0.142 & 7.5/8.65 + 051221 & short & 0.546 & @xmath598 & @xmath599 & . .
& 2.61 & 0.804 & 7.6/8.59 + 060218 & long & 0.034 & @xmath600 & @xmath601 & @xmath602 & 0.05 & -0.061 & 8.13 + 060505 & long ? & 0.089 & @xmath603 & @xmath604 & @xmath605 & 0.43 & -0.777 & 8.44 + 060614 & long & 0.125 & @xmath606 & @xmath607 & . .
& 0.01 & -0.863 & . .
+ 061006 & short & 0.438 & @xmath608 & @xmath609 & . .
& 0.17 & -2.189 & . .
+ 061126 & long & 1.159 & @xmath610 & @xmath611 & . .
& 2.38 & -0.934 & . .
+ [ 2pt ] & & & & & & & & + median & long & 0.75 & @xmath612 & 9.3 & 0.44 & 2.5 & -0.10 & 7.9/8.3 + | we present the most extensive and complete study of the properties for the largest sample ( 46 objects ) of gamma - ray burst ( grb ) host galaxies .
the redshift interval and the mean redshift of the sample are @xmath0 and @xmath1 ( look - back time : 7.2 gyr ) , respectively ; 89% of the hosts are at @xmath2 .
optical - near - infrared ( nir ) photometry and spectroscopy are used to derive stellar masses , star formation rates ( sfrs ) , dust extinctions and metallicities .
the average stellar mass is @xmath3 m@xmath4 , with a 1@xmath5 dispersion of 0.8 dex .
the average metallicity for a subsample of 17 hosts is about 1/6 solar and the dust extinction in the visual band ( for a subsample of 10 hosts ) is @xmath6 .
we obtain new relations to derive sfr from [ oii ] or uv fluxes , when balmer emission lines are not available .
sfrs , corrected for dust extinction , aperture - slit loss and stellar balmer absorption are in the range 0.01 - 36 m@xmath4 yr@xmath7 .
the median sfr per unit stellar mass ( specific sfr ) is 0.8 gyr@xmath7 .
equivalently the inverse quantity , the median formation timescale is 1.3 gyr .
most grbs are associated with the death of young massive stars , more common in star - forming galaxies
. therefore grbs are an effective tool to detect star - forming galaxies in the universe .
star - forming galaxies at @xmath8 are a faint and low - mass population , hard to detect by conventional optical - nir surveys , unless a grb event occurs .
there is no compelling evidence that grb hosts are peculiar galaxies .
more data on the subclass of short grb are necessary to establish the nature of their hosts . |
it is well known that wzw models are some of the few known exact string theory backgrounds .
they are important elements of the near - horizon geometry of various string theory configurations .
for example the background of @xmath0 ns5 branes and @xmath1 fundamental strings has a near horizon geometry given by the @xmath2 wzw model . on the other hand
the near - horizon geometry of a circular distribution of ns5 branes is t - dual to an orbifold of the @xmath3 gauged wzw model .
in addition wzw models and their gauged variations have become increasingly important in studies of the ads / cft correspondence and branes - world scenarios .
d - branes are an essential ingredient of string theory since they play crucial role in various dualities and holography .
one way to study the properties of d - branes in a given background is to construct the boundary state of the d - branes which describe the couplings of closed string modes on their world - volume .
this is in general a difficult question for a general background .
although d - branes in wzw models on compact group manifolds like su(2 ) have been studied extensively , little progress has been achieved towards non - compact cases like @xmath4 .
in fact even the spectrum of these theories is not fully understood .
nevertheless we can get useful information for such d - branes by studying just the semiclassical limit of their boundary states , in other words their low energy effective action .
this has been a fruitful approach for the @xmath4 case @xcite as well as for the cases of compact group manifolds @xcite . in this note
we will study the semiclassical features of d - branes in the gauged @xmath5 wzw model along the lines of a similar study for @xmath6 @xcite .
since we are not in position to give the full boundary state for all d - branes in @xmath5 , we will restrict ourselves to their geometrical features .
finally , for one set of branes , we use a small fluctuation analysis of the low energy effective action to get the semiclassical limit of the open string spectra .
consequently we compare our results with expectations from a cft analysis @xcite .
the @xmath4 wzw model is described by a world - sheet sigma model with target space : @xmath7 and a kalb - ramond field strength given by the volume element of the spacetime .
the constant k is the level of the affine lie algebra that can take real values with @xmath8 , and we used units @xmath9 .
one can construct the @xmath5 coset from the @xmath4 wzw model by gauging a u(1 ) subgroup of @xmath4 @xcite . if this subgroup is non compact one ends with a manifold with lorentzian signature , the 2d black hole . on the contrary
if we choose a compact u(1 ) then we get an euclidean manifold .
there are two possible anomaly free u(1 ) gauge symmetries : the axial , @xmath10 and the vector , @xmath11 , where @xmath12 group elements of @xmath4 and @xmath13 . for a compact u(1 ) the manifold we get by gauging the axial symmetry has the geometry of a cigar : @xmath14 the axial gauging amounts intuitively into gauging the symmetry which corresponds to time translations in @xmath4 .
this will be important later in understanding the geometry of branes in the cigar geometry .
notice that unlike @xmath15 , in @xmath4 we have closed time - like curves .
therefore the u(1 ) associated with time translations is a compact subgroup of @xmath4 . by gauging the vector compact u(1 ) symmetry
we get the geometry of a trumpet : @xmath16 this coset amounts to gauging the u(1 ) symmetry of shifting @xmath17 .
notice that unlike the cigar manifold , that of the trumpet is singular due to the fact that the vector gauging action has fixed points .
the two backgrounds are actually t - dual to each other . to be exact the cigar is t - dual to the trumpet where @xmath18 . finally we would like to summarize the operator spectrum of the theory @xcite since it will become useful in our small fluctuation analysis .
the operators in the axially gauged theory belong to discrete series with @xmath19 and continuous series with @xmath20 , @xmath21 .
the operators are charged under the @xmath4 currents @xmath22 with charges : @xmath23 and @xmath24 . the classical angular momentum of an operator is @xmath25 and the winding @xmath26 .
the conformal weights of the operators are : @xmath27 in addition from the harmonic analysis in @xmath4 we conclude that discrete representations should satisfy : @xmath28 .
we see that we can not have discrete representations with vanishing winding . for the t - dual vector
gauging we need to take @xmath29 which results in @xmath30 in ( [ cw ] ) .
the discrete representations constrain in this case is @xmath31 , where @xmath25 is the angular momentum in the trumpet and @xmath26 the winding . in this case discrete series states with non - zero momentum and zero winding exist .
the continuous representations correspond semiclassically to wave like solutions while the discrete ones to bound states .
so , in the semiclassical limit @xmath32 , where the winding states decouple , we expect bound states for the trumpet and not for the cigar .
these expectations from cft are nicely confirmed by the computation of the low energy spectrum of the two geometries @xcite .
the open string spectra of the d - branes satisfy analogous constrains .
the geometry of d - branes in @xmath33 has been studied in @xcite ( see also i.e. @xcite for related work ) where it was found that solutions of the dbi action correspond to regular and twined conjugacy classes of @xmath4 .
it was found that d1 branes have @xmath34 world volumes and d2 branes come in two types with @xmath35 and @xmath36 world volumes @xcite .
the first type of d2 branes has a d - instanton density in its world - volume while the second one was found to be unphysical due to a supercritical electric field . as explained in @xcite the d - branes in coset models @xmath37 have world - volumes localized on the projection to the coset , of the product of a conjugacy class of @xmath38 with a conjugacy class of @xmath39 . in geometric words we expect the d - branes we will find on the cigar to be projections on a constant @xmath40 plane of the d - branes in @xmath4 and those of the trumpet projections on a constant @xmath17 plane .
we will verify these expectations by solving the dbi for the cigar and trumpet .
we will assume from now on that the level @xmath41 is an integer .
this needs not to be the general case but in most of the interesting ( supersymmetric ) backgrounds the @xmath42 cft appears with a corresponding @xmath43 with the levels @xmath41 and @xmath44 differing by an integer @xcite . since @xmath44 must be an integer then @xmath41 should be one too . for the d1 branes in the cigar and with an embedding @xmath45 we need to minimize the `` energy '' of the system @xmath46 , where @xmath47 is derivative with respect to @xmath17 .
the same method is employed for the d1 branes in the trumpet . for the d2 branes one
has to solve the equations of motion for the gauge field @xmath48 .
the analysis is straight forward as shown in tables 1 and 2 . in tables 1 and 2
we give the embedding equations for the various branes in trumpet and cigar respectively , as well as the @xmath49 branes from which they descent .
the dimensionality of each brane is evident by its name i.e. @xmath50 is an 1-dimensional brane in the trumpet , descending from the @xmath36 brane in @xmath33 .
@xmath51 @xmath52 where they descent from .
, width=377 ] at this point a few comments are in order .
first looking at the table of the cigar and the trumpet we notice that the branes which descent from the @xmath34 and @xmath36 branes have two classical moduli .
unlike these branes , the dbi analysis analogous to @xcite for d2 branes in the cigar , implies that they carry a d0 brane charge which leads to the quantization of the modulus @xmath53 .
also since they have trivial topology they do not admit a wilson line . by t - duality these branes are related to d1 branes in the trumpet which hit the infinite radius circle at @xmath54 . in principal one expects no quantization of their incidence angle given by the modulus @xmath53 and in addition that they carry an extra modulus , the angle @xmath55 where they hit the circle @xmath54 .
t - duality with the cigar suggests that these d1 branes can hit only special points on the @xmath54 circle of the trumpet and it is suggested also ( see @xcite ) that they correspond to bound states of d1 branes orthogonal to the circle with a number of @xmath56 branes .
we should also emphasize that the trumpet has a strong coupling singularity at @xmath54 and therefore it is plausible that our analysis might be valid only for branes which extend far enough from the singularity .
nevertheless the considerable success of our semiclassical analysis and the similar results of @xcite for the spectrum of the theory suggest that there should be a grain of truth in our results .
notice that we have not been able to identify the brane of @xmath49 , if any , where the @xmath57 branes on the cigar and their t - dual @xmath56 descent from .
one might think that the d0 brane of the cigar comes from the point brane of @xmath4 but there seems to be a contradiction when one tries to construct the @xmath56 by descending from this brane to the trumpet .
the point in @xmath4 will still be a point in cigar or trumpet .
these shortcomings of our semiclassical analysis should be eliminated once we are able to have the complete boundary state for these branes .
finally notice that since the cigar is actually t - dual to an orbifold of the trumpet , one needs to rescale the string coupling by @xmath58 when comparing quantities under t - duality .
the boundary states for branes in @xmath4 has been a long and rather subtle problem . the non - compactness of the group results in infinities for most physically relevant quantities which must be dealt with caution . on top of
that only recently @xcite was a significant aspect of the spectrum , that of spectral flow , understood .
nevertheless one can work in the euclidean analog of @xmath15 and try to construct branes in the @xmath59 hyperbolic space .
this method was employed in @xcite and the boundary state of a brane with @xmath34 world - volume was constructed among other branes of the background .
it was claimed that due to the time independence of this particular set of branes on can descent trivially to the corresponding branes in the coset @xmath60 which describes the cigar geometry .
these branes couple only to closed string momentum modes as expected by their geometry .
this in turn implies , due to the discrete representation constrains , that only closed strings in the continuous representations couple to the brane .
we will work along the lines of @xcite to study the semiclassical approximation to the open string spectrum of the branes .
it is easy to understand by the shape of the @xmath61 branes ( see also @xcite ) that open string winding states should appear in the spectrum and not momentum ones . since open string states satisfy the same constrains for the allowed quantum numbers as closed ones , continuous states as well as discrete ones with winding @xmath62 are allowed . in the semiclassical limit
only the winding zero state survives since all other ones become heavy and decouple .
then computing the open string spectrum of these branes using the dbi in the cigar geometry we expect to see only continuous representations since the discrete ones are forbidden for @xmath63 .
the purpose of the following analysis is to compute the small fluctuation spectra on these branes and confirm the expected spectrum .
in addition we will compute the semiclassical reflection coefficient and determine the density of open string states .
we use the dbi lagrangian for d1 strings with embedding @xmath64 where @xmath65 and we assume a fluctuation @xmath66 around the classical background .
the time is @xmath67 not to be confused with @xmath17 .
expanding the dbi action we get : @xmath68 \nonumber\end{aligned}\ ] ] the first term is the energy density of our d - brane .
the first order variation vanishes when we integrate it on the world - volume of the brane since to have a small fluctuation @xmath69 for @xmath70 in cartesian coordinates .
the equations of motion for the quadratic fluctuations give a hypergeometric equation .
the solution shows no poles which would signify bound states and it turns out that for the wave like solutions with the correct asymptotic behavior the solutions is : @xmath71 where we have assumed ( semiclassical approximation ) @xmath72 , for @xmath73 .
the term @xmath74 describes an outgoing wave and @xmath75 an incoming one .
the ratio of the coefficients of these two terms is the semiclassical approximation to the reflection coefficient of a state coming from infinity .
now , we can compute the semiclassical limit of the relative density of states between a brane with modulus @xmath76 and a reference brane @xmath77 : @xmath78 where the function @xmath79 is the semiclassical limit of the reflection coefficient for an open string state .
we compute a relative density of states since the density of states for a single brane is infinite due to the non - compactness of its world - volume .
our results agree with @xcite .
in this note we have determined solutions of the dbi for d0 , d1 and d2 branes in the cigar and trumpet geometries . some aspects of those branes in realistic string backgrounds will be further discussed in @xcite.we have also computed the small fluctuations spectrum for one kind of them ( the @xmath61 ) and found agreement with geometrical and cft expectations .
one might wonder if we can get useful information for the winding states on d1cig , by studying the spectrum of momentum modes in its t - dual , the @xmath80 .
we have calculated the small fluctuation spectra of @xmath80 and found both discrete series states as well as continuous representations .
the discrete series states appear with degeneracy one . for the continuous part of the spectrum the density of states turns out to be the same as ( [ pst ] ) . as a final remark , in @xcite the spectral flow of semiclassical
solution gives strings which wind half a time around the center of @xmath15 .
these in the cigar should descent to half - winding open string states which is intuitively obvious from the shape of the @xmath61 .
this implies that on the trumpet the @xmath80 should allow states which have half - angular momentum and are periodic only under a @xmath81 rotation .
this deserves further investigation and we hope to return to this point in the future . c. bachas and m. petropoulos , jhep * 0102 * ( 2001 ) 025 [ arxiv : hep - th/0012234 ] .
p. m. petropoulos and s. ribault , jhep * 0107 * ( 2001 )
036 [ arxiv : hep - th/0105252 ] .
p. lee , h. ooguri , j. w. park and j. tannenhauser , nucl .
b * 610 * ( 2001 ) 3 [ arxiv : hep - th/0106129 ] . c. bachas
, m. r. douglas and c. schweigert , jhep * 0005 * ( 2000 ) 048 [ arxiv : hep - th/0003037 ] .
bordalo , s. ribault and c. schweigert , jhep * 0110 * ( 2001 ) 036 [ arxiv : hep - th/0108201 ] . j. m. maldacena , g. w. moore and n. seiberg , jhep * 0107 * ( 2001 ) 046 [ arxiv : hep - th/0105038 ] .
s. ribault , v. schomerus and j. teschner ; work in progress e. witten , phys .
d * 44 * ( 1991 ) 314 .
r. dijkgraaf , h. verlinde and e. verlinde , nucl .
b * 371 * ( 1992 ) 269 .
a. rajaraman , arxiv : hep - th/0109200 .
g. sarkissian , jhep * 0207 * ( 2002 ) 033 [ arxiv : hep - th/0205097 ] . t. quella , jhep * 0212 * ( 2002 ) 009 [ arxiv : hep - th/0209157 ] .
s. fredenhagen and v. schomerus , jhep * 0202 * ( 2002 ) 005 [ arxiv : hep - th/0111189 ] .
k. gawedzki , annales henri poincare * 3 * ( 2002 ) 847 [ arxiv : hep - th/0108044 ] .
s. elitzur and g. sarkissian , nucl .
b * 625 * ( 2002 ) 166 [ arxiv : hep - th/0108142 ] .
m. a. walton and j. g. zhou , nucl .
b * 648 * ( 2003 ) 523 [ arxiv : hep - th/0205161 ] .
e. kiritsis and c. kounnas , phys .
b * 320 * ( 1994 ) 264 [ addendum - ibid .
b * 325 * ( 1994 ) 536 ] [ arxiv : hep - th/9310202 ] .
e. kiritsis , c. kounnas and d. lust , int .
j. mod .
phys . a * 9 * ( 1994 ) 1361 [ arxiv : hep - th/9308124 ] .
c. bachas , a. fotopoulos , m. petropoulos and s. ribault ; work in progress j. m. maldacena and h. ooguri , j. math .
* 42 * ( 2001 ) 2929 [ arxiv : hep - th/0001053 ] . b. ponsot , v. schomerus and j. teschner , jhep * 0202 * ( 2002 ) 016 [ arxiv : hep - th/0112198 ] .
lee , h. ooguri and j. w. park , nucl .
b * 632 * ( 2002 ) 283 [ arxiv : hep - th/0112188 ] . | in this note we examine some semiclassical features of d - branes in the sl(2)/u(1 ) gauged wzw model and determine the small fluctuation spectra for one class of branes .
we compare our results with expectations from the cft side .
cpht - rr 014.0403 + imperial / tp/2 - 03/19 + hep - th/0304015 + |
the epoch of reionization ( eor ) represents a milestone in the evolution of our universe , during which hydrogen gas in most of its volume changed from fully neutral to fully ionized .
observations of the cosmic microwave background support that reionization started at redshift @xmath12 while quasars observations demonstrate that the process was completed at @xmath13 ( e.g. * ? ? ?
* ; * ? ? ?
one key to new insights in this area are observations of high redshift galaxies , as they will allow for the construction of a complete physical picture of the reionization process
. observations of distant lyman break galaxies ( lbgs ) and lyman @xmath0 emitters ( laes ) in the redshift range @xmath14 have provided the most prominent windows on the star formation history at these early epochs ( see e.g. * ? ? ? * and references therein ) .
lbgs are galaxies that have been selected using broad - band surveys , which are sensitive to the ` lyman break ' in the spectra of star forming galaxies .
this break is a consequence of typical effective temperatures of o & b stars , and of interstellar and intergalactic absorption of flux blueward of the ly@xmath0 wavelength @xcite .
surveys for lbgs ( or ` drop - out ' galaxies ) are most sensitive to uv rest - frame continuum emission of galaxies .
laes are galaxies that have been selected from narrow - band surveys which are sensitive to high - redshift galaxies that have strong ly@xmath0 line emission ( e.g. * ? ? ?
the ly@xmath0 line originates in hii regions that surrounded o & b stars . as a result , the ly@xmath0 line flux ( and strength ) in laes are more directly connected to the ionizing photon budget .
the ` strength ' of the lyman @xmath0 emission line is quantified by its rest - frame equivalent width ( ew ) , which corresponds to the ratio between the line intensity and the continuum .
the ly@xmath0 ew is an important physical quantity associated with galaxies during the eor for ( at least ) three reasons : ( _ i _ ) narrowband surveys for laes are typically sensitive to galaxies for which the ly@xmath0 ew exceeds some threshold value ( which can range between ew@xmath15 , see e.g. * ? ? ?
the ly@xmath0 ew is clearly an important physical quantity that links the lbg and lae populations ; ( _ ii _ ) the ` first ' galaxies that formed from pristine gas are expected to be characterised by unusually large ew ( well in excess of ew@xmath16 ) , which may be one of their most prominent observational signatures ; ( _ iii _ ) recent observations of laes & lbgs in the redshift range @xmath17 have revealed an apparent non - monotonic evolution in the ew distribution , with significantly less ly@xmath0 being observed from galaxies beyond redshift @xmath18 @xcite .
the redshift evolution in the ew distribution at @xmath18 is likely to be connected with the reionization process rather than with secular evolution process in the galaxies ( * ? ? ?
* also see @xcite ) .
although current observations do not allow for robust conclusions , analyzing ew distributions provides a powerful technique to probe the population of star forming galaxies as well as the ionization state of the igm during and after reionization has been completed ( e.g. * ? ? ?
* also see @xcite ) .
recently , @xcite showed that at low levels of star formation ( sfr @xmath19 yr@xmath3 ) , the effects of the statistical description of the stellar initial mass function ( imf ) and the cluster mass function ( cmf ) emerge . among the expected effects
gathered under the name of ` stochasticity ' the most relevant for our discussion is that the flux ratio between ionizing and far uv photons fluctuates around the mean expected value .
@xcite show how the clustering of star formation together with the sampling of the initial mass function ( imf ) cause the ratio of the h@xmath0 flux and fuv flux density to fluctuate around the mean significantly for local dwarf galaxies .
this effect should be also noticeable in the ratio of ly@xmath0 flux to fuv flux density i.e. the ew .
in other words , stochasticity may affect the _ intrinsic _ ly@xmath0 ew - pdfs of faint lbgs ( @xmath20 ) . because ly@xmath0 ew plays an important role in understanding high redshift galaxy populations and constraining the reionization process
, it is important to understand quantitatively the impact of stochasticity .
furthermore , observational constraints on the ` escape fraction ' of ionizing photons from star forming galaxies are derived from the observed ratio of lyc flux density to fuv flux density , which is subject to the clustering of star formation and stochastic sampling of the imf , especially at sfr@xmath21 yr@xmath3 .
it has long been realized that such galaxies can provide the bulk of the ionizing photons needed to reionize the igm and to maintain it ionized ( at a fixed escape fraction , see e.g. fig 31 of * ? ? ?
a more recent analysis by @xcite echoes these findings .
@xcite show that in plausible scenarios for reionization , star forming galaxies with very faint magnitudes @xmath22 ( corresponding to sfr @xmath23yr@xmath3 ) contribute significantly to the overall budget of ionizing photons .
while high redshift galaxies with such low star formation rates remain too faint to be detected even by future facilities , the observation of galaxies with sfrs@xmath24 yr@xmath3 at lower redshifts @xmath25 can help us to understand the physics of such systems .
for instance , the escape fraction of ionizing photons from galaxies with sfr within the upper end of that range have already been constrained by observations at @xmath26 ( * ? ? ?
* who constrained the escape fraction down to sfr@xmath27 yr@xmath3 ) .
the possible dependence of these estimates on stochasticity provides an additional motivation to study the consequences of the stochastic sampling of the imf . in this paper we set out to quantify the effects of stochasticity in the ew distribution of dwarf ly@xmath0 emitting galaxies .
we implement highly simplified models for dwarf galaxies at high redshift to isolate the stochastic effect .
the sfh is constant , single metallicity , without any extinction effects .
we stress that our objective is not to explain or interpret current observations of laes and lbgs , but instead simply to isolate and highlight the impact of stochasticity on predicted values for the ly@xmath0 ew .
this paper is structured as follows . in [ sec : ew ] we summarize the expected effects on the equivalent width stochasticity and present the numerical experiments we perform to quantify this effect . in [ sec : results ] we present the raw results that help us in [ sec : discussion ] to asses the impact of stochasticity in different aspects of the interpretation of observations of laes and lbgs at high redshift .
finally , we summarize our conclusions in [ sec : conclusions ] throughout this paper we will study the ` lyman continuum ' ( @xmath28 ) , ionizing continuum ( @xmath29 ) , and far uv ( fuv ) continuum ( @xmath30 ) , where all wavelengths are rest frame .
whenever we refer to uv we imply fuv . also , the ly@xmath0 equivalent widths are measured in the galaxies restframe .
to be consistent with observations , we define the ly@xmath0 rest - frame equivalent ew as the ratio between the intensity of the ly@xmath0 line to the flux density redwards of the line resonance , i.e. at @xmath31@xmath32 .
the flux density at this wavelength is @xmath33 times larger than at @xmath34 , where @xmath35 @xcite denotes the observed slope of the uv continuum at the redshift and uv - magnitudes of interest .
therefore , if we had corrected for the slope in the spectrum of the uv continuum , then we would have found ew@xmath36 for the @xcite star formation calibrators ( e.g. * ? ? ?
] @xmath37 where subscript @xmath38 or @xmath39 indicates the flux density in frequency or wavelength , respectively . the wavelengths and frequencies associated with the ly@xmath0 transition and the uv continuum ( from the central wavelength for the filter used in slug ) are @xmath40 , @xmath34 , @xmath41hz and @xmath42hz .
we calculate the intrinsic luminosity of ly@xmath0 assuming that a fraction @xmath43 of the ionizing photons are converted into ly@xmath0 .
we obtain l@xmath44=@xmath45 where @xmath46 erg photon@xmath3 .
we then have @xmath47 in order to compare results on the same grounds results from different metallicities or star formation rates we introduce a dimensionless constant @xmath48 where @xmath49 is the equivalent width expected value for high values of the star formation rate when the stochasticity effects are negligible .
we point out that under this convention ew@xmath50 when considering a standard conversion between ly@xmath0 luminosity , uv luminosity density @xcite , which are l@xmath51sfr@xmath52 erg s@xmath3 hz@xmath3 and @xmath53sfr@xmath54 erg s@xmath3 .
we emphasize that an implicit assumption in the conversion is a salpeter imf in the mass range 0.1 - 100 for solar abundance . in our numerical experiments ,
ew@xmath55 , which is a result of the lower metallicities and the higher upper limit in the imf mass range assumed in slug ( details are given in [ sec : slug ] ) .
these minor systematic corrections due to different conditions in stellar metallicity , imf or star formation history wo nt affect our main conclusions . throughout this paper , we will adopt ew@xmath56 .
the description of star formation is probabilistic . at its core
we find the concept of the initial mass function ( imf ) , @xmath57 , which describes the relative abundance of stars of a given mass to be formed .
the imf can be mostly determined by two pieces of information .
the first one is the mass interval , @xmath58 that the newly formed stars can have .
the second is the change of relative abundance for different stellar masses , which is usually parametrized as @xmath59 , if @xmath60 less massive stars are more probable to form that high mass stars . for large masses of
newly formed stars it is possible to guarantee a good sampling of the imf . for a low rate of star formation
the poor sampling of the imf makes possible that the rarer most massive stars near @xmath61 do nt appear in a burst of star formation .
star formation is also thought to proceed in clusters .
this provides another important astrophysical element in the statistical description of star formation that can be impacted by stochasticity effects . in analogy to the stellar mass function , one can also parametrize the abundance of stellar clusters by the cluster mass function ( cmf ) .
low values of the star formation rate can also affect the sampling of the cmf .
furthermore , stars of different mass have different evolutionary tracks .
the inhomogeneous imf sampling will produce fluctuating stellar populations with different luminosity histories , this can impact the ratio of ionizing to far uv photons produced as a function of time . the effect of stochasticity has been explored before in the production of h@xmath0 in dwarf galaxies in the local universe , but not on the effect on the ew of lyman@xmath0 emitting galaxies .
our aim is to quantify this impact and estimate how current and future observational campaigns can be affected by these effects .
we used slug ( ` stochastically light up galaxies ' ) a fully stochastic code for synthetic photometry of galaxies @xcite .
the code samples the cmf and imf of a star forming galaxy by forming individual star clusters and following their spectral evolution . at low values of a time
averaged continuous sfr the series of bursts associated with the cluster formation start to be visible .
furthermore , low sfr values imply a low number of stars to sample the imf , lowering the probability of finding massive stars .
all the effects of finite sampling in mass and time have an impact on the instantaneous strength of the lyman@xmath0 emission line .
slug assumes that a fraction @xmath62 of the total stellar population is formed in clusters . in that case
, the clusters also sample a cluster mass function(cmf ) @xmath63 .
the stars in the cluster are then populated according to the selected imf . in this stochastic framework ,
a galaxy with a time averaged constant sfr does not level to a constant sed , but instead it continuously fluctuates in time .
for a sample of independent galaxies with the same average sfr one could thus associate a distribution of line and continuum intensities .
all models used in this paper are based on a salpeter imf in the mass intervals @xmath64@xmath65 .
we consider two kinds of models of varying metallicity @xmath66 ( which corresponds to @xmath67 , @xmath68 ) .
we assume that all the stars are formed in clusters , @xmath69 . assuming unclustered star formation , @xmath70 , reduces the scatter in the flux fluctuations we report here by about an order of magnitude @xcite .
the cmf is described by the cluster mass interval is @xmath71and the slope is @xmath72 .
the stellar libraries are padova asymptotic giant branch ( agb ) tracks . the stellar spectral energy distributions ( sed ) correspond to @xcite where the wolf - rayet stars treatment follows @xcite and @xcite .
we run 500 galaxies for each metallicity and star formation rate comprised in the range @xmath1@xmath2 /yr .
table 1 lists the exact values for the sfr and metallicity that were used .
each model runs for a period of 200 myr with outputs every 1 myr .
running for a period longer than 100 myr ensures a stable base level for the mean fuv and ionizing fluxes .
we use the last 10 outputs per galaxy to slightly increase the statistical sample .
being pessimistic , one can assume that each simulated system contributes only 1 independent datum instead of 10 , from that point of view , one can be sure that all the results for the pdf and derived quantities are accurate within @xmath73 .
in other words , one might be concerned that the 10 outputs of the same galaxy may not be independent and might bias the pdf . to quantify this
we perform the following test : for 4000 galaxies with a star formation rate @xmath74 yr@xmath3 we calculate the pdf in two different ways .
first by taking 10 outputs of 400 galaxies and second by using only 1 output for each one of the 4000 galaxies .
we find that the absolute difference between the pdf(@xmath75 ) constructed from the two different samples is always on the order of @xmath76 and conclude that a possible bias is negligible .
we perform an additional test by taking 1 output and 2 outputs for 4000 galaxies , with spectral outputs spaced by @xmath77myr during @xmath78myr and find that the difference between the pdf(@xmath75 ) constructed from these samples is always on the order of @xmath79 and conclude that a possible bias from the chosen timesteps is also negligible . for each galaxy
we obtain the flux per unit frequency @xmath80 at a wavelength of 1540 , together with the ionizing photon rate @xmath81 .
these quantities are used to calculate the ew as described by eq ( [ eq : def ] ) .
the value for ew@xmath8 is calculated as the arithmetic average ew for 30 models of galaxies with a sfr of @xmath82 yr@xmath3 , we obtain ew@xmath83 where the error quoted is estimated from the standard deviation from the mean .
[ table : fit ] [ cols="^,^,^,^,^,^",options="header " , ] , the normalized ew .
each symbol corresponds to the binned slug results for a different sfr=@xmath1 , @xmath84 , @xmath85 and @xmath2yr@xmath3 and a metalliciticy f @xmath86 .
the lines are the best fit of eq .
( [ eq : pdf - ew ] ) to the points.,scaledwidth=50.0% ] the net effects of stochasticity introduce a sfr dependent scatter around a central value .
we describe this effect as a probability density distribution ( pdf ) for the @xmath87 variable .
we parametrize the pdf by a double power law in order to better quantify how it depends from changes in the sfr and the metallicity , @xmath88^{-1}. \label{eq : pdf - ew}\ ] ] figure [ fig : pdf - ew ] presents the results for the pdf @xmath89 , the symbols represent the results from the simulations and the lines the fit to those points .
given the limited number of models the pdf is reliable only for values @xmath90 , likewise the deviation of the fit with respect to the simulation points is on the order of @xmath91 .
the star formation rate enters through the constants @xmath92 , @xmath0 , @xmath93 , and @xmath94 .
the expected result for a perfectly sampled imf is a delta function around @xmath95 @xcite , while our results show a wide distribution .
each line in figure [ fig : pdf - ew ] corresponds to a different sfr@xmath96 , @xmath84 , @xmath85 and @xmath97 in units of yr@xmath3 for a metallicity of @xmath98 .
the best fit values together with their uncertainty from the fitting procedure for all the slug models are listed in table 1 . in these values
we get a precise quantification on how the pdf shape changes with the sfr . from these results
we confirm that the effect of the metallicity is a second order effect shaping the ew pdf when compared to the influence of the sfr
. for high values of the sfr@xmath99 yr@xmath3 the low end , @xmath100 of @xmath101 increases sharply compared to the slower decrease in the high end @xmath102 . in this regime the peak of the distribution is around the expected value of @xmath103 for a perfectly sampled imf .
this quantitative behaviour presents consequences for the interpretation of high ew systems . on the upper end , on
can achieve values close to what is expected to be limit of star forming galaxies : @xmath104 which corresponds to @xmath105 in our models . for sfr@xmath106 yr@xmath3 , the low end of the pdf shows a bulkier shape that confirms the existence of systems with @xmath100 purely as a consequence of their intrinsic star formation stochasticity .
we quantify and elaborate in the next section the implications of these results .
each line represents different threshold in ew@xmath107 , @xmath108 , @xmath109 and @xmath110 .
this figure quantifies the impact of the stochasticity on the fraction of confirmed laes based on a minimal ew detection threshold.,scaledwidth=50.0% ] the ensuing discussion will focus on the interpretation of high and low ew systems that naturally emerge through stochasticity .
high ew vales might lead one to propose the presence of very low metallicity ( or population iii ) stars while low ew values could cause some galaxies not to be selected as a lae in very deep narrowband surveys .
a common assumption in the search for galaxies with population iii stars is the following : there is an upper ew value to consider a star forming galaxy to be normal .
this threshold value is taken to be ew@xmath111 but , as we show in the previous section , galaxies with similar or larger ew values can be expected by stochasticity effects without demanding pop iii stars . the threshold value ew@xmath112 corresponds to @xmath105 ( eq . [ eq : m ] ) .
we can then quantify the fraction of such galaxies that have ew that exceed the threshold by using eq .
( [ eq : pdf - ew ] ) . for galaxies with star formation rates in the range
@xmath113@xmath2 yr@xmath3 we find that there would be between @xmath114 of systems with ew values larger than ew@xmath115 .
we calculate same fraction directly from the @xmath116 points used to build the pdf and find that between @xmath117 to @xmath118 among them have an ew larger than @xmath119 .
this suggests that a minimal threshold in ew is not enough to assert that an individual galaxy presents signatures of population iii stars .
however , also the effects of stochasticity can not be observationally evaluated from the measurement of an individual galaxy
. only statistical studies of galaxy samples are an appropriate tool to find clear signposts either of special stellar populations or stochasticity .
narrowband surveys for laes often employ color cuts that can be approximated by requiring that the ew of a galaxy exceeds threshold value , which typically lies in the range ew@xmath120 (see [ sec : introduction ] ) .
the result of previous sections show that low values of the sfr@xmath106 m@xmath121 yr@xmath3 the ew can be lower than half of the expected ew@xmath8 . in a survey that adopts color - cuts that correspond to a high ew@xmath122 , one could simply miss those objects due to their low intrinsic ew .
this shows how stochasticity can induce selection bias in laes observations .
not surprisingly , the impact of stochasticity is highly sfr dependent .
the strong dependence of the shape of the @xmath75pdf on sfr causes a larger fraction of systems to be found below a fixed @xmath75 threshold for detection as the sfr decreases . in figure
[ fig : ew_cut ] we show the fraction of laes that are missed as a function of the star formation rate , for different cuts in the ew expressed by ew@xmath123 .
these fractions are calculated directly from the slug models , providing an uncertainty on the order of @xmath124 in the reported fraction of detected galaxies , the discrepancy of calculating the same fractions from the fitting function is in all cases smaller than @xmath76 .
the value of ew@xmath107 corresponds to a common cut used in present day observations ( e.g. * ? ? ? * ) .
a cut of ew@xmath125 represents the cut used in the early observational attempts to find lyman @xmath0 emitting galaxies ( e.g. * ? ? ? * ) . in
[ sec : lfew ] we investigate how such a bias may affect predicted number densities of laes in simplified ( but often employed ) models of laes .
in this section we study the potential impact of stochasticity on the predicted ly@xmath0 luminosity functions & equivalent width distribution of narrowband selected galaxies ( i.e. laes ) .
we compare two very simple models to isolate and highlight the impact of stochasticity . in the first model we ignore stochastic ffects .
we further assume that each star forming galaxy has a ly@xmath0 emission line with ew@xmath126ew@xmath127 (see [ sec : ew ] ) . in the second model
, we instead assume that the ly@xmath0 emission line of a star forming galaxy is drawn from a pdf which is set entirely by stochasticity effects .
we therefore assume that the function @xmath128 is given by eq [ eq : pdf - ew ] for the second model .
finally , we further assume that there is a one - to - one correspondence between sfr and @xmath129 ( given by the standard conversion factor between sfr and uv continuum flux , see
[ sec : ew ] ) , and therefore , and the proper ew - pdf for a fixed @xmath129 is @xmath130 .
because the pdf @xmath128 varies relatively weakly with sfr , this proper calculation barely changes our results at all . ] that @xmath131 @xmath132 . under these assumptions
, we can predict the lae ly@xmath0 luminosity function , @xmath133 , for both models as @xcite @xmath134 where @xmath135 denotes the comoving number density of laes in the luminosity range @xmath136 .
furthermore , @xmath137 denotes the comoving number density of star forming galaxies with absolute uv - magnitude in the range @xmath138 .
the equivalent width ew@xmath139 denotes the ew which corresponds to a ly@xmath0 line of luminosity @xmath4 for an absolute uv magnitude @xmath129 ( see dijkstra & wyithe 2012 for details ) .
we focus our analysis on @xmath140 and compare to the observed lae ly@xmath0 luminosity functions at this redshift by @xcite .
we take the observed @xmath137 at @xmath141 from @xcite and ` rescale ' it to @xmath140 assuming that @xmath142 evolves with redshift as @xmath143 @xcite .
figure [ fig : lf ] shows the luminosity function for the first model ( no stochasticity ) as the _ black solid line _ , and the second model ( with stochasticity ) as the _ red dashed line_. the data points represent observations by @xcite . both models are remarkably similar .
the reason for this similarity is non - trivial : stochasticity decreases the ` most likely ' ew for star forming galaxies compared to models that do not include this effect ( see fig [ fig : pdf - ew ] ) .
this decrease alone would reduce the predicted ly@xmath0 luminosity functions .
however , it is compensated for by the fact the @xmath75-pdf has a tail to values @xmath144 .
the presence of this tail allows for an additional contribution to @xmath133 ( at a given @xmath4 ) by uv - faint galaxies that would not be able to contribute in the absence of stochasticity .
our prediction further assumes that @xmath145=@xmath146 , and therefore likely overestimates the impact of stochasticity at sfr@xmath147 yr@xmath3 ( which corresponds to @xmath148 erg s@xmath3 ) .
we therefore safely conclude that stochasticity barely impacts predictions of ly@xmath0 luminosity functions .
this result is not sensitive to the uncertainties in the fitting functions for the pdf(@xmath149 ) used here . also shown for completeness is the observed @xmath140 lae luminosity function from ouchi et al .
both models clearly predict more laes than is observed .
this is not surprising as we have not attempted to correct for dust extinction and/or intergalactic scattering .
both these effects likely reduce the predicted ly@xmath0 luminosity at a fixed number density of objects , which would bring down the predicted luminosity functions . [ fig : lf ] luminosity function of laes for simplified models with ( without ) the effects of stochasticity included as the _ red dashed _ ( _ black solid line _ ) . also shown
are the observations taken from ouchi et al .
( 2008 , _ black filled circles _ ) .
the main point of this plot is that stochasticity barely affects the predicted luminosity functions .
both models predict more laes than is observed , as we have not attempted to correct for dust extinction and/or intergalactic scattering , which would both reduce the predicted luminosity functions.,title="fig:",scaledwidth=50.0% ] the ew - pdf for a sample of ly@xmath0-selected galaxies is given by ( dijkstra & wyithe 2012 ) @xmath150 where @xmath151 ( @xmath152 ) denotes the minimum ( maximum ) ly@xmath0 luminosity of the galaxies in the sample .
furthermore , @xmath153 denotes the uv magnitude for which a ly@xmath0 line of luminosity @xmath4 corresponds to an equivalent width ew ( again see dijkstra & wyithe 2012 for details ) .
we focus on the impact of stochasticity on the predicted pdf for @xmath154 and @xmath155 ( _ blue dotted line _ ( which corresponds approximately to @xmath156sfr@xmath157 , and roughly to the range of observed ly@xmath0 luminosities ) . for comparison , we also show a case with @xmath158 and @xmath159 ( _ red dashed line _ , which corresponds approximately to @xmath160sfr@xmath161 ) .
the choice @xmath158 is a bit arbitrary , but was motivated by the notion that galaxies with sfr@xmath162 yr@xmath3 can contribute significantly to the overall ionizing photon budget during the epoch of reionization ( see kuhlen & faucher - giguere 2012 ) .
[ fig : ewpdf ] selected galaxies .
the _ solid black line _ shows the dirac delta function for the model with no stochasticity effects .
the _ red dashed line _ ( _ blue dotted line _ ) shows the predicted ew - pdf for models which include stochasticity effects for ly@xmath0 selected galaxies with @xmath163 erg s@xmath3 ( @xmath164 erg s@xmath3 , see text ) .
also shown for comparison is the ew - pdf observed by ouchi et al .
( 2008 ; the _ red filled circles _ represent their photometric data , while the _ blue filled circles _
represent their spectroscopic sample ) .
this figure illustrates that stochasticity may help explain the large observed values of ew of laes at z=5.7 , although it fails to fully explain the observed ew - pdf.,title="fig:",scaledwidth=50.0% ] the _ black solid line _ in figure [ fig : ewpdf ] shows the predicted ew - pdf for the model with no stochastic effects ( which is simply a delta - function ) .
the predicted pdf for the models that include stochasticity peak at the lower values of ew , and have tails that extend to values that greatly exceed ew@xmath165 .
also shown is the observed ew - pdf for a sample of photometrically selected ( _ red filled squares _ ) and spectroscopically confirmed ( _ blue filled circles _ , taken from ouchi et al .
2008 ) laes .
clearly , stochastic effects may help to explain large observed values of ew .
however , additional processes such as dust extinction and/or intergalactic scattering , which impacts differently the uv continuum and the ly@xmath0 line @xcite , are needed to fully explain the observed ew - pdf of laes at z=5.7 .
the distribution of ew at a given averaged sfr is the result of fluctuations in the fluxes of ionizing and far uv continuum ( non ionizing ) photons .
the ionizing photon production is dominated by short lived o & b stars , while less massive a stars can contribute significantly to the non - ionizing photon flux . because of the different lifetimes of these stars of different masses , the ionizing and non ionizing fluxes and therefore their ratio
vary on different timescales and with different amplitudes .
simultaneous fluctuations in the ionizing and non - ionizing fluxes can thus impact observational estimates of the ionizing escape fraction , which is usually estimated from a _ relative escape fraction _ :
@xmath166 where the @xmath167 is the intrinsic flux ratio for a stellar population model , @xmath168 is the observed flux ratio and @xmath169 is the lyc optical depth of the igm at the redshift of emission . based on subaru / suprime - cam imaging
@xcite claimed a detection of 10 laes and 3 lbgs at @xmath170 with a significant emission in lyc out of a sample of 125 laes and 73 lbgs .
recent measurements seem to confirm such measurement for 8 laes using spectroscopy with vlt / vimos and subaru / focas .
the results indicate a lyc excess with respect to non - ionizing uv at @xmath30 expected by standard stellar populations @xcite .
however , observation targeting lyc in lbgs at @xmath171 using deep hubble space telescope imaging do not find any excess @xcite .
the sources present an excess of a factor of @xmath172 times larger than the expected intrinsic stellar ratio for @xmath173 .
some theoretical analysis requires a very top heavy imf to explain the observed ratios or fractions of @xmath174 in mass of very young @xmath175 myr of extremely metal poor @xmath176z@xmath121 or metal free stellar populations @xcite the stochasticity effects present an important element to be considered in such analysis : the ratio @xmath87 can be directly interpreted as the excess of the @xmath177 ratio with respect to the constant expected value .
the star formation rates for the objects with lyc excess are on the order of @xmath178 yr@xmath3 ( @xmath179 , see fig 5 of siana et al . 2010 and where we ignore dust in the conversion from @xmath129 to sfr ) . in this range
@xmath180 of the galaxies will show an excess @xmath181 times the expected stellar @xmath177 ratio only due to stochasticity ( figure 1 ) , this is consistent with the fraction of galaxies observed with those characteristics .
the stochasticity might also explain why the lyc excess is not observed in the recent @xmath171 hubble imaging by @xcite .
the lbgs observed at that redshift have systematically restframe magnitudes @xmath182 , while the lyc excess sources are fainter than this limit .
this transition roughly corresponds to a sfr @xmath183 yr@xmath3 assuming a standard conversion factor , while the fainter systems reported with an excess have @xmath184 and sfr @xmath185 yr@xmath3 .
the dependence of the stochasticity effects on the sfr make it more likely to detect the lyc excess in the fainter sample .
probably the stronger effect is of statistic origin , given that the observations at @xmath171 target 15 lbgs , while at @xmath186 a set of 198 galaxies were observed , this provides one order of magnitude less objects to sample the @xmath101 distribution and find galaxies with a lyc excess .
in this paper we have investigated the quantitative impact of stochastic sampling of the stellar initial mass function ( imf ) on the ly@xmath0 , and ionizing photon emissivity of dwarf galaxies .
we have used the slug code to simulate the spectral energy distribution for restframe wavelengths @xmath187 of galaxies with time - averaged star formation rates in the range @xmath188 yr@xmath3 . from our predicted far uv continuum flux density ( at @xmath189 ) and the ionizing flux density ( at @xmath190 ) we construct probability distribution function ( pdf ) for the variable @xmath87 that represents the ratio of the measured restframe equivalent width ( ew ) of the ly@xmath0 line to the expected constant value in the absence of stochasticity ( ew@xmath191 ) , i.e. @xmath192ew / ew@xmath191 .
we find that the @xmath75pdf can be represented by a double power law in @xmath87 , whose parameters are highly dependent on the sfr .
we emphasize that the results derived from these experiments are sensitive to the degree of the clustering properties of the stars . in the case of @xmath70 where the star formation is completely unclustered ,
the dispersion around the mean is reduced by a factor of 10 @xcite .
the results are also sensitive to the star formation history assumed : in the case of a bursty star formation history , one could expect the scatter around the mean for @xmath75 to be higher than what we derive here .
our results show that it is possible for galaxies to have both extremely low and high @xmath87 values especially at increasingly low sfr and investigate implications of this result .
in particular , we show how the existence of galaxies with low @xmath87 values can induce an observational bias if galaxies are selected by cuts in the equivalent width ( as in narrowband surveys for laes ) .
this bias may cause dwarfs with lower sfrs @xmath193 yr@xmath3 more likely to be missed in narrowband surveys . on the other hand
, the existence of galaxies with high @xmath87 implies that a high ew can not unambiguously be interpreted as a defining characteristic of primordial population iii galaxies .
finally , we use the framework presented in dijkstra & wyithe 2012 to quantify the impact of stochasticity on the predicted luminosity function and ew - pdf of a sample of ly@xmath0 emitters ( laes , i.e. narrowband selected galaxies ) at a fixed redshift .
we find that while stochasticity does not appreciably affect the lae abundance modeling , it may help to explain the large ew ( ew@xmath9 ) observed for a fraction of laes .
finally , we have also shown that stochasticity can affect inferred constraints on the escape fraction of ionizing radiation , most strongly in galaxies that form stars at rates sfr@xmath194 yr@xmath3 .
given that such galaxies are thought to dominate the reionization process at @xmath18 , it is important to consider the effects of stochasticity when using ` local ' observations to put constraints on their ionizing luminosities and escape fractions .
in particular , the observed anomalies in the ratio of lyc to non - ionizing uv radiation in so - called ` lyman - bump ' galaxies at @xmath10 reported by @xcite can be explained naturally in the context of the stochastic sampling of the imf , without resorting to unusual stellar populations .
moreover , the absence of such objects in a sample of @xmath195 lbgs at @xmath171 @xcite is also expected as these galaxies are forming stars at rates at which the stochastic sampling of the imf becomes negligible .
jef - r acknowledges support from the peter and patricia gruber foundation through their fellowship , administered by the international astronomical union .
we thank diederik kruijssen for useful discussions on stellar clusters and imf sampling .
l. , fontana a. , vanzella e. , castellano m. , grazian a. , dijkstra m. , boutsia k. , cristiani s. , dickinson m. , giallongo e. , giavalisco m. , maiolino r. , moorwood a. , paris d. , santini p. , 2011
, apj , 743 , 132 | observations of distant galaxies play a key role in improving our understanding of the epoch of reionization ( eor ) . the observed ly@xmath0 emission line strength quantified by its restframe equivalent width ( ew ) provides a valuable diagnostic of stellar populations and dust in galaxies during and after the eor . in this paper
we quantify the effects of star formation stochasticity on the predicted ly@xmath0 ew in dwarf galaxies , using the publicly available code slug ( used to ` stochastically light up galaxies ' ) .
we compute the number of hydrogen ionizing photons , as well as flux in the far uv for a set of models with star formation rates ( sfr ) in the range @xmath1-@xmath2 yr@xmath3 . from these fluxes
we compute the luminosity , @xmath4 , and the ew of the ly@xmath0 line .
we find that stochasticity alone induces a broad distribution in @xmath4 and ew at a fixed sfr , and that the widths of these distributions decrease with increasing sfr .
we parameterize the ew probability density function ( pdf ) as an sfr dependent double power law .
we find that it is possible to have ew as low as @xmath5ew@xmath6 and as high as @xmath7ew@xmath8 , where ew@xmath8 denotes the expected ew in the absence of stochasticity .
we argue that stochasticity may therefore be important when linking drop - out and narrow - band selected galaxies , when identifying population iii galaxies , and that it may help to explain the large ew ( ew@xmath9 ) observed for a fraction of ly@xmath0 selected galaxies .
finally , we show that stochasticity can also affect the inferred escape fraction of ionizing photons from galaxies .
in particular , we argue that stochasticity may simultaneously explain the observed anomalous ratios of the lyman continuum flux density to the ( non - ionizing ) uv continuum density in so - called lyman - bump galaxies at @xmath10 , as well as the absence of such objects among a sample of @xmath11 drop - out galaxies .
galaxies : high - redshift - galaxies : star formation - line : formation |
magnetically dominated solar plasma consists of a variety of structures , such as active regions , coronal loops , x - ray bright points , coronal holes etc . anchored in the photosphere ,
coronal loops populate both active and quiet regions , and form the basic building blocks of the sun s atmosphere .
the physics of all kinds of loops , therefore , holds the key to understanding coronal heating , solar wind acceleration and the flow of mass and energy in the region . despite extensive work on the development of theoretical modelling
( see e.g. , * ? ? ?
* ; * ? ? ?
* for a review ) the energy source , structure maintenance and mass balance in coronal loops are not yet fully understood .
the ultimate solution of this problem requires precise measurements of physical plasma parameters such as flows , electron temperature , density and filling factors in spatially resolved coronal structures .
the measurement of above mentioned parameters have been performed using earlier spectrometers . however , the measurements were limited by spectral and spatial resolutions .
moreover the simultaneous temperature coverage has always been an issue .
the extreme - ultraviolet imaging spectrometer ( eis ; * ? ? ?
* ) on board hinode spacecraft provides an excellent opportunity to measure these parameters in greater detail . in this paper
we have studied overall intensity and flow structure in an active region .
moreover , we have studied flows , electron temperature , density and filling factor along a well resolved quiescent active region coronal loop .
the rest of the paper is structured as follows . in section 2 ,
we provide observations and data reduction .
we present out results in section 3 with a brief summary and conclusion in section 4 .
the extreme - ultraviolet imaging spectrometer ( eis ; * ? ? ? * ) on board hinode is an off - axis paraboloid design with a focal length of 1.9 meter and mirror diameter of 15 cm .
it consists of a multi - toroidal grating which disperses the spectrum on two different detectors covering 40 each .
the first detector covers the wavelength range 170 - 210 and second covers 250 - 290 providing observations in a broad range of temperature ( @xmath0 5.8 - 6.7 mk ) .
the eis has four slit / slot options available ( 1 , 2 , 40and 266 ) .
the eis provides monochromatic images of the transition region and corona at a high cadence using a slot ( wide slit ) .
high spectral resolution images can be obtained by rastering with a slit .
the eis observed an active region on the sun s disk center on may 19 , 2007 using the observation sequence _
ar_velocity_map_. this sequence uses the 1slit with an exposure time of 40 seconds with a delay of 12 seconds .
the eis raster used in this analysis started at 11:41:23 ut and finished at 16:35:01 ut .
the left panel in fig .
[ context ] displays an image recorded by the transition region and coronal explorer ( trace ; * ? ? ?
* ) in its 171 pass - band showing the complete active region taken at 13:39:35 ut . the over - plotted box on the trace image represents the field - of - view ( fov ) of the eis raster . the right panel of fig .
[ context ] shows an image recorded by the eis in @xmath6195 . during
this observation hinode entered into eclipse mode three times , thereby causing data gaps .
after the first eclipse , eis observed a few well resolved loop structures between x=[0:50 ] arcsec .
the white line demarks the slit location at 13:39:12 ut corresponding to the time of trace image shown in the left panel .
the sequence used for this study comprises very many useful spectral lines . for this particular study
, we have only selected a few windows with lines such as @xmath6275 ( log t[k ] = 5.8 ) , @xmath6278 & 280 ( log t[k ] = 5.8 ) , @xmath6186 ( log t[k ] = 5.8 ) , @xmath6184 ( log t[k ] = 6.0 ) , @xmath6258 & 261 ( log t[k ] = 6.1 ) , @xmath6195 & 186 ( log t[k ] = 6.1 ) , @xmath6196 & 203 ( log t[k ] = 6.2 ) , @xmath6264 & 274 ( log t[k ] = 6.3 ) and @xmath6284 ( log t[k ] = 6.4 ) .
the selection of these lines provides diagnostics for the active region in a broad range of temperatures .
the data were first processed using the standard processing routine _ eis_prep.pro _ provided in _ solar software ( ssw ) _ tree .
we have fitted all the lines at each pixel of the eis raster using the routine _ eis_auto_fit.pro _ which is also provided in _
ssw_. note that the @xmath6278 , and @xmath6274 are blended with a line .
this blending issue is taken care of by using the branching ratios with @xmath6275 .
the @xmath6203 is blended with a line which is taken care of by fitting two gaussians .
the two lines @xmath6195 and @xmath6186 are self blended and care must be taken while determining electron density using these line ratios ( see * ? ? ?
* ; * ? ? ?
it has been pointed out that images obtained from the two eis detectors have an offset of about 20@xmath7 in y ( n - s ) direction and about a couple of arcsec in x ( e - w ) direction .
note that the eis slit is oriented along n - s .
therefore a co - alignment is necessary when comparing the outputs obtained from the two detectors . for this purpose
we have used images obtained at similar temperatures ( log t[k ] = 6.1 ) from two ccds ( e.g. , @xmath6261 & @xmath6195 ) and co - aligned them using a cross - correlation technique . in order to derive physical parameters such as plasma flows , temperature , electron density and filling factors ,
a number of different spectroscopic techniques can be applied to eis observations . for a review on different spectroscopic techniques
see e.g. @xcite .
the intensity of an optically thin emission line can be written as @xmath8 where a(z ) is the elemental abundance , @xmath9 is the electron temperature , and @xmath10 is the electron density .
g(@xmath9 , @xmath10 ) is the _ contribution function _ which contains all the relevant atomic parameters for each line , in particular the ionization fraction and excitation parameters . in quiet solar conditions
it is usually assumed that ionization equilibrium holds and the g(@xmath9 , @xmath10 ) can be obtained using equilibrium ionization balance calculations .
for this purpose we have used chianti , v5.2 @xcite .
@xmath11 is known as _ differential emission measure _ which is defined as : @xmath12 where dh is an element of column height along the line - of - sight ( los ) . to determine the average temperature along the los for a given point along the loop we have used the _ em - loci _ method ( see * ? ? ? * and references therein ) . in this method
we plot the ratios of observed intensities with the contribution function i@xmath13/(a(z ) g(n@xmath14 , t@xmath14 ) ) obtained using ionization balance calculations as a function of temperature .
if the plasma is isothermal along the los then all the curves would cross at a single location indicating a single temperature .
more precisely , this method is useful to find out if the plasma along the los is isothermal or multi - thermal .
the eis wavelength bands were chosen to cover many density diagnostic lines ( see e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
we provide measurements of average electron density along the los using line ratios simultaneously from different line pairs .
once the electron densities have been estimated , the plasma filling factor ( @xmath15 ) - yet another important quantity - can be determined as @xcite @xmath16 where em is the total emission measure which is calculated by integrating over the width of the contribution function for a given spectral line , and h is the thickness of the loop along the los .
obtaining the filling factor provides an indication of whether the loop structures are composed of multiple strands or not .
measuring the filling factors is very difficult due to the accumulated errors in estimating electron density , total emission measure ( i.e. , ionization equilibrium calculations ) and the column depth . the main uncertainty , however , comes from the estimation of the column depth of the emitting plasma .
one of the main problems when studying the plasma parameters in coronal loops is to identify a loop which does not show substantial changes during the observing period .
investigating the observation sequence taken by the trace , we find that the well defined loop structures seen after the first hinode eclipse do not change significantly , at least within the time period when the eis was rastering over them . although there was a filament eruption at around 12:56 ut , one of the legs of which was rooted just west of the active region loops , the loops studied in this paper do not show any significant variations which could affect our study of the physical parameters . to be absolutely sure of this ,
we have also studied a raster by the eis taken before the eruption ( start : 09:42:12 , end : 10:31:24 ut ) .
we find that the loop structure has not changed from that raster to the one for which we have derived the physical parameters .
we did not use the earlier raster to derive the physical parameters because the exposure time was short and therefore , we did not have enough counts for a good diagnostic study . since this study provides measurements of the physical parameters along a _ quiescent coronal loop _
, the results obtained can be reliably used as a benchmark against theoretical models of active region loops .
one of the most important features of the eis is to provide simultaneous observations at many different temperatures with an excellent spectral resolution ( @xmath0 3 km s@xmath1 ) .
this , therefore , helps us enhance our understanding of the overall temperature structure and plasma flows in active regions .
[ int_vel ] , displays intensity ( first and third rows ) and corresponding velocity maps ( second and fourth rows ) taken simultaneously at different wavelengths .
the three black stripes in the intensity and velocity images are data gaps due to hinode eclipse .
it is evident from the intensity images ( top row ) that high - lying coronal loop structures are more prominent at cooler temperatures .
the core of the active region is seen in high temperature lines .
the active region shape become more compact and bright at temperatures such as log t[k]=6.3 ( ) and log t[k]=6.4 ( ) .
this implies that the core of the active region is hotter confirming previous observations by e.g. , @xcite using data from the coronal diagnostic spectrometer ( cds ; * ? ? ?
* ) aboard the solar and heliospheric spectrometer ( soho ; * ? ? ?
* ) and @xcite using eis / hinode observations . the footpoint emission from the high - lying loops marked with arrows in top row images in fig . [ int_vel ] are most intense in low temperature lines such as ( top left image in fig .
[ int_vel ] ) ( and , not shown here ) with little emission in high temperature lines such as , and .
this was also shown by @xcite using hinode / eis observations .
the high - lying loop structures are best defined in the lower temperature lines such as ( log t[k]=6.0 mk ) .
the loop structures merge into the diffuse background emission with increasing temperature ( see also * ? ? ?
in hot temperature lines such as ( and , not shown here ) only the core of the active region remains visible . in order to obtain the los velocity information ( see the bottom rows of in both panels of fig . [ int_vel ] ) , instrumental effects such as the eis _ slit - tilt _ and spacecraft _
orbital variation _ have to be compensated for .
it is known that the eis slits are not vertical on the detectors ( e.g. , * ? ? ?
the 1slit has a tilt of about 1.2@xmath1710@xmath18 / pixel .
therefore , while deriving velocities , it is essential to compensate for this effect .
this is done using the idl routine _ eis_slit_tilt.pro _ provided in the ssw tree . after compensating for the slit - tilt
, we applied a correction for the orbital variation using _
eis_orbit_spline.pro _ also provided in _
ssw_. exploring a large number of datasets it has been concluded that the spacecraft orbital variation changes from one eis study to another . therefore , there is no automatic way to compensate for this effect and it has to be estimated for each and every study separately . to estimate the orbital variation ,
it is necessary to select a quiet region in the raster . by doing so ,
we ensure that we do not remove ( add ) any original ( artifact ) values from ( to ) the data . for this purpose , we selected the bottom 50 pixels in the quiet - sun region in this raster . by studying the average variations of the line centroids , we could determine the effects of orbital variation . applying the above mentioned corrections
, we obtained a corrected set of data , to enable us to estimate the doppler shift in the coronal structures .
to summarize , the at rest reference wavelength has been obtained by considering the bottom 50 pixels . based on sumer measurements
, @xcite and @xcite found that the spectral lines show non - zero line shifts in quiet - sun observations .
the cooler temperature lines ( t @xmath19 0.5 mk ) showed red shifted emission while the hotter lines ( t @xmath20 0.5 mk ) were blue shifted .
ideally , while investigating the doppler shift measurements this effect should be considered .
however , we note that sumer does not observe spectral lines forming at temperature higher than log t[mk ] = 6.1 i.e. temperature corresponding to line . therefore , this effect can only be taken in account for the line such as , and in our analysis and not for high temperature lines studied in this paper such as , and .
since we are interested in comparing the flows observed in these lines with respect to each other , it is not worthwhile taking such effects into account for some lines but not for higher temperature lines . the bottom row images in fig .
[ int_vel ] show the derived doppler velocity maps for different lines .
the velocities maps are plotted in the range of @xmath21 30 km s@xmath1 .
the velocity maps show redshifts and blueshifts all over the active region .
the redshifts ( downflows ) are more prominent in the core region of the active region and blue shifts ( up flows ) are more prominent at the boundary of the active region similar to the observations made by e.g. , @xcite .
the plasma flow structures in high lying loops are better defined at lower temperatures in a similar way to structures seen in the intensity maps . to compare the flow structure in a spatially resolved coronal loop , we have marked a chosen loop with an arrow in the velocity maps in top panel of fig .
[ int_vel ] .
plasma downflow is seen at all temperature along the loop .
however , downflows are localized towards the foot point regions in cooler lines such as and ( not shown in figure ) and was as high as ( @xmath0 60 km s@xmath1 ) .
downflows are seen all along the loop in and lines being stronger towards the foot point region . at high temperatures such as log t[k]=6.3 ( ) and log t[k]=6.4 ( )
the downflows are stronger toward the loop top regions with almost no downflow signature at foot points .
we note that these velocities are along the los
. therefore the actual velocities would be much higher than these values . from the intensity and velocity maps
, it appears that the region showing red - shift in cooler spectral lines shows blue - shift in high temperature lines . to show this quantitatively
we have selected two regions labelled region 1 and region 2 ( see top panel of fig .
[ vel_plot ] ) and plotted the average velocities obtained at different lines ( bottom panel of fig .
[ vel_plot ] ) .
these two regions are located at the foot points of the high lying coronal structures .
as it is evident from the plot , in line both the regions show redshifted emission ( downflows ) turning to blueshifted emission ( upflows ) with increasing temperatures .
we note that both the selected regions correspond to footpoint regions of coronal loop .
observation of blueshifted emission in high temperature lines at the base of coronal loops provides evidence of upflows in loops .
it is worth pointing out that if we take into account the results obtained from sumer measurements @xcite , the average downflows in would be weaker by about @xmath0 3 km s@xmath1 and average upflows seen in and would be stronger by about @xmath0 4 and @xmath0 9 km s@xmath1 respectively .
however , we do not have measurements for other higher temperature lines from sumer to make such a comparison .
after the first hinode eclipse , the eis recorded a few very well defined vertical loop structures ( see fig .
[ context ] right panel x=[0:50 ] arcsec ) before going into another eclipse .
the orientations of these loops ( almost north - south ) are such that rastering across the loops does not take a long time ( less than 20 mins ) .
therefore , this loop system offers us an excellent opportunity to study the physical plasma parameters along active region loops at a given time . from now on we will only focus on the region @xmath0
x=[-10:65 ] arcsec and @xmath0 y=[70:220 ] .
note that the eis rasters a region on the sun from west to east .
[ loop_difftemp ] ( top panel ) shows monochromatic images of the sub - region from the eis raster .
the spectral lines for which these images were created are labelled for each frame . as is evident from the figure , the coronal loops in the active region are more sharply defined in the images observed at lower temperature lines such as , and than those in the higher temperature lines such as , , and .
structures in are the sharpest and are fuzziest in .
the structures seen in are similar to those seen in the trace 171 image ( see fig .
[ context ] left panel ) .
however the small difference is due to somewhat better spatial resolution of trace ( 1 ) than that of eis ( 3 - 4 ) .
the active region loop structures are clearly discernible in the line .
the images obtained in the and lines lack clarity while noting that the loops seen in line may have partial overlap with line . based on these observations it is plausible to conclude that the coronal structures are fuzzy at higher temperatures .
this in itself has been a matter of debate amongst the solar physics community for many years ( since skylab days ) .
the question concerns whether the solar corona really is fuzzy at high temperatures or simply appears to be fuzzy because the instrumental spatial resolution and line intensities are insufficient at higher temperatures .
the high temperature corona was observed with the soft x - ray telescope ( sxt ; * ? ? ?
* ) aboard yohkoh , however cooler temperature emission could not be observed with yohkoh .
the soho / cds instrument could observe the corona simultaneously at low temperature ( @xmath0log t[k ] = 5.0 mk ) as well as high temperatures @xmath0 log t[k ] = 6.4 mk .
however , the cds spatial resolution ( @xmath0 5 - 6 ) was not enough to unambiguously resolve this issue . similarly with soho / eit which could observe the sun at low temperature using its channel and at high temperature using the channel but did nt have sufficiently good spatial resolution . to clarify this point , we chose a region ( marked between the two horizontal white lines in the top panels of fig .
[ loop_difftemp ] ) and plotted the corresponding intensity profile ( summed in the y - direction ) in bottom panels of fig .
[ loop_difftemp ] .
it is straightforward to see that the intensity fluctuations due to the presence of spatially resolved coronal features are highly pronounced in low temperature emission , but much less obvious at higher temperatures .
the intensity images in and lines , shows a spike at location x=40 which is basically due to a spatially resolved loop , marked by an arrow as loop b. another active region loop a is also marked by arrows in the image .
for both these loops we can identify enhancements in the intensity profiles for lines such as , and .
however in the image , the intensity corresponding to loop b has almost disappeared in the fuzzy background , but the intensity enhancement corresponding to loop a is still clearly discernible . the intensity fluctuation corresponding to the loop a is still just about discernible in and .
however shows a sharp decrease in intensity corresponding to the position of loop a. it is interesting to note that the intensity profile corresponding to the line shows a dip for both the coronal loops marked a and b. however , it also shows a sharp enhancement ( barely identifiable in lower panel plots ) between these two dips where the intensity profiles corresponding to cooler lines show a dip .
the plot also demonstrate the possible partial overlaps between the loop structures seen in line and those seen in line .
this basically implies that active regions are comprised of coronal loops seen at different temperatures and intermingled with each other ( see also * ? ? ?
one of the most important issues when studying the problem of active region heating and the coronal heating in general is whether the coronal loops are isothermal or multi - thermal along the los .
this issue has been studied extensively using data from trace ( see e.g. , * ? ? ?
* ) , and cds ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the main controversy started when using the spectroscopic techniques such as dem , @xcite showed that the loops were not isothermal either along the loop length or along the los . later on using an observation which was taken by trace and cds simultaneously @xcite and @xcite
( see also * ? ? ?
* ) showed that the loops were isothermal along the los and the proper treatment of the diffuse background is important .
given the excellent spatial resolution of eis and the broad temperature coverage , we have estimated temperature along the los at many given points along the loop length .
for this purpose we have applied the _ em - loci _ method using a set of spectral lines which are almost density insensitive . in the _ em - loci _ method
we plot @xmath22/(a(b ) g(n@xmath14,t@xmath14 ) ) as a function of temperature .
the _ contribution function _
g(n@xmath14 , t@xmath14 ) was calculated with chianti , v.5.2 , using coronal abundance of @xcite with a density of 10@xmath23 @xmath3 and the ionization fraction of @xcite .
similar results were obtained using the ionization fraction of @xcite .
it is evident from fig .
[ loop_difftemp ] that the loops are not sharply defined at all temperatures .
they are best seen at lower temperatures .
based on a visual inspection , therefore , we have chosen two loops namely loop a and loop b ( labelled in fig.[image_datapoints ] ) in and used these as a proxy for other lines where the loops are not sharply defined .
for the background subtraction we chose a loop - shaped structure running parallel to the loop b , shifted slightly to the east and labelled _ bg _ in fig .
[ image_datapoints ] .
we chose this location to avoid the eruptive material which might contaminate the background emission to the west of the active region .
furthermore we have smoothed out the intensity along the loop and background intensity before subtraction to avoid any local fluctuations .
it is worth mentioning here that we selected different regions for background subtraction and the results were similar with a difference of about 5 - 10% .
[ temperature ] displays the em - loci plots for loop a measured at points ( 1- 20 , point 1 being at the foot - point and 20th towards the loop top ) along the loops shown in fig .
[ image_datapoints ] , from south to north .
the em - loci plots clearly demonstrate that the coronal loop studied here is nearly isothermal along the los and the temperature increases along the loop .
close to the foot points of the loop the temperature is @xmath0 0.8 mk and increases to @xmath0 1.5 mk at the projected height of about 75 mm . for the first few points ( towards the loop foot point ) , the em - loci for is not consistent with other lines .
the discrepancy is due a problem with the atomic data for @xcite .
a similar trend is found for loop b. in order to derive electron densities along the loop we have selected several diagnostic line pairs such as @xmath6280 & @xmath6278 , @xmath6258 & @xmath6261 and @xmath6186 & @xmath6195 @xcite .
the electron density values were obtained using the theoretical line intensity ratios calculated using chianti , v5.2 @xcite .
however , we note that the @xmath6278 and @xmath6274 lines are blended with a line and therefore we have used the @xmath6275 line , which is a strong line , to remove the blend . the two lines are self blended and care must be taken while deriving densities using line ratios .
figure [ image_datapoints ] shows the intensity image obtained in @xmath6184 line overplotted with asterisks showing selected pixels along the loops and the background . in order to obtain a valid electron density for a coronal loop ,
it is necessary to subtract the background intensities .
it has been shown that the results can be significantly different if a proper background subtraction is not performed ( see e.g. , * ? ? ?
note that since the intensity along loop b were not significantly different from the background intensities , it was difficult to obtain an accurate background subtracted intensities for the weak diagnostic lines and to derive densities . hence for density and filling factor estimation we have only considered loop a. figure [ dens_plot ] shows the variation of electron density along loop a with the projected height of the loop .
the squares represent the density values derived from @xmath6(280/278 ) ( log t[k ] = 5.8 ) line ratios , asterisks represent those from @xmath6186/195 ( log t[k ] = 6.1 ) , and triangles are for @xmath6258/261 ( log t[k ] = 6.1 ) . for and
lines we did not have enough counts at higher altitudes along the loop and therefore , we could not estimate the density for the higher loop locations .
error bars on densities are estimated by considering a 10% errors in line intensities .
it is immediately evident from fig .
[ dens_plot ] , that the electron densities decrease with altitude along the loop for and .
however for it does not show as sharp a decrease .
the densities obtained from and are within consistent error bar . note that for the results are based on the atomic calculation by @xcite .
the electron densities obtained using are significantly different .
in fact the densities using these lines could be only measured towards the foot - point location . at the projected height of about 20 mm
the densities using become similar to those of and . another important parameter to determine
is the filling factor ( defined by eq .
3 ) for coronal loops .
this is an indicator of whether loop structures are composed of multiple strands or not @xcite .
the most problematic parameter in the filling factor calculation is the thickness of the loop along the along the los . in this study
we have considered the diameter of the loop to be equivalent to the thickness .
the loop s diameter measured using image was used as thickness when estimating the filling factors using and lines .
the loops , however , in transition region lines such as appear to be fatter towards the footpoint .
we have measured the width of the loops using the image obtained in @xmath6278 and used this as the thickness along the los for the filling factor calculation using densities .
table 1 provides from left to right : projected height of the loop , thickness ( diameter ) of the loop along the los , electron densities , filling factors and the path length for three lines such as , , .
the filling factors for densities is about 0.02 at the footpoint and about 0.8 at a projected height of 40 mm . for lines ,
the filling factors were similar to those of lines but become more than 1 at a projected height of 25 mm .
we believe that this is because of the less reliable densities obtained using lines ratios .
we find substantially different results for with a filling factor close to 1 .
for , the filling factor decreases with altitude giving reliable values only up to around 15 mm .
the filling factor values using lines were close to 1 similar to the soho / cds measurements using lines ( see * ? ? ?
* ; * ? ? ?
the eis on board hinode provides an excellent opportunity to study the physical plasma properties over a range of temperatures simultaneously in spatially resolved coronal structures .
this is extremely important when studying the problem of active region heating . in this paper
we have studied the overall temperature structure and flows in an active region and in a particular spatially resolved coronal loop .
in addition we have also derived densities and filling factors along this loop .
the main results are summarized below . *
the high - lying loop structures are seen more distinctly in the images obtained using lower temperature lines . with increasing temperature
the core of the active region becomes more apparent and overall structure becomes fuzzier .
the active region is comprised throughout of upflows and downflows . the downflows are predominantly seen in the core of the active region .
however the upflows are seen at the boundary of the active region in the low emission regions . with increasing temperature the regions showing red shifted emission in low temperature at the boundary of active region turns towards blue shifted emission .
( see figs .
[ int_vel ] , [ vel_plot ] ) * in a well resolved coronal loop we observed downflows ( red shifts ) along the loop at all temperatures . the downflows are seen only towards the footpoint in and are very strong @xmath060 km s@xmath1 . however , in and downflows are seen all along the loop being stronger towards the footpoint . in high temperature lines such as and
the downflows are localized towards the loop top ( see fig . [ int_vel ] ) .
we note that the velocities measured at quiet sun vary with temperature @xcite . taking account of this variation would not substantially affect our results . *
the active region structures are more clearly defined in lower temperature lines and appear fuzzy at higher temperature .
this is consistent with indications from previous solar observations .
however , confirmation of this required spectroscopic observations with the same instrument over a range of temperatures , with good spatial resolution , as provided by the eis ( see fig . [ loop_difftemp ] ) . * using the em - loci method we find that that both loops studied here are almost isothermal along the los . the temperature at the footpoint is about 0.80 mk and increases to about 1.5 mk at a projected height of about 75 mm ( see fig . [ temperature ] ) .
* the derived electron densities decrease with loop altitude . towards the footpoint ,
the electron density estimated using ratios is log n@xmath24=9.8 @xmath3 which decreases to log n@xmath24=8.5 @xmath3 at an altitude of about 75 mm . the densities obtained using were similar to densities .
however , the estimated densities using line ratios were smaller than those obtained using towards the footpoint locations and reached a value similar to that of at a projected height of about 25 mm ( see fig . [ dens_plot ] ) . * using the densities obtained by , and line ratios and the loop s diameter as the thickness along the los , we find that the filling factor increases with projected height .
the filling factor obtained using densities is about 0.02 at the footpoint and about 0.8 at a projected height of 40 mm . for lines ,
the filling factors were similar to those of lines but become more than 1 at a projected height of 25 mm .
we find substantially different results for ( see table 1 ) , closer to unity .
an understanding of the problem the coronal heating is a long lasting one and it is now widely believed that studying the physics of all kinds of loops holds the key .
there have been many recent theoretical developments to explain the heating in coronal loops and in active regions .
however , the observational confirmations of these theories have not been possible unambiguously for various reasons . for example the nanoflare model originally proposed by @xcite and latter on globally modelled by @xcite ( see also * ? ? ?
* ) suggests that the coronal loops are comprised of many sub - resolution strands and they are all being heating separately .
this predicts that the loops are multi - thermal along the los .
in addition , this model also predicts high velocity evaporative upflows at high temperatures .
using the idea of multi - stranded coronal loops in a 1d hydrodynamic simulation , @xcite suggested that these high velocity upflows can been seen as non - thermal broadening with a particular enhancement in blue wing in high - temperature spectral lines .
observational confirmation of these predictions can only be unambiguously performed using spectroscopic observations with high spectral and spatial resolution with a broad temperature coverage .
thanks to the excellent simultaneous spectral and spatial resolution of eis , we are now able to study flows and other physical parameters in an active region and also in spatially resolved coronal structures in detail .
observations of strong downflows at the foot point in cooler transition region temperatures are similar to previous observations using sumer measurements ( see e.g. , * ? ? ?
* ; * ? ? ?
in addition to that we also observed downflows only towards the loop top in higher temperature lines such that and .
the observation of downflows at all temperatures raises the question concerning the origin of the plasma in the loop and suggests that there must be some sort of upflow of plasma filling up the loops .
@xcite performed a forward modelling of emission line for eis using data from a one dimensional hydrodynamic simulations and suggested that the signatures of high speed upflows can be seen as non - thermal broadening and blue - wing enhancement in the line observed by the eis .
unfortunately , the study sequence used in the present analysis did not contain that line with sufficient counts preventing us to perform a direct comparison .
recently @xcite studied line profiles for at foot point of coronal loops .
they observed a slight enhancement in the blue wing suggesting that this could be associated with high velocity upflows at higher temperature as suggested by @xcite .
similar to our observations , @xcite also did not observe any signature of downflows towards the foot point of the loops in high temperature lines and suggested that the absence of downflows towards the footpoint could be attributed to the presence of high velocity upflows .
moreover , in the present study the observations of average flow structures turning from downflows to upflows in regions located at the foot points of the coronal loops provides another argument in favor of upflows in coronal loops at higher temperatures . in a recent study using the idea of nano - flare foot - point heating ,
@xcite predict upflows of plasma in line which seem to match with our observations .
the idea of multi - stranded impulsively heated coronal loop suggests that the filling factor values should be less than unity .
this is precisely what our measurements show . in our measurements
we find that the filling factors are much less than unity at log t[mk ] = 6.1 and is closer to unity at transition region temperature ( log t[mk ] = 5.8 ) towards the foot point . however , with increasing projected height , the filling factors at log t[mk ] = 6.1 increases while that for log t[mk]=5.8 decreases .
this means that towards the foot points of the loop most of the plasma is at transition region temperature .
however , with increasing projected height , the amount of plasma at log t[mk ] = 6.1 increases .
this suggests an increasing temperature with projected height of the loop .
this is exactly what we obtained using the em - loci analysis .
the em - loci analysis ( see fig . [ temperature ] ) shows that the temperature along the loop increases with the projected height of the loop being @xmath00.8 mk at the foot point and @xmath01.5 mk at a projected height of about 75 mm .
the em - loci analysis also shows that the plasma along the los is close to isothermal .
however , the different flow structures and densities obtained using different spectral lines and filling factors values less than unity emphasize the multi - thermality of the plasma along the los .
it is not trivial to reconcile these two results together .
we note that the multi - thermality of plasma along the los is one of the most important ingredient in the model of multi - stranded impulsively heated coronal loop .
however , so far we do not know how multi - thermal or how isothermal the plasma is observationally and also theoretically . in this analysis ,
e.g. , at least for the first few point from the foot point of the loop , the loops does not appear to be exactly isothermal .
there appears a narrow distribution of temperatures .
based on observations of a number of loops using eis data , @xcite showed that the coronal loops which are bright at have a narrow distribution of temperature but they are not isothermal .
the observations of high speed downflows and concurrent upflows in coronal loops in addition to very low values of filling factors and their nearly - isothermal structure along the los provide very important constraints in the modelling of coronal loops .
given the capabilities of eis , more measurements should be performed for many other coronal loops to obtain more statistically accurate values of these parameters .
we thank the referee for constructive comments .
dt , hem and gdz acknowledge stfc .
bnd aclnowledges hem for support through a rolling grant from stfc at damtp .
hinode is a japanese mission developed and launched by isas / jaxa , collaborating with naoj as a domestic partner , nasa and stfc ( uk ) as international partners .
scientific operation of the hinode mission is conducted by the hinode science team organized at isas / jaxa .
this team mainly consists of scientists from institutes in the partner countries .
support for the post - launch operation is provided by jaxa and naoj ( japan ) , stfc ( u.k . ) , nasa , esa , and nsc ( norway ) 0 & 7 & 9.83 & 0.02 & 0.18 & 10.00 & 0.00 & 0.02 & 20 & 9.14 & 1.79 & 37.69 + 5 & 8 & 9.77 & 0.02 & 0.22 & 9.91 & 0.01 & 0.10 & 17 & 9.13 & 1.71 & 30.40 + 9 & 6 & 9.63 & 0.06 & 0.36 & 9.70 & 0.06 & 0.44 & 15 & 9.10 & 1.48 & 23.33 + 14 & 10 & 9.48 & 0.07 & 0.77 & 9.49 & 0.14 & 1.47 & 15 & 9.07 & 0.84 & 13.16 + 18 & 10 & 9.32 & 0.15 & 1.59 & 9.35 & 0.26 & 2.78 & 15 & 9.02 & 0.56 & 8.78 + 22 & 13 & 9.22 & 0.20 & 2.69 & 9.18 & 0.43 & 5.93 & 10 & 9.07 & 0.28 & 2.91 + 26 & 12 & 9.16 & 0.29 & 3.71 & 9.02 & 1.03 & 12.99 & & & & + 30 & 11 & 9.13 & 0.37 & 4.32 & 8.88 & 2.11 & 24.41 & & & & + 34 & 11 & 9.04 & 0.47 & 5.48 & 8.79 & 2.86 & 33.08 & & & & + 38 & 10 & 8.90 & 0.79 & 8.30 & 8.72 & 3.25 & 34.17 & & & & + 42 & 11 & 8.73 & 1.30 & 15.06 & 8.60 & 4.14 & 47.81 & & & & + 46 & 10 & 8.58 & 2.63 & 27.65 & 8.42 & 7.40 & 77.62 & & & & + 49 & 7 & 8.53 & 4.52 & 33.19 & 8.22 & 18.73 & 137.54 & & & & + 53 & 12 & 8.52 & 3.11 & 39.26 & & & & & & & + 56 & 10 & 8.55 & 3.66 & 38.38 & & & & & & & + 60 & 10 & 8.52 & 4.21 & 44.12 & & & & & & & + 63 & 9 & 8.52 & 4.57 & 43.12 & & & & & & & + 66 & 11 & 8.50 & 4.41 & 50.88 & & & & & & & + 70 & 12 & 8.51 & 3.80 & 47.87 & & & & & & & + 73 & 12 & 8.50 & 3.89 & 48.94 & & & & & & & + 184 .
overplotted asterisks show two loops and the background .
the right most is loop a and the middle is loop b. the left most one , labelled as bg , is used for background subtraction .
[ image_datapoints],scaledwidth=60.0% ] | we have carried out a study of active region loops using observations from the extreme - ultraviolet imaging spectrometer ( eis ) on board hinode using 1 raster data for an active region observed on may 19 , 2007 .
we find that active region structures which are clearly discernible in cooler lines ( @xmath0 1mk ) become fuzzy at higher temperatures ( @xmath0 2mk ) .
the active region was comprised of red - shifted emissions ( downflows ) in the core and blue - shifted emissions ( upflows ) at the boundary .
the flow velocities estimated in two regions located near the foot points of coronal loop showed red - shifted emission at transition region temperature and blue shifted emission at coronal temperature .
the upflow speed in these regions increased with temperature . for more detailed study we selected one particular
well defined loop .
downward flows are detected along the coronal loop , being stronger in lower temperature lines ( rising up to 60 km s@xmath1 near the foot point ) .
the downflow was localized towards the footpoint in transition region lines ( ) and towards the loop top in high temperature line ( ) . by carefully accounting for the background emission we found that the loop structure was close to isothermal for each position along the loop , with the temperature rising from around 0.8 mk to 1.5 mk from the close to the base to higher up towards the apex ( @xmath0 75 mm ) .
we derived electron density using well established line ratio diagnostic techniques .
electron densities along the active region loop were found to vary from 10 @xmath2@xmath3 close to the footpoint to 10 @xmath4@xmath3 higher up .
a lower electron density , varying from 10 @xmath5@xmath3 close to the footpoint to 10 @xmath4@xmath3 higher up , was found for the lower temperature density diagnostic . using these densities we derived filling factors in along the coronal loop which can be as low as 0.02 near the base of the loop .
the filling factor increased with projected height of the loop .
these results provide important constraints on coronal loop modeling . |
the work was supported by the national natural science foundation of china under grant no .
30600122 and guangdong provincial natural science foundation under grant no . 06025073 . 0.4 in r. d. astumian and p. hanggi , physics today 55 , 33(2002 ) .
p. hanggi , f. marchesoni , and f. nori , ann .
( leipzip ) , 14 , 51 ( 2005 ) .
p. reimann , phys .
361 , 57 ( 2002 ) .
m. o. magnasco , phys .
71 , 1477 ( 1993 ) .
p. hanggi and r. bartussek , nonlinear physics of complex system - current status and future trends , 476 , spring , berlin , ( 1996 ) , 294 ; p. reimann , phys .
290 , 149(1997 ) ; j. d. bao and y. z. zhuo , phys .
lett . a 239 , 228(1998 ) ; b. q. ai , l. q. wang , and l. g. liu , chaos , solitons and fractals 34 , 1256(2007 ) .
p. reimann , r. bartussek , r. haussler , and p. hanggi , phys .
lett . a 215 , 26 ( 1994 ) . c. r. doering , w. horsthemke and j. riordan , phys .
72 , 2984 ( 1994 ) .
j. luczka , r. bartussek , and p. hanggi , europhys .
31 , 431 ( 1995 ) .
d. reguera , g. schmid , p. s. burada , j. m. rubi , p. reimann , and p. hanggi , phys .
lett 96 , 130603 ( 2006 ) ; p. s. burada , g. schmid , d. reguera , j. m. rubi , and p. hanggi , phys . rev .
e 75 , 051111 ( 2007 ) .
b. q. ai and l. g. liu , phys .
e 74 , 051114(2006 ) , b. q. ai , h. z. xie , and l. g. liu , phys .
e 75 , 061126(2007 ) ; b. q. ai and l. g. liu , j. chem . phys .
126 , 204706(2007 ) .
p. jung , j. g. kissner , and p. hanggi , phys .
76 , 3436 ( 1996 ) ; m. barbi and m. salerno , phys .
e 62 , 1988 ( 2000 ) ; j. l. mateos , phys .
84 258 ( 2000 ) ; l. machura , m. kostur , p. talkner , j. luczka , f. marchesoni , and p. hanggi , phys . rev .
e 70,061105 ( 2004 ) ; k. lindenberg , j. m. sancho , a. m. lacasta , and i. m. sokolov , phys .
98 , 020602 ( 2007 ) .
l. machura , m. kostur , p. talkner , j. luczka , and p. hanggi , phys .
98 , 040601 ( 2007 ) .
a. haljas , r. mankin , a. sauga , and e. reiter , phys .
e 70 , 041107 ( 2004 ) .
m. kostur , l. machura , p. hanggi , j. luczka , and p. talkner , physica a 371 , 20(2006 ) ; l. machura , m. kostur , f. marchesoni , p. talkner , p. hanggi , and j. luczka , j. phys . -
condensed matter 17 , s3741(2005 ) .
r. mankin , a. ainsaar , and e. reiter , phys .
e 61 , 6359 ( 2000 ) . | based on the recent work [ phys .
rev . lett .
98 , 040601(2007 ) ] , we extend the study of inertial brownian motors to case of an asymmetric potential .
it is found that some novel transport phenomena appear in the presence of an asymmetric potential . within tailored parameter regimes , there exists two optimal values of the load at which the mean velocity takes its maximum , which means that a load can facilitate the transport in the two parameter regimes .
in addition , the phenomenon of multiple current reversals can be observed when the load is increased .
recently , brownian motors have attracted considerable attention simulated by research on molecular motors .
these systems can be modeled , for instance , by considering brownian particles in a periodic asymmetric potential and acted upon by an external time - dependent force of zero average @xcite .
typical examples are rocking ratchets @xcite , flashing ratchets @xcite , diffusion ratchets @xcite , correlation ratchets @xcite , white - shot - noise ratchets @xcite , and entropic ratchets @xcite .
most studies have referred to consideration of the overdamped case in which inertial term due to the finite mass of the particle is neglected . however ,
there are some pervious works on inertial brownian motors and some novel transport phenomena were observed @xcite . in particular , machura and
co - workers @xcite studied the transport of inertial brownian particles moving in a symmetric periodic potential under the influence of both a time periodic and a constant , biasing driving force .
they found that thermal equilibrium fluctuations can induce the phenomenon of absolute negative mobility .
absolute negative mobility is the rather surprising opposite behavior in the form of a permanent motion against a ( not too large ) static force of whatever direction .
haljas and co - workers @xcite had found that an interplay of three - level colored and thermal noises can also generate the phenomenon of absolute negative mobility . in this brief report
, we extend the previous work @xcite to case of an asymmetric potential . we emphasize on finding whether a load can facilitate the transport of inertial brownian particles .
let us consider the one - dimensional of inertial brownian particles driving by a time - dependent external force and a load , under the influence of an asymmetric periodic potential .
the particle is driven by an unbiased time - periodic monochromatic force of strength @xmath0 and an angular frequency @xmath1 .
the equation of motion reads as @xcite @xmath2 where the dot and prime denote derivatives with respect to @xmath3 and @xmath4 respectively .
the parameter @xmath5 denotes the friction coefficient , @xmath6 is temperature , and @xmath7 is boltzmann constant .
@xmath8 is the external asymmetric periodic potential and has the period @xmath9 and a barrier height @xmath10 .
the quantity @xmath11 denotes the external , constant force .
the thermal fluctuations due to the coupling of the particle with the environment are modeled by a zero - mean , gaussian white noise @xmath12 with autocorrelation function @xmath13 satisfying einstein s fluctuation - dissipation relation . upon introducing characteristic length scale and time scale @xcite ,
eq .
( 1 ) can be rewritten in dimensionless form @xmath14 with @xmath15 and the zero - mean white noise @xmath16 has auto - correlation function @xmath17 . in the following , mostly for the sake of simplicity
, we shall only use dimensionless variables and shall omit the `` hat '' notion in all quantities .
in addition , our emphasis is on finding whether a load can facilitate the transport , so we take a new quantity @xmath18 . for the asymmetric ratchet potential @xmath8 ,
we choose @xmath19 where @xmath20 is the asymmetry parameter of the potential .
we restrict the discussion here to a set of optimal driving parameters @xcite , reading @xmath21 , @xmath22 , and @xmath23 .
our emphasis is on finding the asymptotic mean velocity which is defined as the average of the velocity over the time and thermal fluctuations .
the fokker - planck equation corresponding to eq .
( 2 ) can not be analytically solved , therefore , we carried out extensive numerical simulations @xcite .
we have numerically integrated eq .
( 2 ) by the stochastic runge kutta method of the second order with time step @xmath24 .
the initial condition of @xmath25 is taken from a uniform distribution over the dimensionless periodic @xmath26 of the ratchet potential and the initial condition of @xmath27 is chosen at random from a symmetric , uniform distribution over the interval @xmath28 $ ] .
the data obtained were average over 500 different trajectories and each trajectory evolved over @xmath29 periods .
the numerical results were shown in figures 1 and 2 .
vs the external force @xmath30 for different values of the asymmetry parameter @xmath20 at @xmath21 , @xmath22 , @xmath23 , and @xmath31 . ,
width=377,height=302 ] figure 1 shows the asymptotic mean velocity @xmath32 as a function of the external force @xmath30 for different values of the asymmetry parameter @xmath20 .
when the potential is symmetric ( @xmath33 ) , there are two peaks in the curve , one above the zero velocity line and the other below the zero velocity line .
the peak above the zero velocity line ( the left peak ) is already investigated by machura and co - workers @xcite , which shows the phenomenon of absolute negative mobility .
the peak below zero velocity line ( the right peak ) is induced by the high - frequency driving ( the time - dependent periodic external force ) ratchet effect .
when the asymmetry parameter is increased , the position of the right peak moves to small value of the load @xmath30 and its height increases .
the two peaks may be above the zero velocity line ( @xmath34 ) , which shows that a load can facilitate the transport in the two different regimes .
however , for large value of the asymmetry parameter ( @xmath35 ) , the two peaks disappear and the velocity decreases monotonically with increasing the load @xmath30 .
it is obvious that in this case both the absolute negative mobility effect and the high - frequency driving ratchet effect disappear and the transport is dominated by the noise driving ratchet effect . in addition , we also find that multiple current reversals occur as the load is increased .
the brownian particle changes its direction three times for @xmath36 .
it must be pointed out that multiple current reversals can be detected in overdamped brownian systems driven by the colored three - level markovian noises @xcite .
vs the external force @xmath30 for different values of @xmath37 at @xmath21 , @xmath22 , @xmath23 , and @xmath38.,width=377,height=302 ] in figure 2 , we depict the asymptotic mean velocity as a function of the external force @xmath30 for an asymmetric potential ( @xmath38 ) . as the noise intensity
is increased , the high - frequency driving ratchet effect decreases and the noise driving ratchet effect increases .
thus , the height of the right peak decreases . for very large values of @xmath37 , both the absolute negative mobility effect and
the high - frequency driving ratchet effect disappears and the two peaks disappear .
in conclusion , in this brief report we study the transport of inertial browinan particles moving in an asymmetric potential . within tailored parameter regimes given in ref .
12 , we extend their work to the case of an asymmetric potential .
it is found that there are two optimal values of the load , at which the mean velocity takes its maximum , which indicates that the load facilitate the transport in the two different regimes .
the first optimal value ( the left peak ) is from absolute negative mobility @xcite and the second one is induced by the high - frequency driving ratchet effect . when the noise intensity @xmath37 or the asymmetry parameter @xmath20 is very large ,
the two optimal values disappear .
in addition , when the load is increased , multiple current reversals may occur within tailored parameter regimes .
for example , for @xmath36 , the particle can change its direction three times as the load is increased . |
most of the multiparticle production is currently described in terms of color strings stretched between the projectile and target , which decay into new strings and subsequently hadronize to produce observed hadrons .
color strings may be viewed as small areas in the transverse space , @xmath1 , @xmath2=0.2 fm , filled with color field created by the colliding partons .
particles are produced via emission of @xmath3 pairs in this field @xcite .
with growing energy and/or atomic number , the number of strings grows , and they start to overlap , forming clusters , very much similar to disks in the two dimensional percolation theory @xcite . at a certain critical density , @xmath4
a macroscopic cluster appears that marks the percolation phase transition . in ( [ ec1 ] )
@xmath5 is the number of strings and @xmath6 is the total transverse area of the scattering of the colliding nuclei .
the percolation theory governs the geometrical pattern of the string clustering .
its observable implications , however , require introduction of some dynamics to describe string interaction , i.e. , the behavior of a cluster formed by several overlapping strings .
there are several possibilities @xcite . here , we assume that a cluster behaves as a single string with a higher color field @xmath7 corresponding to the vectorial sum of the color charge of each individual @xmath8 string . the resulting color field covers the area @xmath9 of the cluster . as @xmath10 , and
the individual string colors may be oriented in an arbitrary manner respective to one another , the average @xmath11 is zero and @xmath12 . knowing this charge color @xmath7
, one can compute the particle spectra produced by a single color string of area @xmath9 using the schwinger formula @xcite .
for the multiplicity @xmath13 and the average @xmath14 of particles @xmath15 produced by a cluster of @xmath16 strings one finds @xmath17 where @xmath18 and @xmath19 are the mean multiplicity and @xmath14 of particles produced by a simple string with area @xmath20 . for strings , just touching each other @xmath21 and hence @xmath22 , @xmath23 , as expected . in the opposite case of maximum overlapping @xmath24 and therefore @xmath25 ,
@xmath26 equation ( [ ec2 ] ) is the main tool of our calculation @xcite . in order to compute the multiplicities
we generate strings according the monte carlo code of @xcite .
each string is produced at an identified impact parameter . from this , knowing the transverse area of each string , we identify all the clusters formed in each collision and subsequently compute for each of them its multiplicity in units @xmath18 .
the value of @xmath18 was fixed by normalizing our results to the sps wa98 results for central pb - pb collisions @xcite . in color strings models
both @xmath18 and @xmath27 are assumed to rise with energy due to increase of the rapidity interval available and to the hard contribution .
however this rise is very weak in the range @xmath28=17 - 200 gev and we can neglect it , in the first approximation .
the comparison of our results for the dependence of the multiplicity on the number of participants with the sps wa98 data , the rhic phenix @xcite and phobos @xcite data at @xmath28=130 gev and @xmath28=200 gev , is presented in fig .
[ figure1].a , where also is presented our prediction for lhc .
= 130 gev / n , and rhic phenix data at @xmath28=200 gev / n ( filled star ) .
the upper solid line represents our prediction for pb - pb collisions at 5500 gev / n .
b ) comparison of our predictions of ratios of invariant multiplicities of neutral pions for pb - pb collisions of different centralities to minimum bias distributions as a function of @xmath29 and sps wa98 data .
meaning of centrality classes and grey bands can be found en reference @xcite . ] in order to compute the transverse momentum distribution we make use of the parametrization of the @xmath30 data at 130 gev @xmath31 where @xmath32,@xmath33 and @xmath34 are parameters fitted to data .
the standardly expected distribution for au - au collision is @xmath35 ) we have @xmath36 this implies that the same parametrization ( [ ec3 ] ) can be used for nucleus - nucleus collisions with the only change @xmath37 in fig .
[ figure1].b we show the comparison of our results for eight different centrality events in pb - pb collisions with wa98 data @xcite . in fig .
[ figure2].a we show the distribution ( [ ec3 ] ) for charged particles for central ( 5% ) au - au collisions compared to the phenix data @xcite .
also the distribution expected from the independent string picture , eq .
( [ ec4 ] ) , is shown .
a very good agreement is obtained .
we also obtain a good agreement for peripheral collisions , see reference @xcite . in fig .
[ figure2].b we show our prediction for central au - au collisions at @xmath28=5500 gev compared with the formula ( [ ec4 ] ) . in this case , contrary to rhic energy , a crossing at low @xmath0 is predicted .
distribution using the percolation of strings model ( solid line ) for 5% central au - au and pb - pb collisions at 130 gev / n and 5500 gev / n respectively .
also it is shown the expected distribution from the independent string picture ( dotted line ) . ]
we observe that for a cluster , @xmath38 , which at high density translate into @xmath39 the formula ( [ ec7 ] ) shows an universal relation between multiplicities and @xmath27 .
the scaling laws expressed by the formula ( [ ec3 ] ) with the change ( [ ec6 ] ) and by the formula ( [ ec7 ] ) are very similar to other scaling laws using saturation of gluons . to conclude , a general comment .
we obtain a good description of the multiplicities and transverse momentum distribution at different energies and different centralities including a low and high @xmath0 suppression .
however we can not reproduce the additional high @xmath0 suppression reported at this conference by several collaborations .
if the explanation is jet quenching , our approach can provide an unified framework to add the phenomenology of jet quenching .
this work was done under contract fpa2002 - 01161 of ministerio de ciencia y tecnologia of spain .
9 n. s. amelin , m. a. braun and c. pajares , z. phys .
c 63 ( 1994 ) 507 ; n. s. amelin , n. armesto , c. pajares and d. sousa , eur .
j. c 22 ( 2001 ) 149 ; n. armesto , c. pajares and d. sousa , phys . lett .
b 527 ( 2002 ) 92 .
n. armesto , m. a. braun , e. g. ferreiro and c. pajares , phys .
77 ( 1996 ) 3736 ; m. nardi and h. satz , phys .
b 442 ( 1998 ) 14 .
m. a. braun , c. pajares and j. ranft , j. mod .
( 1998 ) 2689 ; m. a. braun and c. pajares , eur .
j. c 16 ( 2000 ) 349 .
m. a. braun and c. pajares , phys .
85 ( 2001 ) 4864 ; m. a. braun , f. del moral and c. pajares , eur .
j. c 21 ( 2001 ) 557 .
j. schwinger , phys .
82 ( 1951 ) 664
. m. a. braun , f. del moral and c. pajares , phys .
c 65 ( 2002 ) 024907 .
m. m. aggarwal et al .
, wa98 collaboration , eur .
j. c 18 ( 2001 ) 651 .
k. adcox et al . ,
phenix collaboration , phys .
86 ( 2001 ) 3500 .
b. b. bock et al . ,
phobos collaboration , phys .
85 ( 2000 ) 3100 .
k. adcox et al . ,
phenix collaboration , phys .
88 ( 2002 ) 022301 .
m. m. aggarwal et al . , wa98 collaboration , eur .
j. c 23 ( 2002 ) 225 . | the dependence of the multiplicity and the transverse momentum distribution on the number of collisions are studied for central and peripheral au - au collisions at sps , rhic and lhc energies in the framework of percolation of strings .
a scaling law relating the multiplicity to the mean transverse momentum is obtained .
our results are in overall agreement with the sps and rhic data , obtaining a suppression on @xmath0 distribution even for @xmath0 larger than 1 gev / c . |
this paper presents a new approach to the constructive design of a robust nonlinear dynamic state feedback controller using an integral quadratic constraint ( iqc ) approach .
the idea of using a copy of the plant nonlinearity in the controller is used previously in the literature @xcite . however in this paper , we apply a new methodology to construct a controller which uses linear state feedback guaranteed cost control and copies of the plant nonlinearities to form a dynamic state feedback robust nonlinear controller .
this approach provides robust performance in the case where uncertainties and nonlinearities are present in the plant .
consider a class of uncertain nonlinear systems described by the following state equations : @xmath0 where @xmath1 is the state , @xmath2 is the control input , @xmath3 , @xmath4 , @xmath5 , @xmath6 are the uncertainty outputs , @xmath7 , @xmath8 , @xmath5 , @xmath9 are the uncertainty inputs , @xmath10 are the nonlinearity outputs , @xmath11 are the nonlinearity inputs . also , @xmath12 the nonlinearity inputs and outputs are related as follows : @xmath13 where the nonlinear functions @xmath14 satisfy the following generalized monotonicity conditions : @xmath15n_i\left [ \begin{array}{c } \psi_i(\nu_1)-\psi_i(\nu_2)\\ \nu_1-\nu_2 \end{array}\right]\geq0\ ] ] for all @xmath16 , @xmath17 and @xmath18 .
also , @xmath19 are given symmetric matrices representing the monotonicity or global lipschitz conditions ; see @xcite .
furthermore , we assume that @xmath20 and the uncertainty in the system satisfies the following integral quadratic constraints , ( see @xcite ) : @xmath21 m_{1,j } \left [ \begin{array}{c } \xi_{1,j}(t)\\ \zeta_{1,j}(t ) \end{array}\right ] dt + x(0)^t s_{1,j } x(0)\geq 0,\ ] ] for all @xmath22 . here the @xmath23 are given symmetric matrices and the @xmath24 are given positive definite matrices .
let us define @xmath25;\quad \tilde{u}(t){\stackrel{\triangle}{=}}\left [ \begin{array}{c } u(t ) \\
\tilde{\nu}_1(t)\\ \vdots\\ \tilde{\nu}_g(t)\\ \bar{z_1}\\ \vdots\\ \bar{z_g } \end{array}\right].\ ] ] hence the class of controllers considered here are nonlinear dynamic state feedback controllers which contain a copy of the plant nonlinearities ( see fig .
[ fig : nonlinear ] ) ; @xmath26 where @xmath27 and @xmath28 also @xmath29 is a controller gain matrix .
the following iqcs , which follow from ( [ eqmon1 ] ) , are satisfied : @xmath30 n_i\left [ \begin{array}{c } { \mu}_i-\tilde{\mu}_i\\ \nu_i-\tilde{\nu}_i \end{array}\right]dt\nonumber \\ & + x(0)'\breve{s}_{i,1}x(0)\big\}\geq0;\\ \label{eqiqc3 } e&\big\{\int_0^t[\mu_i \quad \nu_i ] n_i\left [ \begin{array}{c } { \mu}_i\\ \nu_i \end{array}\right]dt+x(0)'\breve{s}_{i,2}x(0)\big\}\geq0;\\ \label{eqiqc4 } e&\big\{\int_0^t[\tilde{\mu}_i \quad \tilde{\nu}_i ] n_i\left [ \begin{array}{c } \tilde{\mu}_i\\ \tilde{\nu}_i \end{array}\right]dt+x(0)'\breve{s}_{i,3}x(0)\big\}\geq0;\end{aligned}\ ] ] for all @xmath31 . here
the @xmath32 , @xmath33 , @xmath34 are any positive definite matrices .
now , we first move the controller nonlinearities ( [ eqnon2 ] ) into the plant description and introduce new notation as follows : @xmath35;\quad \tilde{b}_{2 } { \stackrel{\triangle}{=}}\left [ \begin{array}{ccc } \bar{b}_2 & 0 & 0\\ 0 & 0 & i \end{array}\right];\\ & \tilde{\xi}_{2,i}=\left [ \begin{array}{c } \mu_i \\ \tilde{\mu}_i \end{array}\right];\quad \tilde{\zeta}_{2,i}=\left [ \begin{array}{c } \nu_i \\ \tilde{\nu}_i \end{array}\right];\quad \tilde{c}_{1,j}=\left [ \begin{array}{ccc } \check{c}_{1,j } & 0 \end{array}\right];\\ & \tilde{b}_{1,j } { \stackrel{\triangle}{=}}\left [ \begin{array}{c } \check{b}_{1,j } \\ 0 \end{array}\right];\quad \tilde{c}_{2,i}=\left [ \begin{array}{ccc } \bar{c}_{1,i } & 0 \\ 0 & 0 \end{array}\right];\\ & \tilde{d}_{2,i}=\left [ \begin{array}{ccc } \bar{d}_{1,i } & 0 \\ 0 & 0 \end{array}\right];\quad \tilde{d}_{1,j}=\left [ \begin{array}{c } \check{d}_{1,j } \\ 0 \end{array}\right ] , \tilde{a}=\left [ \begin{array}{cc } a & 0 \\ 0 & 0 \end{array}\right ] , \end{split}\ ] ] for all @xmath36 and @xmath37 . using the above notation and ( [ eqnx ] ) , a new system can be written as follows : @xmath38 @xmath39 and @xmath37 .
also , the iqcs ( [ eqiqc2])([eqiqc4 ] ) for the nonlinear uncertainty terms can be written as follows : @xmath40 \tilde{m}_{i , p } \left [ \begin{array}{c } \tilde{\xi}_{2,i}\\ \tilde{\zeta}_{2,i } \end{array}\right ] dt + x(0)^t \tilde{s}_{i , p } x(0)\geq 0,\ ] ] for @xmath41 @xmath42 , where @xmath43 and @xmath44 are positive definite matrices .
we consider the following cost functional associated with the system ( [ eqsystemc1 ] ) : @xmath45 where @xmath46 and @xmath47 are positive - definite symmetric matrices .
[ obs1 ] it is observed that the nonlinearities ( [ eqnon1 ] ) and ( [ eqnon2 ] ) satisfy the iqcs ( [ eqiqc2])([eqiqc4 ] ) .
hence , it follows that if the linear uncertain system ( [ eqsystemc1 ] ) and ( [ eqiqc ] ) , with the linear controller ( [ eqnestimator ] ) , leads to an upper bound on the cost function ( [ eqnfcost ] ) then the same controller ( [ eqnestimator ] ) will yield the same upper bound for the uncertain system ( [ eqsystem ] ) , ( [ eqnon1 ] ) , ( [ eqiqc1 ] ) and ( [ eqnon2 ] ) .
furthermore , it follows from the above discussion that the system ( [ eqsystem ] ) , ( [ eqnon1 ] ) , ( [ eqiqc1 ] ) with controller ( [ eqnestimator ] ) and ( [ eqnon2 ] ) will also lead to the same upper bound on the cost .
we first write the iqc ( [ eqiqc ] ) in the following form which is parametrized by a set of multipliers @xmath48 for @xmath49 : @xmath50 \tilde{m}_{i}(\lambda_i ) \left [ \begin{array}{c } \tilde{\xi}_{2,i}\\ \tilde{\zeta}_{2,i } \end{array}\right ] dt + x(0)^t \tilde{s}_{i , p}(\lambda_i ) x(0)\geq 0,\ ] ] where @xmath51 , @xmath52
. furthermore , @xmath53 $ ] where all @xmath54 .
note that we only consider those @xmath55 which satisfy the following condition @xmath56 where @xmath57 represents the number of negative eigenvalues of @xmath58 and @xmath59 is the matrix of eigenvectors corresponding to the negative eigenvalues .
if condition ( [ eqcondq ] ) is satisfied , then there exists a matrix @xmath60 such that the matrix @xmath61 is a diagonal matrix @xmath62.\ ] ] we define @xmath63 where @xmath64 $ ] , @xmath65 $ ] and @xmath66 $ ] . here
@xmath67 are the negative eigenvalues and @xmath68 are the positive eigenvalues of the matrix @xmath69 . now a change in variables is introduced as follows : @xmath70=t_i \left [ \begin{array}{c } \bar{\xi}_{2,i}(t)\\ \bar{\zeta}_{2,i}(t ) \end{array}\right];\ ] ] @xmath71=t_i^{-1}\left [ \begin{array}{c } \tilde{\xi}_{2,i}(t)\\ \tilde{\zeta}_{2,i}(t ) \end{array}\right]=\left [ \begin{array}{cc } \tilde{t}_{11 } & \tilde{t}_{12}\\ \tilde{t}_{21 } & \tilde{t}_{22 } \end{array}\right ] \left [ \begin{array}{c } \tilde{\xi}_{2,i}(t)\\ \tilde{\zeta}_{2,i}(t ) \end{array}\right].\ ] ] the iqcs ( [ eqiqcm ] ) for a given @xmath72 can now be modified by incorporating new variables as given below ( from now on we remove the argument ( @xmath73 ) from the equations wherever possible for the sake of brevity ) : @xmath74\left [ \begin{array}{cc } -\textbf{i } & 0\\ 0 & \textbf{i } \end{array}\right]\left [ \begin{array}{c } \bar{\xi}_{2,i}\\ \bar{\zeta}_{2,i } \end{array}\right]dt + x(0)'\tilde{s}_{2,i}(\lambda)x(0)\geq0 . \end{split}\ ] ] hence , @xmath75 we have @xmath76 and @xmath77 , which imply the following relation : @xmath78 hence , we obtain @xmath79=\left [ \begin{array}{cc } \tilde{t}_{11}^{-1 } & -\tilde{t}_{11}^{-1}\tilde{t}_{12}\\ \tilde{t}_{21}\tilde{t}_{11}^{-1 } & \tilde{t}_{22}-\tilde{t}_{21}\tilde{t}_{11}^{-1}\tilde{t}_{12 } \end{array}\right ] \left [ \begin{array}{c } \bar{\xi}_{2,i}(t)\\ \tilde{\zeta}_{2,i}(t ) \end{array}\right].\ ] ] substituting for @xmath80 and @xmath81 into ( [ eqsystemc1 ] ) gives the following dynamical system : @xmath82 for all @xmath83 , and @xmath36 where @xmath84 also in order to deal with the @xmath85 terms in ( [ eqc1system ] ) we use standard loop shifting ideas @xcite where we require that the following condition is satisfied for all @xmath83 and @xmath36 : @xmath86 for this purpose , we first define the following quantities : @xmath87 by using the definition in ( [ eqphis ] ) , we define the transformed uncertainty inputs and outputs as follows : @xmath88;\\ \check{\zeta}_{2,i } & { \stackrel{\triangle}{=}}\bar{\phi}_i^{-1/2 } [ \bar{c}_{2,i } x + \bar{d}_{2,i}\tilde{u } ] . \end{split}\ ] ] hence , @xmath89 $ ] and ( [ eqnniqc ] ) can be rewritten as follows for all @xmath36 : @xmath90 now , we re - write ( [ eqc1system ] ) using ( [ eqndef ] ) as follows : @xmath91 for all @xmath83 , and @xmath36 where @xmath92 in order to obtain a bound on the cost function ( [ eqnfcost ] ) , we design a guaranteed cost controller for the system ( [ eqc2system ] ) .
the theory of guaranteed cost controllers can be found in @xcite . in order to apply a guaranteed cost controller of the form ( [ eqnestimator ] )
, a parameter dependent algebraic riccati equation is required to be solved for different values of the multipliers @xmath93 and @xmath55 for all @xmath37 and @xmath36 .
this riccati equation is given below :
@xmath94 where @xmath95,d_\tau=\left [ \begin{array}{c } 0\\ g^{1/2}\\ \sqrt{\tau_{1,1 } } \tilde{d}_{1,1}\\ \vdots\\ \sqrt{\tau_{1,k } } \tilde{d}_{1,k}\\ \check{d}_{2,1}\\ \vdots\\ \check{d}_{2,g}\\ \end{array}\right],g_\tau = d_\tau^t d_\tau,\\ & b_{2\tau } = \left [ \begin{array}{cccccc } \frac{\tilde{b}_{1,1}}{\sqrt{\tau_{1,1 } } } & \cdots & \frac{\tilde{b}_{1,k}}{\sqrt{\tau_{1,k } } } & \check{b}_{2,1 } & \cdots & \check{b}_{2,g } \end{array}\right ] .
\end{split}\ ] ] the parameters @xmath96 and @xmath55 are chosen such that the riccati equation ( [ eqare ] ) has a positive definite solution @xmath97 .
[ ass1 ] for any @xmath93 and @xmath55 @xmath98 and @xmath36 satisfying conditions ( [ eqcondq ] ) , ( [ eqcondd1 ] ) , the following conditions are satisfied : 1 .
the riccati equation ( [ eqare ] ) has a symmetric nonnegative definite solution @xmath99 .
if the conditions ( [ eqcondq ] ) , ( [ eqcondd1 ] ) along the assumption [ ass1 ] are satisfied then a controller of the form ( [ eqnestimator ] ) for the system ( [ eqc2system ] ) can be obtained as follows : @xmath100\tilde{x}\ ] ] and hence we define @xmath101 $ ] .
the corresponding bound on the cost function is obtained as follows : @xmath102 suppose there exist constants @xmath93 and vectors @xmath55 @xmath98 and @xmath36 such that conditions ( [ eqcondq ] ) , ( [ eqcondd1 ] ) , along with assumption 1 are satisfied . then : 1 . if the controller defined by ( [ eqare ] ) , ( [ eqdefares ] ) , and ( [ eqcontrol2 ] ) is applied to the uncertain system defined by ( [ eqfiqc ] ) , ( [ eqc2system ] ) , ( [ eqfiqc ] ) , then the cost functional ( [ eqnfcost ] ) satisfies the bound @xmath103 .
2 . if the controller defined by ( [ eqare ] ) , ( [ eqdefares ] ) , and ( [ eqcontrol2 ] ) is applied to the uncertain system defined by ( [ eqnniqc ] ) , ( [ eqc1system ] ) , then the cost functional ( [ eqnfcost ] ) satisfies the bound @xmath103 .
* _ proof : _ * the first part of the theorem follows directly from the main results of @xcite [ theorem 5.3.1 ] .
the second part of the theorem is a result of condition ( [ eqcondd1 ] ) which allows for the system ( [ eqnniqc ] ) , ( [ eqc1system ] ) to be written in the form ( [ eqfiqc ] ) , ( [ eqc2system ] ) and hence application of the result in @xcite [ theorem 5.3.1 ] to this system will result in @xmath103 as noted in observation [ obs1 ] . @xmath104
[ th1 ] suppose there exist constants @xmath93 and vectors @xmath55 @xmath98 and @xmath36 such that conditions ( [ eqcondq ] ) , ( [ eqcondd1 ] ) , along with assumption 1 are satisfied . if the nonlinear controller defined by ( [ eqnestimator ] ) , ( [ eqnon2 ] ) , ( [ eqnstate ] ) is applied to the nonlinear uncertain system defined by ( [ eqsystem ] ) , ( [ eqnon1 ] ) , ( [ eqiqc1 ] ) , then the cost functional ( [ eqnfcost ] ) satisfies the bound @xmath103 .
* _ proof : _ * the result directly follows from part ( ii ) of theorem [ th1 ] and the construction of the iqc ( [ eqnniqc ] ) , the system ( [ eqc1system ] ) and the controller ( [ eqnestimator ] ) along with the discussion in observation [ obs1 ] . @xmath104
an example of state feedback control of axial compressor surge is considered in @xcite and is given as follows : where @xmath106 and @xmath107 are the system states , and @xmath108 is the control input . in order to obtain a nonlinearity which is monotonic and sector bounded , we add a linear function to the nonlinearity .
we also add an additional uncertainty satisfying an iqc for robustness purposes .
hence , we obtain where @xmath110 . the uncertainty input
@xmath111 satisfies the iqc @xmath112 for all @xmath113 .
we solve the algebraic riccati equation ( [ eqare ] ) for the steady state stabilizing solution for possible values of @xmath114 and @xmath115 satisfying conditions ( [ eqmon1 ] ) , ( [ eqmon2 ] ) , ( [ eqcondd1 ] ) along with assumption 1 and considering the iqcs ( [ eqiqc1 ] ) , ( [ eqiqc2])-([eqiqc4 ] ) .
these values of @xmath114 and @xmath116 $ ] are chosen so that steady state cost bound ( [ eqbound ] ) is minimized .
the value of bound on the cost functional ( [ eqnfcost ] ) for an initial condition of @xmath117 and @xmath118 is obtained as @xmath119 for the following values of the parameters : the cost bound obtained using this scheme is lower than the cost bound obtained in @xcite for the same example .
this is expected as in the state feedback design we assume that all states are available for measurement .
the nonlinear system with the nonlinear state feedback controller has also been simulated using the above initial conditions and by assuming @xmath121 .
the result of the simulation is presented in fig .
[ fig : simulation ] .
it is observed that the control system performance is satisfactory
. i. r. petersen , `` robust output feedback guaranteed cost control of nonlinear stochastic uncertain systems via an iqc approach , '' _ ieee transactions on automatic control _ , vol .
54 , no . 6 , pp .
12991304 , 2009 . | this paper presents a systematic approach to the design of a robust dynamic state feedback controller using copies of the plant nonlinearities , which is based on the use of iqcs and minimax lqr control .
the approach combines a linear state feedback guaranteed cost controller and copies of the plant nonlinearities to form a robust nonlinear controller . , , |
this work was supported by the bureau national de la mtrologie , the dlgation gnrale de larmement , and the rgion ile de france , the deutsche forschungsgemeinschaft sfb 407 and spp1116 , the rtn cold quantum gases , ist program equip , esf pesc bec2000 + , epsrc , and the ministero dellistruzione , delluniversit e della ricerca ( miur ) .
p.p . also wants to thank the alexander von humboldt foundation and the zip programm of the german government .
we make the following ansatz for the non equilibrium distribution function : @xmath93 with @xmath94 , @xmath95 and @xmath96 .
the dependence on time is contained through the parameters @xmath21 and @xmath22 .
following @xcite , we substitute this ansatz into eq .
( [ beq ] ) and use the equation for the equilibrium distribution @xmath17 .
we find @xmath97 \bigg\}=i_{\rm coll } \label{bs}\end{aligned}\ ] ] performing integration in phase space , we calculate the average moment of @xmath98 , namely @xmath99 d^3rd^3v / n$ ] .
this leads to eq .
( [ s1 ] ) .
note that this equation is not affected by the collision integral since the quantity @xmath98 is conserved by collisions .
to derive eq .
( [ s2 ] ) , we consider instead the average moment of @xmath100 .
this yields : @xmath101 where @xmath82 is the equilibrium temperature . differently from ( [ s1 ] ) , eq . ( [ z1 ] ) depends explicitly on the collision integral . in order to calculate the r.h.s of eq .
( [ z1 ] ) , we use the relaxation time approximation : @xmath102 .
the first term gives : @xmath103 . to obtain a relation among the @xmath104 scaling parameters one uses the identity @xmath105 , from which we deduce @xmath106 .
the contribution of the local equilibrium term to the second term is obtained by noticing that , at local equilibrium , @xmath107 : @xmath108 .
( [ z1 ] ) can be recast in the form ( [ s2 ] ) .
y. castin _
et al . _ ,
lett . * 77 * , 5315 ( 1996 ) ; v.m .
perez - garcia _ et al .
lett . * 77 * , 5320 ( 1996 ) ; k.g .
rokhsar , phys .
lett . * 77 * , 1667 ( 1996 ) ; f. dalfovo _
a * 227 * 259 ( 1997 ) . in the unitary limit of @xmath8-wave collisions ,
the cross section @xmath69 scales as @xmath112 where @xmath113 is the wave vector for the relative velocity which suggests a different scaling for @xmath52 given by @xmath114 . | by means of a scaling ansatz , we investigate an approximated solution of the boltzmann - vlasov equation for a classical gas . within this framework , we derive the frequencies and the damping of the collective oscillations of a harmonically trapped gas and we investigate its expansion after release of the trap .
the method is well suited to studying the collisional effects taking place in the system and in particular to discussing the crossover between the hydrodynamic and the collisionless regimes .
an explicit link between the relaxation times relevant for the damping of the collective oscillations and for the expansion is established .
the favored signature of bose - einstein condensation in weakly interacting gases is the time - of - flight expansion @xcite . in this technique
, the asymmetric trapping potential is switched off and the evolution of the spatial density is monitored . after long time expansion , the observed inversion of the aspect ratio reflects the anisotropy of the initial confinement . in an ideal bose - einstein condensate ( bec ) , this effect is a direct consequence of the heisenberg uncertainty constraint on the condensate wave function . for an interacting bec ,
the inversion is also produced by the anisotropy of the pressure gradients caused by the hydrodynamic forces .
the changes in the shape of the expanding gas can be characterized by scaling factors , which provide an easy quantitative tool for the analysis of on - going bec experiments .
the set of equations for those factors have been derived in many papers @xcite .
a similar effect has also been predicted for a fermi gas in its superfluid phase @xcite .
strong anisotropy has been recently measured in the expansion of a highly degenerate fermi gas @xcite close to a feshbach resonance .
resonance scattering can also give rise to anisotropic expansion in the normal phase as proven in the experiments of @xcite carried out in a less degenerate regime .
some recent experiments on bosonic atoms above the critical temperature have also reached the collisional regime investigating both the oscillations of the low lying quadrupole mode and the expansion in asymmetric traps @xcite .
so far analytic calculations for the expansion of a classical gas have been limited to either the ballistic or to the hydrodynamic regime @xcite . as a consequence ,
it is important to generalize such calculations to all intermediate collisional regimes .
this is precisely the main purpose of this paper .
we begin by an outline of the theoretical description of the thermal gas based on the boltzmann - vlasov equation .
our approach relies on an approximated solution to this equation by means of a scaling ansatz .
this solution is used throughout the paper to investigate two kinds of related problems : the lowest collective oscillation modes and the time - of - flight expansion when the confinement is released .
the boltzmann - vlasov ( bv ) kinetic equation for the phase space distribution @xmath0 takes the form : @xmath1 , \label{bv}\end{aligned}\ ] ] where @xmath2 is the trapping potential chosen of harmonic form : @xmath3 .
interparticle interactions enter eq .
( [ bv ] ) in two different ways @xcite . on the one hand
, they modify the effective potential through the mean field term @xmath4 which affects the streaming part of the boltzmann kinetic equation .
the mean field potential @xmath4 is equal to @xmath5 for bosons and @xmath6 for two fermion species @xcite , where the coupling constant @xmath7 is fixed by the @xmath8-wave scattering length @xmath9 .
the mean field term is linear in @xmath9 and is non dissipative . on the other hand
, two body interaction determines the collision integral @xmath10 $ ] which is quadratic in the scattering length , and describes dissipative processes .
eq .
( [ bv ] ) is valid in the semi - classical limit , namely when the thermal energy is large compared with the separation between the energy eigenvalues of the potential @xcite . in this paper
, we will treat the collision integral within the relaxation time approximation .
this model should suffice to capture the essential physics of the problem .
we consequently write : @xmath11\approx-\frac{f - f_{\rm le}}{\tau},\end{aligned}\ ] ] where @xmath12 is the relaxation time related to the average time between collisions , and @xmath13 is the local equilibrium density in phase space . as a consequence , @xmath13 has a spherical symmetry in velocity space , _ i.e. _ it depends on the velocity through @xmath14 ^ 2 $ ] where @xmath15 is the local velocity field . the dynamics of the gas will be described by the following scaling ansatz for the non equilibrium distribution function : @xmath16 where @xmath17 is the equilibrium distribution function which satisfies the equation ( @xmath18=0 $ ] ) : @xmath19 the dependence on time of @xmath20 is contained in the dimensionless scaling parameters @xmath21 and @xmath22 .
the parameter @xmath21 gives the dilatation along the @xmath23 direction , while @xmath22 gives the effective temperature in the same direction .
such an ansatz generalizes the one used in @xcite .
we recall that in this method the shape of the cloud does not explicitly enter the equations .
this is the reason why the solutions are equally valid for a dilute bose gas above the critical temperature , a dilute fermi gas in its normal phase , and a classical gas . following ref .
@xcite , one can derive the set of equations for the scaling parameters @xmath21 and @xmath22 ( see appendix a ) : @xmath24 , \label{s2}\end{aligned}\ ] ] where the dimensionless parameter @xmath25 accounts for the mean field interaction @xcite and @xmath26 is the average temperature , for a classical gas @xmath27 .
the parameter @xmath28 is expected to be small for dilute gases ( @xmath29 ) since the ratio @xmath30 scales as @xmath31 , where @xmath32 is the de broglie wavelength and @xmath33 the mean density @xcite .
eq .
( [ s2 ] ) shows that the dissipation occurs when the temperature is not isotropic and the relaxation time @xmath12 has a finite value .
eqs ( [ s1 ] ) and ( [ s2 ] ) are the main result of this paper .
the collisionless regime is obtained by taking @xmath34 . in this limit
, we have the simple relation @xmath35 between the scaling parameters , and we recover the equations derived in @xcite . in the opposite limit ( hydrodynamic regime ) , local equilibrium is always ensured because of the high collision rate . as a consequence , we have @xmath36 and the eqs ( [ s1 ] ) and ( [ s2 ] ) can be recast in the form : @xmath37 for @xmath38 ( no mean field ) , we recover the equations first derived in @xcite .
note that in both the collisionless and the hydrodynamic regimes , the collisional term does not contribute since there is no dissipation in these limits .
we next focus our attention on the intermediate regimes where the collision term enters explicitly the equations of motion .
let us first study the breathing mode in the case of spherical harmonic trapping with angular frequency @xmath39 . in this case
, we find a solution with @xmath40 and @xmath41 .
for such a solution the collision term identically vanishes in all intermediate collisional regimes
. our approach can be readily generalized to lower dimensions leading to the frequency @xmath42 for the monopole mode @xcite , where @xmath43 is the dimension of space . in two dimensions
the mean field does not affect the frequency of the monopole .
this comes out from the fact that in this case the ansatz is an exact solution of the bv equations , as already stressed in ref .
@xcite .
we now consider a sample of atoms confined in a three - dimensional cylindrically symmetric harmonic potential .
we denote by @xmath44 the ratio between the axial and radial angular frequencies .
expanding eqs .
( [ s1 ] ) and ( [ s2 ] ) around equilibrium ( @xmath45 ) we get a linear and closed set of equations which can be solved by looking for solutions of the type @xmath46 .
the associated determinant yields the dispersion law : @xmath47+\frac{1}{\tau_0}b[\omega]\bigg ) \bigg(c[\omega]+\frac{1}{\tau_0}d[\omega]\bigg)=0 , \label{disp}\ ] ] where @xmath48=\omega^2(\omega^2-\omega_{\rm cl + } ^2)(\omega^2-\omega_{\rm cl-}^2)$ ] , @xmath49=\omega(\omega^2-\omega_{\rm hd + } ^2)(\omega^2-\omega_{\rm hd -}^2)$ ] , @xmath50=\omega(\omega^2-\omega_{\rm cl } ^2)$ ] and @xmath51=(\omega^2-\omega_{\rm hd}^2)$ ] and @xmath52 is the value of the relaxation time @xmath12 calculated at equilibrium and @xmath53 here @xmath28 is the parameter accounting for the mean field effects and ( cl ) and ( hd ) refer to the collisionless and hydrodynamics regimes respectively .
the solution of eq .
( [ disp ] ) interpolates the frequencies of the low lying modes for all collisional regimes ranging from the collisionless to the hydrodynamic . as the confinement is cylindrically symmetric around the @xmath54 axis , we can label the modes by their angular azimuthal number @xmath55 .
the first factor of the l.h.s . of eq .
( [ disp ] ) gives the frequencies of the two @xmath56 modes , while the second factor gives that of the quadrupole ( @xmath57 ) mode .
the roots of @xmath58 and @xmath59 have already been obtained in @xcite , and correspond to the frequencies of the low lying modes of a collisionless gas in presence of mean field .
eq .
( [ disp ] ) for @xmath38 has been derived in ref .
@xcite and the corresponding frequencies have been investigated experimentally @xcite . for @xmath60 , corresponding to @xmath61 , we find @xmath62 , @xmath63 , and the frequencies coincide with the ones predicted for a bose - einstein condensate in the thomas - fermi regime @xcite .
so far , we have not given the explicit link between the relaxation time entering eq .
( [ disp ] ) and the collision rate .
following ref .
@xcite , we can establish this link for a classical gas by means of a gaussian ansatz for the equilibrium distribution function @xmath64 .
one obtains @xmath65 where @xmath66 is the classical collision rate where @xmath67 is the thermal velocity , @xmath68 is the peak density and @xmath69 is the cross section which is assumed to be velocity independent . for bosons
the link between the scattering length and the cross section is @xmath70 whereas for two fermions species one has @xmath71 .
we now establish the set of equations that describe the time - of - flight expansion . in the collisionless regime where the mean free path
is very large with respect to the size of the trapped cloud and in the absence of mean field contribution , we readily obtain the exact equations @xmath72 which admit the solutions @xmath73 , leading to isotropic density and velocity distributions after long time expansion .
when the effect of collisions is important the physics of the expansion changes dramatically . as an example , the radial directions of a cigar - shaped cloud expand faster than the longitudinal one finally resulting in an anisotropic velocity distribution .
so far , an analytic approach has been proposed only in the full hydrodynamic regime @xcite . however
, this approach assumes that the hydrodynamic equations are always valid during the expansion . in general
, this can not be the case since the density decreases during the expansion reducing the effect of collisions .
alternatively , the expansion of an interacting bose above @xmath74 gas has been investigated by means of a monte carlo simulation @xcite . in our approach
, we provide an interpolation between the two opposite collisionless and hydrodynamic regimes using the scaling formalism .
the decrease of the collision rate during the expansion yields a non constant relaxation time @xmath75 that depends explicitly on the scaling parameters and reflects the changes of the density and the temperature during the expansion . as a result
, the expansion is described by the following set of 6 non - linear equations @xmath76 the dependence of the relaxation time @xmath12 on the scaling parameters is obtained by noticing that the collision rate @xmath77 scales as @xmath78 . using the scaling transformation @xmath79 and @xmath80 , where @xmath81 and @xmath82 are the initial density and temperature respectively , we deduce @xmath83 where @xmath52 is the average time of collision at equilibrium @xcite . since both the results ( [ disp ] ) for the dispersion of the linear oscillations and eqs .
( [ expan ] ) and ( [ tauexp ] ) for the expansion have been derived starting from the same scaling equations ( [ s1 ] ) and ( [ s2 ] ) the relaxation time @xmath52 entering the two processes is the same . as a consequence ,
the combined investigation of the expansion and of the quadrupole oscillations can provide a useful check of the consistency of the approach and , possibly , useful constraints on the value of the cross section .
the time evolution of the aspect ratio @xmath84 in the absence of mean field is depicted on fig . 1 for different values of the product @xmath85 and for an initial aspect ratio @xmath86 .
in the collisionless regime ( @xmath34 ) , the aspect ratio tends asymptotically to unity reflecting the isotropy of the initial velocity distribution . for other collisional regimes ,
the asymptotic aspect ratio is larger that one as a consequence of collisions during the expansion .
we find a continuous transition from the collisionless to the hydrodynamic prediction as we decrease @xmath52 from infinity to zero .
we then conclude that in general the expansion can not be described with either the hydrodynamic or the collisionless prediction @xcite , but requires a full solution of our equations ( [ expan ] ) .
similar conditions have been already encountered experimentally for classical or almost classical gases @xcite .
we also notice that it is very important to take into account the time dependence of the relaxation time , accounted for by the scaling law ( [ tauexp ] ) .
for example by simply using @xmath87 during the whole expansion , the curve @xmath88 of fig . 1 ( solid line ) would be shifted upward and the resulting prediction would result much closer to the hydrodynamic curve ( dotted line ) .
_ aspect ratio as a function of the normalized time @xmath89 for different collisional regimes ( initial aspect ratio @xmath86 ) : collisionless ( dashed line , @xmath90 ) , intermediate collisional regime ( solid line , @xmath91 ) , hydrodynamic regime ( dotted line , @xmath92 ) .
_ [ fig1 ] let us finally comment on the effect of quantum statistics on the calculation of the relaxation time .
for a bose gas at temperature above @xmath74 the problem has been investigated in @xcite where it has been shown that statistical effects do not play a significant role .
in contrast , the relaxation time in a harmonically trapped dilute fermi gases has been shown to be strongly affected by pauli blocking at low temperature @xcite .
the effects of collisions in a strongly interacting fermi gas , including the unitarity limit , have been recently addressed in @xcite . |
since its formulation , quantum mechanics ( qm ) has always peaked the interest of individuals with its inherent mystery and its seemingly paradoxical , but experimentally reproducible , results . at the heart of this mystery
is the superposition principle and the measurement problem .
the situation has been nicely described by wallace @xcite : solutions of the measurement problem are often called ` interpretations of qm ' , the idea being that all such ` interpretations ' agree on the formalism and thus the experimental predictions " , and thus the interpretation of qm in entropic dynamics @xcite _ _ _ _ will inevitably be discussed in the context of the measurement problem .
ed differs from most physical theories . in most approaches to qm one starts with the formalism and the interpretation is appended to it , almost as an afterthought . in ed one starts with the interpretation , that is , one specifies what the ontic elements of the theory are , and only then one develops a formalism appropriate to describe , predict and control those ontic elements .
for instance , to derive non - relativistic qm one starts from the assumption that particles have definite yet unknown positions and then one proceeds to derive ed as a non - dissipative diffusion .
the solution to the measurement problem should address two problems : one is the problem of _ definite outcomes _ ( when and how does the wavefunction collapses ) , the other is the problem of the _ preferred basis _ , which is also known as the _ basis degeneracy problem_. ( for a review see @xcite . ) both problems as well as the interpretation of qm in ed will be addressed in the body of this paper .
entropic dynamics ( ed ) derives laws of physics as an example entropic inference .
this view is extremely constraining .
for example , there is no room for quantum probabilities probabilities are neither classical nor quantum , they are tools for reasoning that have universal applicability . related to this is the fact that once one claims that the wavefunction is an epistemic object ( _ i.e. _ , @xmath0 is a probability ) then its time evolution the updating of @xmath1 is not at all arbitrary
. the laws of dynamics must be dictated by the usual rules of inference .
there is one feature of the unified bayesian / entropic inference approach that turns out to be particularly relevant to the foundations of quantum theory .
the issue in question is von neumann s proposal of two modes of evolution for the wavefunction , either continuous and unitary as given by the schrdinger equation , or discontinuous and stochastic as described by the projection postulate
the collapse of the wavefunction .
once one adopts the broader inference scheme that recognizes the consistency of bayesian updating with entropic updating @xcite the apparent conflict between the two modes of wavefunction evolution disappears : the wavefunction is an epistemic object meant to be updated on the basis of new information .
when the system is isolated the evolution is obtained by a continuous entropic updating which leads to a unitary evolution described by the schrdinger equation .
when , in a measurement , the information is in the form of discrete data the bayesian update is discontinuous and the wavefunction collapses .
both forms of updating coexist harmoniously within the framework of ed @xcite . in the von neumann measurement scheme ,
the macroscopic measurement device is treated quantum mechanically @xcite .
the initial state of the measuring device is generally given as @xmath2 in indicating the device is in the ready or reference state .
an interaction between the system , in an initial state @xmath3 and the measuring device leads to a state in which the system becomes entangled with the device s
pointer variable , @xmath4 the standard von neumann measurement scheme involves interactions such that the system and measurement device evolve into a special entagled state called a biorthogonal state ( _ i.e. _ , there are no cross terms in eq.([vn a ] ) , @xmath5 for @xmath6 ) schlosshauer .
thus , a measurement that finds the pointer variable in state @xmath7 seemingly allows the observer to infer that the system is in state @xmath8 .
however , the right hand side of eq.([vn a ] ) can be expanded in some other basis .
there is no ontological matter of fact about wihich outcome has been obtained .
thus we have problems : not only we have a never observed phenomenon of macroscopic entanglement but we also have a _ problem of no definite outcomes _ also called the _ preferred basis problem _ or the _ basis degeneracy problem _ @xcite@xcite . in ed
the collapse of the wave function involves a straightforward application of bayes theorem to find @xmath9 , the probability the system is in state @xmath8 given a detection of the pointer variable in state @xmath10 . in ed it is possible and convenient to introduce observables other than position ( _ e.g. _ , momentum , energy , and so on ) but these are not attributes of the system , they are attributes of their probability distributions .
while positions are ontic elements , these other observables are purely epistemic ; they are properties of the wave function , not of the system @xcite .
these ideas can be pushed to an extreme when discussing the notion of the weak value @xcite-@xcite of an operator @xmath11 in which the system is prepared in an initial state @xmath12 and post - selected in state @xmath13 , @xmath14weak values are complex and therefore strictly they can not be observed , and yet they may still be _
inferred_. from the perspective of an inference framework such as ed the common reference to ` observables ' is misguided .
it should be replaced by bell s term ` beables ' @xcite for ontic elements such as particle positions , and perhaps ` inferables ' for those epistemic elements associated to probability distributions .
after a brief review of ed and of the simplest or direct type of measurement @xcite , we discuss the less direct types of measurement the von neumann and weak measurements in which information about the system is inferred by observation of another system , the pointer device , to which it has been suitably correlated .
finally we discuss weak values in the context of ed .
here we give a brief overview of ed emphasizing those ideas relevant to the measurement problem .
( for a review with more extended references see ed . )
the system consists of @xmath15 particles living in a flat euclidean space @xmath16 with metric @xmath17 .
the positions of the particles have _ definite _ values which are however _ unknown_. we already see that ed differs from the copenhagen interpretation :
in the latter , positions become definite only as a result of the measurement process .
positions are denoted @xmath18 where @xmath19 denotes the particle and @xmath20 denotes the spatial coordinates .
the microstate @xmath21 , where @xmath22 , is a point in the @xmath23 dimensional configuration space .
the objective is to predict the particle positions and their motion .
the first assumption is that particles follow continuous trajectories so that the motion can be analyzed as a sequence of short steps .
we use the method of maximum entropy to find the probability @xmath24 that the particles take infinitesimally short steps from @xmath18 to @xmath25 .
the information that the steps are infinitesimal is given by @xmath15 constraints , @xmath26(eventually we take the limit @xmath27 imposing continuous motion ) . to introduce correlations
we impose one additional constraint , @xmath28where @xmath29 , @xmath30 is another small constant , and @xmath31 is the drift potential .
it is this single constraint acting on the @xmath23 dimensional configuration space that leads to quantum effects such as interference and entanglement .
the physical nature of the @xmath31 potential need not concern us here .
we merely postulate its existence and note @xmath31 is closely related to the phase of the wave function .
affects the motion of particles it plays the role of a pilot wave or an electromagnetic field .
indeed , @xmath31 is _ as real as _ the vector potential @xmath32 .
the close relation between them is the gauge symmetry ( see @xcite ) . ]
the result of maximizing entropy with an appropriate choice of lagrange multipliers leads to @xmath33~ , \label{prob xp / x a}\]]where @xmath34 is a normalization constant , @xmath35 is the time interval between @xmath36 and @xmath37 , the @xmath38 s are particle - specific constants called masses , and @xmath39 is a constant that fixes the units of time relative to those of length and mass @xcite . once we have @xmath24 for an infinitesimal step we can ask how a probability distribution @xmath40 evolves from one step to the next step @xmath41 by considering , @xmath42 iterating this process for very short steps allows one to find the evolution of @xmath43 to be given by the fokker - planck equation ( an explication can be found in @xcite ) , @xmath44where @xmath45 is the velocity of the probability flow in configuration space or _ current velocity _ , @xmath46equation ( [ fp b ] ) describes a standard diffusion . to obtain a non - dissipative dynamics the choice of drift potential @xmath31 must be revised after each step @xmath35 , which means that the drift potential , or equivalently , the phase @xmath47 , is promoted to a dynamical degree of freedom .
the correct updating , @xmath48 , follows from requiring that a certain functional @xmath49 $ ] be conserved .
the result is that the coupled evolution of @xmath50 and @xmath47 is given by a conjugate pair of hamilton s equations , @xmath51the ensemble hamiltonian @xmath52 is chosen so that the first equation above reproduces ( [ fp b ] ) in which case the second equation in ( [ hamilton ] ) becomes a hamilton - jacobi equation .
a more complete specification of @xmath52 suggested by information geometry , @xmath53=\int dx\,\left [ \frac{1}{2}\rho m^{ab}\partial _ { a}\phi \partial _ { b}\phi + \rho v+\xi m^{ab}\frac{1}{\rho } \partial _ { a}\rho \partial _ { b}\rho \right ] ~ , \label{hamiltonian}\]]includes the kinetic term that reproduces ( [ fp b ] ) , a potential @xmath54 , and a quantum potential .
the parameter @xmath55defines the value of planck s constant @xmath56 .
the formulation of ed is now complete : combining @xmath50 and @xmath47 into a single complex function , @xmath57 with @xmath58 allows us to rewrite ( [ hamilton ] ) as a _ linear _ schrdinger equation,@xmath59 at this point we can adopt the standard hilbert space formalism to represent the epistemic state @xmath60 as a vector , @xmath61we can express @xmath62 in bases other than the position basis @xmath63 and we can write the schrdinger equation as a manifestly unitary evolution , @xmath64 for some suitably chosen @xmath65 .
the particles have definite positions but their brownian paths are not smooth ; there is no momentum associated to the trajectories .
one can define the usual quantum momentum the operator that generates translations of the wavefunction .
in general it does not have definite values .
its values , which are inferred by the process of measurement and detection , are not an ontic attribute of the particles but an epistemic attribute of the wave function .
the simplest type of quantum measurement , studied in @xcite , consists of the direct detection of the particle s position .
since the wavefunction is purely epistemic , the collapse problem does not arise .
the collapse is nothing but a bayesian update from a prior to a posterior and in general the preferred basis problem for a single orthonormal vector does not apply @xcite . in this context ,
a question of interest in ed is how do observables other than position arise was also answered in @xcite .
consider , for simplicity , a particle that lives on a discrete lattice , so that if the state is @xmath66 in a more complicated measurement the particle is subject to additional interactions right before reaching the position detector .
such a complicated setup @xmath67 is described by a particular _ unitary _ evolution @xmath68 .
the particle will evolve to the position @xmath69 with certainty provided it was initially in a state @xmath70 such that @xmath71 .
since the set @xmath72 is orthonormal and complete , the set @xmath73 is also orthonormal and complete . to figure out the effect of @xmath67 on some generic initial state vector @xmath62 , expand as @xmath74where@xmath75 .
then the state @xmath62 evolves according to @xmath68 into the new state at a later time @xmath76 , @xmath77which , invoking the born rule for _ position _ measurements , implies that the probability of finding the particle at the position @xmath78 is @xmath79 . from a physics perspective
there is nothing more to say but we can adopt _ a different language _ : we can _ say _ that the particle has been measured _ as if _ it had earlier been in the state @xmath70 .
thus , the setup @xmath67 is a complicated device that in principle measures all operators of the form @xmath80 where the eigenvalues @xmath81 are _ arbitrary _ scalars .
note that there is no implication that the particle previously had or now currently has the value @xmath81 this is just a figure of speech .
von neumann measurements are explored briefly and bayes rule is used as an instance of collapse " in amplification processes @xcite .
in this section , the von neumann and weak measurement schemes are naturally implemented into ed .
we will attempt to be particularly careful in our word choice such that a _ measurement _ in ed involves first an _ interaction _ with an apparatus or with another auxiliary system with which it becomes entangled .
this step does not collapse the wave function .
then there is the actual _ detection _ , which is the step that collapses or localizes the distribution to the value detected by the observer .
here we present ed s solution to the problem of _ definite outcomes _ and discuss the collapse " of the wavefunction .
entropic dynamics is a theory of inference and the rules for probability updating are precisely the same as in any other inference problem .
there is no such thing as quantum probabilities . the problem of physical " collapse is never encountered in ed for the same reason that it is not encountered in epistemic classical probability theory .
no - one asks how the probability distribution of a die role collapses during measurement ; this just follows from the inductive logic expressed by bayes rule .
when a particle with wave function @xmath82 is detected at @xmath83 with certainty the prior probability @xmath84 is updated to @xmath85 .
a more realistic detector is represented ( for the purpose of inference ) by a likelihood function @xmath86 that gives the probability of a detection @xmath87 when the particle is at @xmath36 .
then the probability that the particle was at @xmath36 given the detection event @xmath87 is given by bayes rule , @xmath88 an idealized device detects the presence of a particle in an infinitesimally small region centered around @xmath83 so the likelihood function in this case is @xmath89 and indeed @xmath90 .
as mentioned in the introduction , a von neumann measurement scheme @xcite is one in which a detection of the pointer variable in state @xmath91 gives @xmath92 with certainty and it potentially suffers from the _ preferred basis problem _ or equivalently the _ basis degeneracy problem _ @xcite .
we circumvent this problem by using bayes theorem and a specific unitary evolution to entangle the system and measurement device via the weak measurement scheme @xcite . in ed
the only positions are beables , all other
observables " must be inferred from detections of position and strictly are properties of the wavefunction or probability distribution . in this sense ,
quantities such as momentum or energy are not ontic ( in ed ) but rather are seen to be convenient parameters for expressing the probability distribution in question , that is , they are epistemic quantities verifying @xcite . for instance , an expectation value can be a useful parameter for expressing our knowledge of the state of a system but in general it is a property belonging to the probability distribution rather than being something persistently expressed through the system as an ontic variable is .
the simplest pointer measurement device we can imagine in this framework uses the position of a single test particle @xmath36 as its pointer variable .
the construction of a macroscopic pointer measurement device is an exercise in the central limit theory for a many particle system @xcite . to avoid clutter we will neglect normalization factors until we reach the level of probability distributions near the end of the calculation .
an ideal pointer measuring device in ed is one in which a test particle has a definite initial state , @xmath93 in position space and we will let @xmath94 for simplicity .
we consider a more general case in which the test particle is in the initial state , @xmath95 which reproduces the ideal measurement device when @xmath96 after normalization .
using the completeness relationship @xmath97 , the state of the test particle can be represented in momentum space as , @xmath98 the quantum system to be measured with the pointer device is a preselected or prepared superposition state of the eigenvectors of @xmath99 , @xmath100 we will consider a situation in which the measuring device is coupled to the system to be measured by a strong coupling or interaction hamiltonian , @xmath101 where @xmath102 is the canonical conjugate of the pointer variable and thereby generates translations in the pointer variable @xmath103 and @xmath104 is a function with compact support near the time of measurement and integrates to unity @xcite .
we can assume that the coupling hamiltonian will dominate over the full hamiltonian for the , assumed small , period of measurement .
the time evolution of our entangled system is , @xmath105 @xmath106 which , in the position space representation , is @xmath107 this is a superposition of potentially overlapping gaussian distributions having peaks at the eigenvalues of @xmath11 .
when the gaussian distributions overlap we have a so - called weak measurement " @xcite ; when the gaussian distributions are neatly resolved we have a `` strong '' or von neumann measurement .
the joint probability of the system , @xmath108 in the present case , we may infer the probability the system of interest is accurately _
describable _ by a particular eigenvector @xmath92 by making detections of the position of the pointer particle following notions from the detection in ed : ... " section and @xcite .
this is given by the conditional probability , @xmath109 thus the detection of the test particle at @xmath110 allows us to infer the probability that the system is in a state with eigenvalue @xmath111 .
bayes rule leads to a partial collapse because the initial state of the pointer variable is itself uncertain .
in the limit that @xmath112 for @xmath113 ( after normalization ) we have an ideal ( von neumann ) measurement .
the measurement always produces one of its eigenstates with probability @xmath114 , and from a detection of the position @xmath110 we can infer the eigenstate @xmath92 with probability 1 using ( [ 22a ] ) .
because position is ontic in ed it always plays the role of the preferred basis , which eliminates the basis degeneracy problem .
so far we have taken into account the uncertainty in the preparation of the pointer variable but we have assumed that the pointer variable @xmath36 of the test particle has been measured precisely . if the detection of @xmath36 is noisy then there is a second source of uncertainty quantified by @xmath115 given by equation ( [ detect ] ) .
the probability of the value @xmath111 is @xmath116 by compounding a unitary detector @xmath117 from ( [ unitary ] ) , with the weak measurement devices above one can map a non - position pointer variable @xmath118 to a position pointer variable @xmath103 which in principle allows for a more robust set of operators @xmath119 to be measured .
we may then consider coupling hamiltonians @xmath120 , where @xmath121 is the fourier conjugate to @xmath118 , and before detection apply @xmath122 to put the pointer variable in the position basis .
the weak value @xmath123 of a hermitian operator @xmath99 was first introduced by aharonov , albert , and vaidman @xcite ( aav ) in 1989 as an interesting application of a weak measurement which gives nonintuitive results ( recent review @xcite ) .
what is particularly significant in aav s paper @xcite is not that they defined an odd quantity associated tor @xmath99 , but rather they found , after a series of approximations ( see @xcite ) a way in which @xmath124 could be `` measured . ''
a weak value is a complex number which may lie outside the set of eigenvalues of @xmath99 when @xmath125 is sufficiently small . due to this
, the interpretation of weak values has had a
colorful history `` @xcite , much of which can be summarized by the question , ' ' are weak values ontic properties of a particle ? " . in ed , variables other than position either are inferred from position ,
are useful parameters in the position probability distribution of the particle , or may be a particularly convenient basis for representing the position space probability distribution @xmath126 due to the completeness relations . from this perspective
it is clear that inferring @xmath124 does not indicate that the particle is ontically expressing @xmath124 .
consider the final state of an entangled detector from ( [ ent ] ) , given we consider the state is post - selected into @xmath127 : @xmath128 which in the position space representation gives , @xmath129 given the set up configuration is in the desired range of validity @xcite . in the ed framework
we are interested in the probability distribution of @xmath36 post - selected ( just like one may preselect a quantum state ) into @xmath130 , @xmath131)^2}{2\bigtriangleup^2}\big),\end{aligned}\ ] ] where @xmath124 is taken as a feature of the probability distribution of @xmath36 .
because @xmath124 appears as a parameter in the probability distribution of @xmath36 , we may consider @xmath132 and invert the problem to ask ,
what is the probability the parameter @xmath124 has a certain value given a detection of the pointer particle at @xmath110 " that is we may use @xmath133-\overline{x})^2}{2\bigtriangleup^2/n}\big)$ ] and the parameter estimation scheme in @xcite to find @xmath124 in agreement with @xcite .
there are several potentially interesting weak values but in particular consider the operator @xmath134 post selected in a momentum state , @xmath135 which is proportional to the full wavefunction and @xmath136 is a constant that will be removed after normalization @xcite .
et al _ show that the real and imaginary parts of @xmath137 are proportional to the position and momentum shifts of the pointer variable and claim they are measuring the wavefunction . from the ed perspective it is more appropriate to state that the value of the wavefunction at each @xmath36 is being inferred .
if the value of @xmath137 is inferred with certainty then it is possible to solve for the phase of the wavefunction ( up to an additive constant @xmath138 ) and also infer the values of the drift potential and its derivatives .
this provides a link to wiseman s use of weak values to measure the probability current in bohmian mechanics @xcite . especially in cases as exotic as the above , ed takes the standpoint that weak values and quantities other than position ( energy , momentum , etc . )
are best considered as epistemic
inferables " rather than ontic beables or observables .
we have discussed the von neumann and weak measurement schemes in the framework of entropic dynamics and have offered solutions to the problems of _ definite values _ and of the _ preferred basis_. from the perspective of ed quantum mechanics is a framework for updating probabilities , for processing information .
when the quantum system is left undisturbed it evolves smoothly and the appropriate tool for updating is the schrdinger equation ; when the system is coupled to a measuring device the appropriate tool to handle information in the form of data is bayes theorem .
there is no conflict : information of different types is handled differently .
the fact that position plays a privileged role positions are `` beables '' provides us with a natural pointer variable . in ed quantities
normally called `` observables '' such as energy and momentum , and the quantities called `` weak values '' are not ontic properties of the system .
they are epistemic properties of the wave function .
none of these quantities are beables and they can not be observed directly . instead they can be inferred by appropriate position measurements either of the system itself or of another system with which it is suitably entangled .
such quantities should more appropriately be referred as " inferables .
it is the explicit recognition of the ontic nature of positions vs the epistemic nature of other inferables that allows one to circumvent the paradoxes and confusion that have surrounded the problem of measurement .
we would to like the reviewers and also : m. abedi , d. barolomeo , n. carrara , a. fernandes , t. gai , s. ipek , k. knuth , s. nawaz , p. pessoa , j. tian , j. walsh , and a. yousefi for their many discussions on entropy , inference , and quantum mechanics .
a. caticha , _ entropic inference and the foundations of physics _
( monograph commissioned by the 11th brazilian meeting on bayesian statistics
ebeb-2012 ) ; http://www.albany.edu/physics/acaticha-eifp-book.pdf . j. von neumann , _ mathematische grundlagen der quantenmechanik _ ( springer - verlag , berlin , 1932 ) [ english translation : _ mathematical foundations of quantum mechanics _
( princeton university press , princeton , 1983 ) ] .
a. demme , a. caticha , _ the classical limit of entropic quantum dynamics _
, presented at maxent 2016 , the 36th international workshop on bayesian inference and maximum entropy methods in science and engineering ( ghent , belgium 2016 ) , in these proceedings . | the problem of measurement in quantum mechanics is studied within the entropic dynamics framework .
we discuss von neumann and weak measurements , wavefunction collapse , and weak values as examples of bayesian and entropic inference . |
let @xmath8 be a solution of the linear differential equation @xmath9 where @xmath10 are functions , rational in the independent variable @xmath11 .
we are interested in determining the equation of minimal degree @xmath12 @xmath13 such that * @xmath3 is a solution * the coefficients @xmath14 are algebraic functions in @xmath11 recall that a function @xmath15 is said to be algebraic in @xmath11 if there exists a polynomial @xmath16 with coefficients in @xmath17 , such that @xmath18 .
we shall suppose that , more generally , @xmath1 is an arbitrary differential field of characteristic zero with algebraically closed field of constants , @xmath2 is its algebraic closure , and @xmath19 , @xmath20 . to find the equation @xmath21 we consider the differential galois group @xmath22 of @xmath23 and its connected subgroup @xmath24 containing the unit element of @xmath22 .
our first result , theorem [ main ] , says that the orbit of @xmath3 under the action of @xmath24 spans the solution space of @xmath21 .
a particular attention is given further to the case in which ( [ aa ] ) is of fuchs or picard - fuchs type .
theorem [ main ] is re - formulated in terms of the action of the corresponding monodromy groups in theorem [ th : main1 ] and theorem [ th : algeq ] .
in the last part of the paper , section [ sec : ex ] , we apply the general theory to some abelian integrals appearing in the study of perturbations of the lotka - volterra system .
these integrals have the form @xmath25 where @xmath26 is a continuous family of ovals , @xmath27 and @xmath28 is a suitable rational one - form on @xmath29 . in the first case the abelian integral satisfies a picard - fuchs equation of order @xmath30 .
it has been shown by van gils and horozov @xcite , that @xmath31 satisfies also a second order differential equation whose coefficients are functions algebraic in @xmath11 .
this allows to compute the zeros of @xmath31 ( by the usual roll s theorem for differential equations ) and finally , to estimate the number of limit cycles of the perturbed plane foliation defined by @xmath32 where @xmath33 is a real polynomial one - form on @xmath34 . by making use of theorem [ main ]
we provide the theoretical explanation of the phenomenon observed observed first in @xcite , see section [ sec : hvg ] .
another interesting case , studied in the paper is when @xmath35 ( @xmath36 relatively prime ) .
the abelian integral @xmath31 satisfies a picard - fuchs equation of order @xmath37 , which is the dimension of the first homology group of the generic fiber @xmath38 .
we show that the minimal order of the equation ( [ bb ] ) is @xmath37 or @xmath39 or @xmath40 , and that the coefficients @xmath41 are rational in @xmath11 , see section [ sec : parab ] .
the meaning of this is that the differential galois group of the picard - fuchs equation is connected and , in contrast to @xcite , there is no reduction of the degree , which may only drop by one or two , depending on whether @xmath28 has or has not residues `` at infinity '' .
_ acknowledgements _ part of the paper was written while the first author was visiting the university paul sabatier of toulouse .
he is obliged for the hospitality .
let @xmath1 be a differential field of characteristic zero with algebraically closed field of constants @xmath42 , @xmath43 be a picard - vessiot extension for the homogeneous monic linear differential operator @xmath0 @xmath44 and @xmath45 a solution , @xmath46 .
we denote by @xmath47 the algebraic closure of @xmath1 which is also a differential field . _ a homogeneous monic linear differential operator @xmath5 with coefficients in @xmath2 is said to be annihilator of @xmath3 , provided that @xmath48 . the annihilator @xmath5 is said to be minimal , provided that its degree is minimal .
_ the definition has a sense , because the algebraic closure @xmath49 of @xmath50 is a differential field which contains @xmath50 and @xmath2 as differential subfields .
the minimal annihilator obviously exists and is unique , its degree is bounded by the degree of @xmath0 which is an annihilator of @xmath3 .
we are interesting in the following question _ for a given solution @xmath3 as above , find the corresponding minimal annihilator @xmath5 _ to answer , consider the differential galois group @xmath51 , which is the group of differential automorphisms of @xmath50 fixing @xmath1 . recall that @xmath22 is an algebraic group over @xmath42 , and let @xmath24 be the connected component of @xmath22 , containing the unit element ( the identity ) .
the intermediate field @xmath52 , @xmath53 , of elements invariant under @xmath24 is then a finite algebraic extension of @xmath1 .
we denote it by @xmath54 .
let @xmath55 be a basis of the @xmath42-vector space spanned by the orbit @xmath56 and consider the wronskian determinant in @xmath57 variables @xmath58 @xmath3 satisfies the differential equation @xmath59 and because of the @xmath42-linear independence of @xmath60 @xmath61 let @xmath5 be the monic linear differential operator defined by @xmath62 its coefficients are invariant under the action of @xmath24 , and hence they belong to the differential field @xmath63 our first result is the following [ main ] the differential operator @xmath5 ( [ minimal ] ) is the minimal annihilator of the solution @xmath3 .
let @xmath64 be the unique differential operator of minimal degree with coefficients in some algebraic extension @xmath65 of @xmath1 , such that @xmath66 .
denote by @xmath67 the picard - vessiot extension for @xmath64 .
as a first step , we shall show that @xmath67 can be identified to a differential subfield of the picar - vessiot extension @xmath50 for @xmath0 .
the algebraic closure @xmath49 of @xmath50 is a differential field which contains @xmath50 and every algebrtaic extension of @xmath1 ( hence it contains @xmath65 ) .
therefore the compositum @xmath68 of @xmath69 and @xmath70 , that is to say the smallest field containing @xmath50 and @xmath71 , is well defined @xcite .
the differential automorphisms group @xmath72 acts on the compositum @xmath68 and leaves @xmath50 invariant
. therefore @xmath73 leaves @xmath50 invariant too , and the orbit @xmath74 is contained in @xmath50 .
let @xmath75 be a basis of the @xmath42-vector space spanned by this orbit .
then @xmath3 satisfies the differential equation @xmath76 and the coefficients of the corresponding monic linear homogeneous differential operator belong to @xmath65 .
consider the ring of differential polynomials @xmath77\ ] ] in formal variables @xmath78 .
identifying differential operators on @xmath54 to polynomials ( the derivatives @xmath79 correspond to variables @xmath78 ) , we may consider the ideal @xmath80 generated by homogeneous linear differential operators with coefficients in @xmath65 which annihilate @xmath3 .
this is obviously a linear ideal which , according to the general theory ( see ( * ? ? ?
* proposition 1.8 ) ) , is principal in the following sense .
there exists a linear differential operator with coefficients in @xmath65 , which generates @xmath80 .
clearly the generator of @xmath80 is the operator @xmath64 defined above .
it follows that the solution space of @xmath64 can be identified to a @xmath42-vector subspace of the solution space of the operator defined by ( [ wronskian ] ) , which implies @xmath81 ( the first two inclusions hold by definition ) . at the second step of the proof we shall show that @xmath82 . indeed , the automorphisms group @xmath24 leaves fixed the elements of @xmath50 which are algebraic on @xmath1 . in particular , the elements of @xmath65 are fixed by @xmath24 and hence @xmath24 induces differential automorphisms of the picard - vessiot extension @xmath67 .
this shows that the solution space of @xmath64 contains the solution space of @xmath5 and @xmath83 reciprocally , if we consider ( by the construction above ) the ideal in @xmath84 generated by all linear homogeneous differential operators with coefficients in @xmath54 , which annihilate @xmath3 , then this ideal is linear and principal .
the generator of the ideal corresponds to the operator @xmath64 , and hence @xmath85 theorem [ main ] is proved .
to the end of this section we apply theorem [ main ] to fuchs and picard - fuchs differential operators .
the minimal annihilator of a solution is described in terms of the action of the monodromy group .
let @xmath0 be a fuchsian differential operator of order @xmath86 on the riemann sphere @xmath87 , @xmath88 be the set of its singular points .
the field of constants is @xmath89 , the coefficients of @xmath0 belong to the field of rational functions @xmath90 .
denote by @xmath91 the complex vector space of solutions of @xmath92 .
the monodromy group @xmath93 of @xmath0 is the image of the homomorphism ( monodromy representation ) @xmath94 the zariski closure of @xmath93 in @xmath95 is the differential galois group @xmath22 of @xmath0 : @xmath96 a vector subspace @xmath97 is invariant under the action of @xmath22 if and only if it is invariant under the action of @xmath93 .
a subspace @xmath97 is said to be _ virtually invariant _ , provided that it is invariant under the action of identity component @xmath24 of @xmath22 , or equivalently , under the action of @xmath98 . for an automorphism
@xmath99 the set @xmath100 is a vector subspace of the same dimension .
thus @xmath22 acts on the grassmannian space @xmath101 @xmath102 and for every plane @xmath103 the orbit @xmath104 is well defined .
[ virtual ] a plane @xmath103 is virtually invariant , if and only if the orbit @xmath105 is finite .
we have @xmath106 if the orbit @xmath107 is finite , then @xmath108 and hence @xmath109 is finite .
as @xmath24 is a connected lie group , then @xmath110 and @xmath111 is virtually invariant .
suppose that @xmath111 is virtually invariant .
as @xmath112 is a finite group , then @xmath110 implies that the orbit @xmath113 is finite and hence @xmath114 is finite too .
+ let @xmath0 be a fuchsian differential operator as above , and @xmath3 a solution , @xmath46 .
the minimal annihilator of @xmath3 is a differential operator @xmath5 of minimal degree with coefficients in some algebraic extension of @xmath115 .
thus @xmath5 is a fuchsian operator too , but on a suitable compact riemann surface realized as a finite covering of @xmath87 .
let @xmath116 be two virtually invariant planes containing the solution @xmath3 .
then @xmath117 is a virtually invariant plane containing @xmath3 .
this shows the existence of a unique virtually invariant plane @xmath111 of minimal dimension , containing @xmath3 .
we call such a plane minimal . according to lemma
[ virtual ] and theorem [ main ] the minimal annihilator of @xmath3 is constructed as follows .
let @xmath118 be a basis of the minimal virtually invariant plane @xmath111 containing @xmath3 .
consider the fuchsian differential operator defined as in formula ( [ minimal ] ) .
[ th : main1 ] the differential operator @xmath5 is the minimal annihilator of the solution @xmath3 .
the degree of @xmath5 equals the dimension of the minimal virtually invariant plane containing @xmath3 .
suppose finally that @xmath0 is a linear differential operator of picard - fuchs ( and hence of fuchs ) type
. we shall adapt theorem [ th : main1 ] to this particular setting .
let @xmath119 be a bivariate non - constant polynomial .
it is known that there is a finite number of atypical points @xmath120 , such that the fibration defined by @xmath121 @xmath122 is locally trivial .
the fibers @xmath38 , @xmath123 are open riemann surfaces , homotopy equivalent to a bouquet of a finite number of circles .
consider also the associated homology and co - homology bundles with fibers @xmath124 and @xmath125 respectively . both of these vector bundles carry a canonical flat connection . choose a locally constant section @xmath126 and consider the abelian integral @xmath127 where @xmath28 is a meromorphic one - form on @xmath29 which restricts to a holomorphic one - form on the complement @xmath128 .
the milnor fibration ( [ milnor ] ) induces a representation @xmath129 which implies the monodromy representation of the abelian integral @xmath31 .
let @xmath130 be a continuous family of complex vector spaces obtained by a parallel transport .
the space @xmath131 can be seen as a point of the grassmannian variety @xmath132 .
therefore the representation ( [ hmilnor ] ) induces an action of the fundamental group @xmath133 on @xmath132 .
we say that a complex vector space @xmath130 of dimension @xmath134 is virtually invariant , provided that its orbit in the grassmannian @xmath132 under the action of @xmath133 is finite . a virtually invariant space @xmath131 is said to be irreducible , if it does not contain non - trivial proper virtually invariant subspaces .
let @xmath135 be a locally constant section of the homology bundle defined by @xmath121 .
as intersection of virtually invariant vector spaces @xmath136 containing @xmath135 is virtually invariant again , then such an intersection is the minimal virtually invariant space containing @xmath135 .
clearly a virtually invariant minimal space containing @xmath135 need not be irreducible : it might contain a virtually invariant subspace subspace not containing @xmath137 .
consider the abelian integral @xmath138 , where @xmath135 is a locally constant section of the homology bundle and @xmath139 is a meromorphic one - form as above .
denote by @xmath131 the minimal virtually invariant vector space containing @xmath135 .
[ th : algeq ] if @xmath131 is irreducible , then either the abelian integral @xmath31 vanishes identically , or its minimal annihilator is a linear differential operator of degree @xmath140 .
let @xmath141 be the complex vector space of germs of analytic functions in a neighborhood of @xmath11 , obtained from @xmath31 by analytic continuation along a closed path in @xmath142 .
it suffice to check that @xmath131 is isomorphic to @xmath141 .
equivalently , for every locally constant section @xmath143 we must show that @xmath144 .
indeed , the vector space of all locally constant sections @xmath145 with @xmath146 is an invariant subspace of @xmath131 . as @xmath131 is supposed to be irreducible , then this space is trivial .
theorem [ th : algeq ] follows from theorem [ th : main1 ] .
the above theorem is easily generalized .
for instance , the coefficients of the minimal annihilator of @xmath80 are rational functions of @xmath11 if and only if the minimal virtually invariant space @xmath131 containing @xmath147 is monodromy invariant , i.e. its orbit in the grassmannian consists of a single point .
further , it might happen that @xmath131 is reducible .
let @xmath148 be a proper virtually invariant subspace of @xmath131 . if the factor space @xmath149 is irreducible ( does not contain proper virtually invariant subspaces ) , then theorem [ th : algeq ] still holds true , but the minimal annihilator of @xmath31 is of order equal to @xmath150 .
multidimensional abelian integrals ( along @xmath1-cycles ) are studied in a similar way .
let @xmath121 be a real polynomial and @xmath151 a real polynomial differential one - form in @xmath34 .
consider the perturbed real foliation in @xmath34 defined by the infinitesimal 16th hilbert problem asks for the maximal number of limit cycles of ( [ 16th ] ) when @xmath153 as a function of the degrees of @xmath154 .
let @xmath155 be a continuous family of closed orbits of ( [ 16th ] ) .
the zeros of the abelian integral @xmath156 approximate limit cycles ( at least far from the atypical points of @xmath121 ) in the following sense . if @xmath157 , then a limit cycle of ( [ 16th ] ) tends to the oval @xmath158 , when @xmath159 tends to @xmath160 .
the question of explicit computing the number of zeros of abelian integrals remains open ( although a substantial progress was recently achieved , see @xcite and the references therein ) . generically an abelian integral satisfies a picard - fuchs differential equation @xmath161 of order equal to the dimension of the homology group of the typical fiber @xmath38 .
we are interested in the possibility of reducing the degree of this equation , assuming that the coefficients of the equation are algebraic in @xmath11 , @xmath162 .
indeed , the zeros of the solutions of a second order equation are easily studied ( by the rolle s theorem ) . in this section
we study abelian integrals which appear in the perturbations of foliations @xmath163 with @xmath164 and @xmath165 , where @xmath36 are positive integers .
the corresponding foliation @xmath163 is a special lotka - volterra system .
consider first the fibration defined by the polynomial @xmath166 .
we assume that @xmath36 are _ relatively prime_. the base of the fibration is the punctured plane @xmath167 .
each fiber is a sphere with two points removed .
the homology bundle is one - dimensional with trivial monodromy representation .
we investigate the monodromy representation on the _ relative homology _ bundle .
it will be a basic ingredient of the monodromy investigation in more complicated cases .
consider a set of marked points @xmath168 on the complex fibre @xmath38 @xmath169 where @xmath0 is a real positive number .
the relative homology @xmath170 is a free group with @xmath39 generators . a convenient model for the pair @xmath171 consists of a cylinder with some strips attached ; marked points are located at the ends of these strips .
note that there exist a unique pair of positive integers @xmath172 satisfying the following relation @xmath173 let @xmath174 be the strip in complex plane around the real segment @xmath175 $ ] .
let @xmath176 be ring ( cylinder ) which radii @xmath177 and @xmath178 satisfy relations @xmath179 .
the model @xmath180 is a surface constructed with three charts @xmath181 , @xmath182 , @xmath183 : @xmath184 with the following transition functions ( strips @xmath181 are attached to the external circle of radius @xmath178 and strips @xmath182 are attached to the internal boundary of @xmath183 ) @xmath185 the marked points are @xmath186 and @xmath187 at the end of strips . to construct a map
@xmath188 we will use a bump function @xmath189)$ ] which is 0 near @xmath190 and 1 near @xmath191 .
the map @xmath192 reads @xmath193 complex level curves @xmath38 intersect line at infinity in two points : @xmath197 $ ] and @xmath198 $ ] .
the neighborhood of any of them is a punctured disc .
thus , there exists an isotopy of the level curve @xmath38 shrinking it to the region @xmath199 for sufficiently big @xmath178 .
we will assume that @xmath11 is sufficiently close to @xmath200 .
the intersection of @xmath38 with the neighborhood @xmath201 of @xmath202 is a cylinder parametrized by the formula @xmath203 where @xmath204 is a function which will be fixed later .
the intersection of @xmath38 with set @xmath205 decomposes into two connected components @xmath206 and @xmath207 ; one is located close to the @xmath208-plane and another to the @xmath209-plane respectively .
the component @xmath206 is a graph of multi - valued ( @xmath210-valued ) function @xmath211 defined over the ring @xmath212 .
marked points are images of point @xmath213 located on the real axis .
we deform this domain by isotopy to the strip @xmath214 along real line
see figure [ fig : toymod ] .
the model of @xmath207 is constructed in an analogous way . to glue the above map together with parametrization of disk around zero
, we use the auxiliary function @xmath217 .
it must be equal @xmath218 near the internal circle of the ring @xmath219 ( i.e. @xmath220 ) and @xmath221 near the exterior boundary ( @xmath222 ) .
it is easy to check that @xmath223 solves the problem . a 2-dimensional version of figure [ fig : toymon3d ] presenting the model surface @xmath180 is drawn below .
it is obtained from figure [ fig : toymon3d ] by cutting the cylinder along a vertical line .
we will use these planar style of drawing models in subsequent , more complicated cases .
the segments that are unified are marked with arrows .
strips @xmath181 and @xmath182 are enumerated by integers @xmath225 and @xmath226 respectively ; the argument @xmath227 is calculated in point @xmath228 which is glued with the point @xmath229 according to relations .
generators of the relative homology of @xmath180 are also marked .
[ prop : toyhom ] the relative homology is @xmath170of the complex fiber @xmath38 has dimension @xmath39 .
it is generated by cycles @xmath230 with the relations : @xmath231 the monodromy representation on the relative homology space reads : @xmath232 consider the fibration given by a polynomial @xmath233 , where @xmath36 is a pair of positive , relatively prime integer numbers .
thus , they satisfy the relation with a pair of integers @xmath234
. they must be of opposite signs ; we assume @xmath235 and so @xmath236 .
the base of locally trivial fibration in this case is a plane with two points removed @xmath237 , where @xmath238 corresponds to a center @xmath239 of the hamiltonian vector field @xmath240 .
the cycle @xmath241 for @xmath242 is an oval ( compact component ) of the real level curve @xmath38 .
the model of complex fiber is presented on figure [ fig : parab ] .
it consists of two cylinders and @xmath39 strips glued together as shown on the figure .
cylinders are drown as rectangles , with horizontal edges unified . to simplify the combinatorial structure , there are opposite orientations on these two cylinders .
vertical , dotted lines mark another unification .
[ lem : parabmodel ] the surface shown on the figure [ fig : parab ] provides a model @xmath180 for complex fiber @xmath38 .
the homology group @xmath243 has dimension @xmath37 and is generated by cycles @xmath244 with the following relations @xmath245 intersection indices of @xmath147 and other generators of the homology group reads @xmath246 the monodromy around zero takes the form : @xmath247 the idea of proof is similar to the proof of lemma [ lem : toy ] .
we shrink the level curve @xmath38 by isotopy to the region @xmath199 .
we take the value of @xmath11 sufficiently close to 0 .
the intersections of @xmath38 with neighborhoods of saddle points @xmath248 are cylinders ; we parametrize them by formulas similar to .
the remaining part of the fiber @xmath38 splits into two pieces : @xmath249 and @xmath250 , located close to the line @xmath251 and close to the parabola @xmath252 respectively .
the part @xmath249 is the graph of @xmath253-valued function @xmath254 defined over the disc of radius @xmath178 with small discs around points @xmath255 removed : @xmath256 we deform the domain @xmath257 to the strip @xmath214 along the real segment see figure [ fig : pardeform ] .
leaves of function over the strip @xmath214 are numbered by @xmath258 . in
analogous way we deform @xmath250 to the graph of the @xmath210-valued function defined over the strip along real segment of the parabola @xmath259 .
leaves of the function are numbered by @xmath216 .
we glue together both collections of strips with two cylinders in a way analogous to the toy example . indeed , in sufficiently small neighborhood of the point
@xmath260 the pair of functions @xmath261 define a holomorphic chart . in this chart
the function @xmath121 takes the form as in the toy example .
the same is true for the other saddle @xmath262 .
both cylinders are glued by @xmath253 strips going along the line @xmath251 and @xmath210 strips going along parabola @xmath263 .
the monodromy around zero permutes strips according to the rule @xmath264 , @xmath265 , what is compatible with the formula for monodromy in the toy example case .
thus , the monodromy acts on both cylinders around saddles as in the toy example .
the surface shown on figure [ fig : parab ] provides a model for complex fibre @xmath38 .
formulas follow from respective formulas in the toy example . the relations and intersection indices one can read from the figure [ fig : parab ] .
+ the critical value @xmath267 corresponds to a morse critical point of @xmath121 .
the monodromy operator @xmath268 around @xmath269is therefore described by the usual picard - lefschetz formula .
let @xmath270 be the continuous family of cycles , vanishing @xmath269 .
the proof is based on theorem [ th : algeq ] .
let @xmath274 be a @xmath1-dimensional subspace of the ( complex ) homology space @xmath275 and @xmath276 .
assume that the monodromy orbit of @xmath274 in the grassmannian @xmath277 is finite .
we show that the dimension of @xmath274 satisfies @xmath278 .
let @xmath279 be the operator of monodromy around @xmath224 ( i.e. along a loop winding once around @xmath224 ) ; let @xmath268 be a monodromy around the center @xmath267 .
it follows formulas that the @xmath280 is nilpotent operator and its image is one dimensional , generated by @xmath147 . the homology space @xmath281 splits into 2-dimensional @xmath268-invariant subspace @xmath282 and @xmath283-dimensional ; the monodromy @xmath268 restricted to the latter one is the identity .
the matrix of the restricted monodromy operator @xmath284 in a basis @xmath285 has the form @xmath286_{(\gamma,\delta)}=\left(\begin{smallmatrix}1&1\\ 0&1\\ \end{smallmatrix}\right).\ ] ] note that the subspace @xmath282 is not defined uniquely .
it is spanned by @xmath147 and any element @xmath287 such that @xmath288 .
consider the intersection @xmath289 .
the property that @xmath274 has a finite @xmath290 orbit ( see theorem [ th : algeq ] ) implies that the @xmath268-orbit of @xmath291 , @xmath292 , is finite .
thus , the intersection @xmath293 has also finite @xmath268 orbit in @xmath282 .
the form of @xmath284 implies that there are only 3 subspaces with a finite orbit : @xmath294 note that all these subspaces are @xmath268-invariant .
take @xmath299 and consider 2-dimensional , @xmath268 invariant space @xmath282 spanned by @xmath147 and @xmath300 .
the @xmath268 orbit of the @xmath301 space is finite , so the intersection @xmath302 has one of three forms listed in . since @xmath303 then @xmath304 and so @xmath305 .
+ consider the @xmath279-orbit of @xmath276 .
the lemma [ lem : ninvcnd ] ( for @xmath306 ) and the property that the intersection number is preserved by the monodromy implies the following condition @xmath307 consider an element @xmath308 . since @xmath309 , then @xmath310 .
consider elements @xmath311 for @xmath312 .
the intersection index @xmath313 .
so , the cycles @xmath314 have a nonzero intersection indices with @xmath147 .
since @xmath36 are relatively prime the space spanned sums @xmath315 coincide to the space generated by @xmath316 , @xmath317 , ; the latter one is the full space generated by @xmath318 .
both claims follow the fact that @xmath253 and @xmath86 are also relatively prime ( see ) and the following observation [ linalg ] let @xmath111 be a vector space of dimension @xmath253 and let @xmath210 be an integer .
assume that @xmath36 are relatively prime .
let @xmath319 be a basis of @xmath111 .
then the following sums @xmath320 ( all indices @xmath321 assumed ) generate the whole space @xmath111 .
the proof of lemma is based on the following observations .
since @xmath36 are relatively prime , any sum of length @xmath210 appears in a sequence .
the difference of two sums has the form @xmath322 , @xmath323 ; they generate a hyperplane orthogonal to vector @xmath324 .
since the scalar product @xmath325 , the space generated by vectors is a whole @xmath111 .
we have shown that the subspace of the homology group containing @xmath147 , with finite @xmath290-orbit must be necessarily @xmath290-invariant hyperplane in the homology space @xmath281 .
it proves the theorem for a _ generic _
( when the zero subspace @xmath327 ) .
to finish the proof we show that either @xmath328 or @xmath329 .
consider an element @xmath330 and its images under the monodromy around @xmath224 : @xmath331 . since @xmath332 is monodromy invariant , all elements @xmath333 . if the intersection index @xmath334 , then monodromy around the center @xmath267 adds a multiple of @xmath147 , so @xmath335
. then , it follows from the previous analysis that @xmath336 .
assume now that all intersection indices @xmath337 .
the coefficient @xmath338 must then vanish , otherwise @xmath339 adds the cycle @xmath340 which realizes intersection index @xmath341 . consider monodromies @xmath342 , @xmath343 .
it preserves the expression @xmath344 .
vanishing of the intersection indices @xmath345 implies equations @xmath346 the solution of depends on the parity of @xmath253 . if @xmath253 is odd then all coefficients @xmath347 are equal : @xmath348 .
if @xmath253 is even the solution of reads : @xmath349 we repeat then the analogous analysis with iterations of @xmath350 .
we obtain the following form of @xmath332 @xmath351 thus , @xmath352 and so the theorem is proved .
+ we have actually shown that the abelian integral does not satisfy any differential equation with algebraic coefficients of order lower than the fuchs type equation with rational coefficients which follows the general theory .
consider a fibration given by a polynomial @xmath165 .
it defines a locally trivial fibration defined over plane with two points removed @xmath353 , where @xmath354 corresponds to a center .
the cycle @xmath241 for @xmath242 is an oval ( compact component ) of the real level curve @xmath38 .
note , that the fibration has a morse type singularity at @xmath267 and @xmath241 is a vanishing cycle at the center .
below we investigate the fibration and the monodromy representation on the sufficiently small neighborhood of @xmath224 : @xmath355 .
the monodromy around center @xmath267 follows the picard - lefshetz formula .
thus , to determine the monodromy representation it is enough to investigate the monodromy around @xmath224 and intersection indices with the cycle @xmath147 . the model of complex fiber is presented on figure [ fig : hvg ] which should be understood as follows .
any rectangle represents a cylinder ; unification of edges according to arrows assumed .
another unification is assumed on vertical , dotted lines .
[ lem : hvgmodel ] the complex level curve @xmath38 is a surface of genus @xmath356 with @xmath357 points removed ( intersection with the line at infinity ) . the surface shown on the figure [ fig : hvg ]
provides a model @xmath180 for @xmath38 .
the homology group @xmath243 has dimension @xmath30 ; it is generated by cycles @xmath358 with the following relation @xmath359 intersection indices of @xmath147 with other generators of the homology group reads @xmath360 the monodromy representation takes the form ( monodromy around zero assumed ) @xmath361 the proof is analogous to proofs of lemma [ lem : toy ] and [ lem : parabmodel ] .
we modify the level curve @xmath38 by isotopy to the part contained in the compact region @xmath362 .
we consider points @xmath11 sufficiently close to @xmath200 .
we cut the level curve @xmath38 into pieces lying close to lines @xmath251 ,
@xmath363 , @xmath364 and close to saddles @xmath365 .
the analysis of pieces of level curve @xmath38 close to the saddles @xmath366 and @xmath260 and close to the line @xmath364 is completely analogous to the parabola case
see proof of lemma [ lem : parabmodel ] .
the model of this part of the level curve consist of two cylinders joined by a single strip .
now we consider the region which is in finite distance from the line @xmath364 .
the level curve @xmath38 outside the line @xmath364 splits into @xmath253 components defined by the equation : @xmath367 any of these components coincide to the toy example with @xmath368 .
thus , it is isotopic to the cylinder with two strips attached .
as @xmath11 winds around zero @xmath369 we rotate components according to the rule @xmath370 .
the @xmath253-th power of the monodromy ( winding @xmath253 times around zero ) @xmath371 corresponds to the usual monodromy in the toy example ; it follows from the formula ( for @xmath368 ) that @xmath371 adds the generator @xmath372 .
[ pr : hvg ] let @xmath274 be the following 2-dimensional subspace @xmath373 of the ( complex ) homology space @xmath374 .
the orbit of @xmath274 under the monodromy representation @xmath375 in grassmannian @xmath376 consists of @xmath253 elements and so is finite .
denote by @xmath93 and @xmath268 the monodromy around @xmath224 and around the center critical value @xmath267 . by lemma [ lem : hvgmodel ]
, we have @xmath377 the crucial observation is that subspaces @xmath378 for @xmath292 are @xmath268-invariant . indeed , the subspace @xmath274 is @xmath268-invariant since @xmath276 is a vanishing cycle corresponding to the center critical value @xmath267 .
we calculate ( using formulas ) the intersection indices of @xmath147 and generators of @xmath378 for @xmath379 : @xmath380 thus , both generators are @xmath268-invariant .
it proves that the orbit @xmath375 in grassmannian @xmath376 consists of @xmath253 subspaces given in formula .
+ the proposition [ pr : hvg ] provides a geometric explanation of phenomenon described in @xcite . according to the general theory given in theorem [ th : algeq ] and calculations of monodromy given in lemma [ lem : hvgmodel ] and proposition [ pr : hvg ] the abelian integral along cycle @xmath147 satisfies a linear second order equation with algebraic coefficients in variable @xmath11 .
99 v. i. arnold , arnold s problems , springer , 2004 . on the number of zeros of abelian integrals : a constructive solution of the infinitesimal hilbert sixteenth problem , gal binyamini , dmitry novikov , sergei yakovenko , arxiv:0808.2952v3 [ math.ds ] s. a. van gils , e. horozov , _ uniqueness of limit cycles in planar vector fields which leave the axes invariant _ , contemp . math .
* 56 * ( 1986 ) , 117 129 .
centennial history of hilbert s 16th problem .
, 39(3):301354 ( electronic ) , 2002 .
s. lang , algebra , addison - wesley , 1994 .
a. r. magid , lectures on differential galois theory , university lecture series vol .
7 , ams , 1994 .
m. van der put , galois theory of differential equations , algebraic groups and lie algebras , j. symbolic computations * 28 * ( 1999 ) , 441 - 473 .
m. van der put , m. f. singer , galois theory of linear differential equations , 2002 . | let @xmath0 be a linear differential operator with coefficients in some differential field @xmath1 of characteristic zero with algebraically closed field of constants .
let @xmath2 be the algebraic closure of @xmath1 .
for a solution @xmath3 , @xmath4 , we determine the linear differential operator of minimal degree @xmath5 and coefficients in @xmath6 , such that @xmath7 . this result
is then applied to some picard - fuchs equations which appear in the study of perturbations of plane polynomial vector fields of lotka - volterra type . |
gathering information on metal abundance and abundance ratios of many galactic clusters located in different region of the disk and with different ages is mandatory to study the chemical evolution of the galactic disk .
this , in turn , provides us with hints on the formation mechanism of the disk and its relation with the other major components of the galaxy , the halo and bulge ( janes & phelps 1994 , carraro et al .
+ however , a frequent , well known problem in the study of open star clusters is the stellar contamination from the general galactic disk field , which complicates the analysis of the color magnitude diagram ( cmd )
. it also makes it difficult to derive fundamental cluster parameters , especially metallicity , when only a few stars are observed . +
a notorious example in this context is the nearby old open cluster ngc 2112 ( collinder 76 , c 0551 - 0031 , ocl 509 ) , which has a reputation of suffering from heavy field star contamination ( brown et al .
1996 ) , and for this reason its basic parameters remain poorly constrained .
+ the first investigation on ngc 2112 was carried out by richtler ( 1985 , hereinafter r85 ) , who obtained photographic bv photometry for about 80 stars down to v= 15 .
although his photometry barely reaches the cluster turn - off ( to ) , he nevertheless drew attention to this probably old , so far neglected cluster , and he suggested that ngc 2112 has a reddening of @xmath3 0.5 mag and lies @xmath3800 pc from the sun . by analyzing additional stromgren photometry , he proposed that the cluster had to be very metal poor ( [ fe / h ] as low as -1.4 ) . + a more accurate and deeper analysis
was performed a few years later by richtler & kaluzny ( 1989 ) .
they obtained bv ccd photometry for about 500 stars in a field of 200 arcmin@xmath4 .
additionally , they obtained moderate resolution spectra for a handful of bright stars .
their conclusions were that the cluster was very contaminated by field stars .
nevertheless , they were able to strengthen the suggestions of richtler ( 1985 , hereinafter r85 ) by confirming that the cluster is indeed old ( 3 - 5 gyrs ) , is located 700@xmath5800 pc from the sun and has a reddening of e(b - v)=0.60 mag . lacking a membership analysis ,
no further information on the cluster metallicity was provided .
+ photometry in the washington system by geisler(1987 ) and geisler , claria & minniti ( 1991 ) seems to confirm the previous suggestions that ngc 2112 is very metal poor ( [ fe / h ] as low as -1.3 ) .
+ more recently , three spectroscopic campaigns have been carried out in the field of the cluster .
friel & janes ( 1993 , hereinafter fj93 ) present moderate resolution spectra of 6 stars . out of these ,
5 are considered members , and an average @xmath6 $ ] of [email protected] has been found .
this is significantly larger than all the previous determinations .
brown et al . ( 1996 , hereinafter bwgo96 ) used the blanco echelle on ctio to obtain high - resolution spectroscopy of 6 stars .
they accepted just one star as a definitive member , providing an even higher metal content value , [ fe / h]=-0.15 , only slightly lower than the solar value .
finally , mermilliod & mayor ( 2007 , hereinafter mm07 ) enlarged the sample of spectroscopic members to 3 , out of 6 stars observed with coravel .
they underlined the need to obtain radial velocities for a larger sample down to v @xmath7 14.5 to better probe the shape of the red giant branch ( rgb ) . at the same time
, this would also provide much firmer metallicity estimates .
+ in carraro et al .
( 2002 ) we reported a jonhson bvi photometry of the cluster down to magnitude v=20 . assuming the metal content found by brown et al .
( 1996 ) , we found a reddening of e(b - v)[email protected] , a distance of 850@xmath2100 pc , and an age of [email protected] gyr .
we argued on a purely photometric basis that the metallicity can not be much lower than bwgo96 value , and we stressed the need for a new , more detailed spectroscopic investigation of the cluster .
+ in this paper we attempt such an investigation by securing the deepest and widest field coverage multicolor ( @xmath1 ) photometry of ngc 2112 to date . at the same time , we provide moderate and high resolution spectroscopy of 40 stars . with these data at hand , we present robust and updated determinations of the cluster s basic properties . 593 , which has @xmath8 = 05:53:43.75 and @xmath9= + 00:23:56.4 . with filled circles
we indicate stars observed spectroscopically . ]
in this work we present photometry and spectroscopy in the field of ngc 2112 obtained with three different telescopes .
for this reason , the details of data acquisition and reduction are presented in the next three sub - sections .
@xmath10 and @xmath11 images centered on ngc 2112 were obtained at the cerro tololo inter - american observatory 1.0 m telescope , which is operated by the smarts consortium .
the telescope is equipped with a new 4k@xmath124k ccd camera having a pixel scale of [email protected]/pixel , which allows one to cover a field of @xmath0 .
this allows us to cover the entire cluster , which has an estimated diameter of 18 arcmin ( dias et al .
+ observations were carried out on november 30 , 2005 .
three landolt ( 1992 ) areas ( tphoenix , rubin 149 , and pg 0231 + 006 ) were also observed to calibrate the instrumental magnitudes to the standard system .
the night was photometric with an average seeing of 1.1 arcsec .
data were reduced using iraf packages ccdred , daophot , and photcal .
photometry was done employing the point spread function ( psf ) method ( stetson 1987 ) .
the covered area is shown in fig . 1 , while table 1 contains the observational log .
+ the calibration equations read : .journal of photometric observations of ngc 2112 and standard star fields together with calibration coefficients ( november 30 , 2005 ) . [ cols="<,^,^,^,^,^,^ " , ]
now that we have determined the cluster s fundamental parameters , we can use the best fit isochrone as a tool to investigate possible binaries among cluster members from a purely photometric point of view . in fig 10
we provide a zoom of the to region , and indicated with filled circles ( color coded in red ) radial velocity members .
the solid line is the best fit padova isochrone ( see fig .
the dashed curve is the same isochrone , but shifted up by 0.75 mag , to illustrate the locus of unresolved equal mass binaries .
this exercise is suggesting us that several stars that lie off the best fit isochrone may in reality be unresolved binaries .
in fact some of them lie very close to the binary locus , and we indicate them as filled squares ( color coded in green ) in the cmd of fig . 10 .
they are stars @xmath14707 , 824 , 782 and 601 . at the same time
, we indicated as filled triangles ( blue when printed in color ) stars which lie neither in the single star nor in the binary star sequence .
these are stars @xmath14714 , 707 , and 323 .
one may speculate that these are unequal mass binaries or maybe systems with more than two components .
if we refer only to the to region , we are left with 14 single member stars and 7 possibly multiple systems , which implies a rough binary percentage of 33@xmath15 .
in this paper we presented new photometric and spectroscopic data in the field of the old open cluster ngc 2112 .
this new dataset allowed us to revise the cluster s fundamental parameters and clarify a long lasting debate on its properties , which for many years have been poorly constrained due to the high level of field star contamination .
+ by means of multi - fiber spectroscopy , we measured radial velocity for 40 stars and found 21 radial velocity members .
this , in turn , allowed us to clean the cmd , providing a better comparison with stellar models .
+ the most important result of our study is that the cluster has a metallicity much higher than previous determinations , and somewhat higher than the sun .
also the @xmath16 elements are marginally enhanced with respect to the sun , but still compatible with the trends of thin disk stars in the solar vicinity , as are all the other elements we measured .
+ therefore ngc 2112 is typical of old , thin disk star clusters , as metal rich as the hyades ( boesgaard & friel 1990 ) , and located at less than 1 kpc from the sun in the anticenter direction .
+ in addition , we confirmed the age of the cluster ( @xmath3 1.8 gyr ) , previously derived on a purely photometric basis ( carraro et al .
2002 ) . as for the distance and reddening , our values are in agreement within the errors of previous determinations . +
this study stresses the importance of performing detailed membership analysis in galactic open clusters in order to derive more robust estimates of their fundamental parameters .
this research was part of a joint project between universidad the chile and yale university , partially funded by the fundacin andes .
hydra observations were performed remotely from yale by mvm .
cmb acknowledges universidad de chile graduate fellowship support from program mece educacin superior uch0118 and fundacin andes c-13798 .
mvm was supported by an nsf astronomy and astrophysics postdoctoral fellowship , under award ast04 - 011640 .
gc thanks doug geisler , tom richtler and roberto barbon for long and interesting discussions on ngc 2112 , and the anonymous referee who helped us to improve the paper presentation .
this study made use of simbad and webda .
99 alonso , a. , arribas , s. & martnez - roger , c. 1999 , a&as , 140 , 261 bensby , t. , feltzing , s. , lundstrm , i. , ilyin , i. , 2005 , a&a 433 , 185 boesgaard , a.m. , friel , e.d .
, 1990 , apj 351 , 467 brown j.a . , wallerstein g. , geisler d. , & oke j. b. 1996 , aj , 112 , 1551 carraro , g. , ng , y.k . , & portinari , l. 1998 , mnras , 296 , 1045 carraro , g. , barbon , r. , & boschetti , c.s .
2002 , mnras , 336 , 259 carraro g. , villanova , s. , demarque , p. , mcswain , m.v . ,
piotto , g. , bedin , l.r . , 2006 ,
apj 643 , 1151 carraro g. , geisler , d. , villanova , s. , frinchaboy , p.m. , majewski , s.r . , 2008 ,
a&a 476 , 217 demarque p.,wo , y - h . , kim , y - c .
, yi , s.k . 2004 , apjs 115 , 667 dias , w.s . , alessi , b.s . , moitinho , a. , lepine , j.r.d .
2002 , a&a 389 , 871 friel e.d .
, janes k.a . , 1993 , a&a 267 , 75 friel , e.d . ,
jacobson , h.r . , barrett , e. , fullton , l. , balachendran , s.c . , pilachowski , c.a , 2003 , aj 126 , 2372 geisler d. 1987 , aj , 94 , 84 geisler d. , claria j.j .
, minniti .
d 1991 , aj , 102 , 1836 girardi l. , bressan a. , bertelli g. , & chiosi , c. 2000 , a&as , 141 , 371 gratton r. g. , carretta e. , eriksson k. , & gustafsson b. 1999 , a&a , 350 , 955 houdashelt m.l .
, bell r.a . , &
sweigart a.v .
2000 , aj , 119 , 1448 horne , k. , 1986 , pasp 98 , 60 karatas y. , schuster w.j . , 2006 ,
mnras 371 , 1793 kim , y .- c . , demarque , p. , alexander , d.r .
, 2002 , apjs 143 , 499 kurucz r.l . ,
furenlid i. , brault j. , & testerman , larry 1984 , sfat.book , k kurucz r.l .
1992 , iaus , 149 , 225 janes , k.a . ,
phelps , r.l .
, 1994 , aj 108 , 1773 landolt , a.u . , 1992 , aj 104 , 340 mermilliod , j .- c . , mayor , m. , 2007 , a&a 470 , 919 richtler t. 1985 , a&as , 59 , 491 richtler t. , & kaluzny j. 1989 , a&as , 81 , 225 schlegel d.j . , finkbeiner d.p .
, & davis m. 1998 , apj , 500 , 525 schmidt - kaler , th . , 1982 , in schaifers k. , voigt , h.h .
, eds . , numerical data and functional relationships in science and technology , landolt - brbstein , new series , group vi , vol 2(b ) , springer , berlin , p. 14 stetson , p.,b . , 1987 , pasp 99 , 191 zacharias , n. , urban , s.e .
, zacharias , m.i . ,
wycoff , g.l . , hall , d.m . ,
monet , d.g . , rafferty , t.j . , 2004 , aj 127 , 3043 yi , s.k , demarque p.,kim , y - c . , lee , y .-
w , ree , c.h . , lejeune , th . , barnes , s. , 2001 , apjs 136 , 417 | we report on a new , wide field ( @xmath0 ) , multicolor ( @xmath1 ) , photometric campaign in the area of the nearby old open cluster ngc 2112 . at the same time
, we provide medium - resolution spectroscopy of 35 ( and high - resolution of additional 5 ) red giant and turn off stars .
this material is analyzed with the aim to update the fundamental parameters of this traditionally difficult cluster , which is very sparse and suffers from heavy field star contamination . among the 40 stars with spectra , we identified 21 _ bona fide _ radial velocity members which allow us to put more solid constraints on the cluster s metal abundance , long suggested to be as low as the metallicity of globulars .
as indicated earlier by us on a purely photometric basis ( carraro et al .
2002 ) , the cluster [ fe / h ] abundance is slightly super - solar ( [ fe / h ] = [email protected] ) and close to the hyades value , as inferred from a detailed abundance analysis of 3 of the 5 stars with higher resolution spectra .
abundance ratios are also marginally super solar . + based on this result , we revise the properties of ngc 2112 using stellar models from the padova and yale - yonsei groups .
+ for this metal abundance , we find the cluster s age , reddening , and distance values are 1.8 gyr , 0.60 mag , and 940 pc , respectively .
both the yale - yonsei and padova models predict the same values for the fundamental parameters within the errors .
+ overall , ngc 2112 is a typical solar neighborhood , thin disk star cluster , sharing the same chemical properties of f - g stars and open clusters close to the sun .
+ this investigation outlines the importance of a detailed membership analysis in the study of disk star clusters .
[ firstpage ] open clusters and associations : general open clusters and associations : individual : ngc 2112 |
since the degrees of freedom for high reynolds number ( @xmath9 ) turbulence , as can be encountered in geophysical and astrophysical flows , can be very large , the implementation of their numerical modeling can easily exceed technological limits for computations .
furthermore , since truncation of the omitted scales removes important physics , e.g. , of multi - scale interactions , the only approach to a numerical study of such flows is to employ subgrid modeling of those scales .
this is frequently accomplished with large eddy simulations ( les see @xcite for recent reviews ) .
this is of importance for geophysical , astrophysical and engineering applications and can have consequences for meteorological @xcite and climate prediction simulations @xcite , for instance . while realistic reynolds numbers will remain out of reach for the foreseeable future , subgrid
modeling can be an extremely useful tool in the computation of simulations for such applications .
the incompressible lagrangian - averaged navier - stokes equations ( lans@xmath1 , @xmath0model , or also the viscous camassa - holm equation ) @xcite is one possible subgrid model .
it can be derived , for instance , by temporal averaging applied to hamilton s principle ( where taylor s frozen - in turbulence hypothesis is applied as the closure , and also as the only approximation of the derivation ) @xcite .
for this reason , the momentum - conservation structure of the equations are retained . for scales smaller than the filter width
, lans@xmath1 reduces the steepness of steep gradients of the lagrangian mean velocity and limits how thin vortex tubes become as they are transported ( the effect on larger length scales is negligible ) @xcite .
the @xmath0model may also be derived from smoothing the transport velocity of a material loop in kelvin s circulation theorem @xcite .
consequently , there is no attenuation of resolved circulation , which is important for many engineering and geophysical flows where accurate prediction of circulation is highly desirable .
lans@xmath1 has previously been compared to direct numerical simulations ( dns ) of the navier - stokes equations at modest taylor reynolds numbers ( @xmath10 @xcite , @xmath11 @xcite , and @xmath12 @xcite ) .
lans@xmath1 was compared to a dynamic eddy - viscosity les in 3d isotropic turbulence under two different forcing functions ( @xmath13 and @xmath14 ) and for decaying turbulence with initial conditions peaked at a low wavenumber ( @xmath15 ) and at a moderate wavenumber ( @xmath16 ) @xcite .
lans@xmath1 was preferable in that it demonstrated correct alignment between eigenvectors of the subgrid stress tensor and the eigenvectors of the resolved stress tensor and vorticity vector .
lans@xmath1 and a related regularization , the leray model , were contrasted with a dynamic mixed ( similarity plus eddy - viscosity ) model in a turbulent mixing shear layer ( @xmath17 ) @xcite .
lans@xmath1 , was the most accurate of these three les , but the effects of numerical contamination can be strong enough to lose most of this potential .
this could pose some limitations on its practical use .
quantifying those limitations is one of the goals of this present work .
the @xmath0model also describes an incompressible second - grade non - newtonian fluid ( under a modified dissipation ) @xcite . in this interpretation
, @xmath6 is a material parameter which measures the elastic response of the fluid . either from this standpoint , from its status as a regularization of the navier - stokes equations , or , independently of any physically motivation , as a set of partial differential equations with proven unique regular solutions , we may analyze lans@xmath1 without any les considerations . analyzing inertial - range scaling for lans@xmath1 for moderate and large @xmath6 , as well as identifying different scalings at scales larger and smaller than @xmath6 is another of the goals of this work . in this context
we also study the numerical resolution requirements to obtain well - resolved solutions of lans@xmath1 ( i.e. , grid - independent solutions ) which leads to a verification of the predictions of the size of the attractor in lans@xmath1 @xcite .
section [ sec : details ] presents the lans@xmath1 model , our numerical experiments and technique . in section [ scaling ]
we analyze inertial - range scaling for lans@xmath1 . in section [ gridindep ]
we determine the numerical resolution requirements to obtain well - resolved solutions of lans@xmath1 . in section [ les ]
we address the les potential of lans@xmath1 by comparing @xmath0model simulations to a @xmath18 dns ( @xmath19 , @xmath20 ) , a @xmath21 dns ( @xmath22 , @xmath23 ) , a @xmath21 dns ( @xmath24 , @xmath25 ) , a @xmath26 dns ( @xmath27 , @xmath28 ) , and a @xmath29 dns ( @xmath2 , @xmath30 ) .
( the @xmath27 simulation has been previously described in a study of the imprint of large - scale flows on local energy transfer @xcite . ) in section [ results ] , we compare and contrast in more detail lans@xmath1 solutions with dns at @xmath31
. finally , in section [ sec : concl ] we summarize our results , present our conclusion , and propose future directions of investigation .
we consider the incompressible navier - stokes equations for a fluid with constant density , @xmath32 where @xmath33 denotes the component of the velocity field in the @xmath34 direction , @xmath35 the pressure divided by the density , @xmath36 the kinematic viscosity , and @xmath37 an external force that drives the turbulence ( in all results , the time , @xmath38 , is expressed in units of the eddy - turnover time ) .
the lans@xmath1 equations @xcite are given by @xmath39 where @xmath40 denotes the filtered component of the velocity field and @xmath41 the modified pressure .
filtering is accomplished by the application of a normalized convolution filter @xmath42 where @xmath43 is any scalar or vector field . by convention , we define @xmath44 . we choose as our filter the inverse of a helmholtz operator , @xmath45 .
therefore , @xmath46 where @xmath47 is the green s function for the helmholtz operator , @xmath48 ( i.e. , the well - known yukawa potential ) , or in fourier space , @xmath49 .
we solve eqs .
( [ eq : navier - stokes ] ) and ( [ eq : lanssc ] ) using a parallel pseudospectral code @xcite in a three - dimensional ( 3d ) cube with periodic boundary conditions . in most of the runs ,
we employ a taylor - green forcing @xcite , @xmath50 \label{eq : tgscale}\ ] ] ( generally , with @xmath51 ) , and employ dynamic control @xcite to maintain a nearly constant energy with time . this expression eq .
( [ eq : tgscale ] ) is not a solution of the euler s equations , and as a result small scales are generated fast when the fluid is stirred with this forcing .
the resulting flow models the fluid between counter - rotating cylinders @xcite and has been widely used to study turbulence , including studies in the context of the generation of magnetic fields through dynamo instability @xcite .
we also consider some runs with random and abc @xcite forcing .
we define the taylor microscale as @xmath52 and the mean velocity fluctuation as @xmath53 the taylor microscale reynolds number is defined by @xmath54 and the reynolds number based on a unit length is @xmath55 .
for lans@xmath1 , the @xmath56 norm is the quadratic invariant to be identified with the energy , @xmath57 where @xmath58 and @xmath59 as usual , we define the ( omni - directional ) spectral energy density , @xmath60 , from the relation @xmath61 where @xmath62 represents integration over the surface of a sphere .
the @xmath0model possesses a theorem corresponding to the krmn - howarth theorem @xcite for the navier - stokes equations and , as in the navier - stokes case , scaling of the inertial range energy spectra may be derived from it @xcite .
we summarize here the dimensional analysis argument for the lans@xmath1 inertial range scaling that follows from this theorem , beginning from equation ( 3.8 ) in ref .
we use the short notation @xmath63 , @xmath64 and @xmath65 . in the statistically isotropic and homogeneous case , without external forces and with @xmath66 , taking the dot product of eq .
( [ eq : lanssc ] ) with @xmath67 we can obtain the equation @xmath68 the trace of this equation is the fourier transform of the detailed energy balance for lans@xmath1 .
@xmath69 is the second - order correlation tensor while @xmath70 and @xmath71 are the third - order correlation tensors for lans@xmath1 and @xmath72 is the sub - filter scale stress tensor . for @xmath73
this reduces to the well - known relation derived by krmn and howarth .
the energy dissipation rate for lans@xmath1 , @xmath74 , satisfies @xmath75 . by dimensional analysis in eq .
( [ eq : holm38scale ] ) we arrive at @xmath76 for large scales such that @xmath77 , the second right hand term is ignored , @xmath78 , @xmath79 , and we arrive at the scaling of the four - fifths law , @xmath80 @xcite . here ,
@xmath81\cdot{\mathbf}{l}/l$ ] is the longitudinal increment of @xmath82 .
the four - fifths law expresses that the third - order longitudinal structure function of @xmath82 , @xmath83 , is given in the inertial range in terms of the mean energy dissipation per unit mass @xmath84 by @xmath85 or , equivalently , that the flux of energy across scales in the inertial range is constant .
we also obtain the kolmogorov 1941 @xcite ( hereafter , k41 ) energy spectrum , @xmath86 , or , equivalently , @xmath87 for small scales such that @xmath88 , however , @xmath89 and both right hand terms are equivalent in eq .
( [ eq : full_scalesc ] ) , and our scaling law becomes @xmath90 note that this scaling differs in a substantial way from the kolmogorov scaling ( @xmath91 ) . for our small scale energy spectrum
we then have @xmath92 where we used @xmath93 .
the energy spectrum for scales smaller than @xmath6 is then @xmath94 this spectrum can also be derived from phenomenological arguments originally introduced by kraichnan @xcite , and it differs from the navier - stokes spectrum due to the fact that the fluid is advected by the smoothed velocity @xmath95 which does not directly correspond to the conserved energy @xmath96 @xcite .
we test this prediction for lans@xmath1 scaling at a resolution of @xmath18 ( @xmath97 ) by moving both the forcing ( @xmath98 ) and @xmath6 ( @xmath99 ) to large scales in order to increase the number of resolved scales for which @xmath100 . in so doing
, we are assuming that the scaling for large @xmath6 is the same as for small @xmath6 and large @xmath101 ( for evidence to this effect , see @xcite ) .
confirmation as given by eq .
( [ eq : lcubelans ] ) is presented in fig .
[ fig : kminus1 ] where we plot @xmath102 as a function of @xmath103 ( by convention , we plot @xmath104 to reduce cancellation in the statistics ) . the scales identified with an inertial range
@xmath105 $ ] are marked by vertical dashed lines and the predicted scaling , @xmath4 , is indicated by a solid line .
we fit a scaling exponent ( @xmath106 ) and find @xmath107 .
this is significantly steeper than the classical kolmogorov scaling given by eq .
( [ eq : fourfifthslans ] ) ; it can thus be viewed as more consistent with the scaling given by eq .
( [ eq : lcubelans ] ) .
it is also more consistent with @xmath4 than with other possible lans@xmath1 scalings : under the assumption that the turnover time scale of eddies of size @xmath91 is determined by the unsmoothed velocity @xmath82 , we find @xmath108 , and if it is determined by @xmath109 , we find @xmath110 ( see , e.g. , refs .
the observed scaling corresponds to none of these cases , and is actually closer to an evaluation of the turnover time @xmath111 at the scale @xmath103 given by @xmath112 ( with @xmath113 ) . note that for 2d lans@xmath1 , however , it is the case that the scaling is determined by the unsmoothed velocity @xmath82 @xcite .
we note that this is one of many differences between the 2d and 3d cases ( e.g. , ideal invariants and cascades ) .
another difference , which we shall show in section [ results ] , is that in 2d vorticity structures decrease in scale as @xmath6 increases while in 3d there is a change in aspect ratio with structures getting both shorter and fatter .
this may , in fact , be related to the shallower lans@xmath1 energy spectrum for @xmath100 which we show in section [ results ] . while differences are observed between the scaling shown in fig .
[ fig : kminus1 ] and eq .
( [ eq : lcubelans ] ) , the error bars deny a k41 scaling ( as well as the @xmath114 and @xmath115 scalings ) at scales smaller than @xmath6 .
we believe the discrepancy between the observed and predicted scaling can be due to lack of resolution to resolve properly the inertial range at sub - filter scales .
we have less than a decade of inertial range and only @xmath18 points for the statistics . as more computational resources become available ,
this scaling should be re - examined . as a consequence of lans@xmath1 s taylor s frozen - in hypothesis closure , scales smaller than @xmath6 can phase - lock into coherent structures and be swept along by the larger scales ( see , e.g. , @xcite ) .
if we assume , formally , that this `` frozen - in turbulence '' takes the form of `` rigid bodies '' in the smoothed velocity field ( no stretching ) , we arrive at a much different spectrum than @xmath5 , eq . ( [ eq : lansa_spectrumsc ] ) .
all scales smaller than @xmath6 are subject to the frozen - in hypothesis and we expect to find such rigid bodies at these scales .
we note that collections of `` rigid '' portions of the flow ( rotating or non - rotating ) reduce the total degrees of freedom ( _ dof _ ) and make physical sense with lans@xmath1 s relation to second - grade fluids : these rigid bodies can be envisioned as polymerized portions of the fluid . as a matter of fact , in such structures all internal _ dof _ are frozen .
these `` rigid bodies '' follow as well from the consideration of lans@xmath1 as an initial value problem in fourier space , for which we have @xmath49 . in the limit as @xmath6 approaches infinity , all wavenumber ( and spatial ) dependence for @xmath116 is eliminated and the entire flow is advected by a uniform velocity field ( advection without internal degrees of freedom ) . for a rigid body there can be no stretching and , therefore , all the longitudinal velocity increments , @xmath117 , must be identically zero ( @xmath118 from basic mechanics with @xmath119 the rotation vector and , hence , @xmath120 ) .
note that in lans@xmath1 eq .
( [ eq : lanssc ] ) the @xmath121 term contributes only a rotation and not a stretching of @xmath95 .
such polymerization would have two consequences .
firstly , since there is no stretching , these rigid bodies would not contribute to the turbulent energy cascade , @xmath122 secondly , the energy spectrum from dimensional analysis ( @xmath123 , for large @xmath124 : @xmath125 , and @xmath126 ) is @xmath127 this is , in fact , the observed lans@xmath1 spectrum for @xmath128 as is shown in fig .
[ fig : norms ] .
we verified that the spectrum is not the result of under - resolved runs , as is the case , e.g. , in the @xmath129 spectrum observed in truncated euler systems @xcite or in extremely under - resolved spectral simulations of the navier - stokes equations .
indeed , equipartition of the energy among all modes in a truncated euler@xmath1 system should also lead to a @xmath129 spectrum .
along with several experiments with different viscosities and also with statistically homogeneous and isotropic forcing ( not shown here ) , these are assurances that the observed spectrum is not a result of inadequate numerical resolution .
it should be noted that this is the same computation for which the third - order structure function is shown in fig .
[ fig : kminus1 ] .
the third - order structure function is consistent with a @xmath4 scaling ( corresponding to a @xmath5 energy spectrum ) while the spectrum itself is @xmath130 .
( also shown in fig .
[ fig : norms ] are the @xmath131 and the @xmath132 norms which ( through @xmath133 for @xmath128 ) correspond to @xmath5 and @xmath134 spectra , respectively . since the analytical properties of the lans@xmath1 solution are based on the energy balance , @xmath135 , in the @xmath56 norm , we employ this norm for all following results . )
these two different scalings , @xmath4 and @xmath130 , are consistent with a picture where a fluid has both rigid - body portions at scales smaller than @xmath6 ( wherein there is no turbulent cascade ) and spatial regions between these where the cascade does take place . for the structure functions , a non - cascading rigid body does not contribute to the scaling and consequently the cascading contribution , eq .
( [ eq : lcubelans ] ) , dominates .
the energy spectrum , however , for the limit of @xmath101 very large , is dominated by the @xmath136 term , and hence the @xmath5 component is subdominant . for lans@xmath1 and @xmath137 for dns . for all black pixels , the cubed longitudinal increment is less than @xmath138 ( approximately consistent with rigid bodies ) .
on the top is the large-@xmath6 simulation ( @xmath98 , @xmath139 , @xmath97 ) where the filling factor ( computed over the entire 3d domain ) is 0.67 . on the bottom
is a dns of navier - stokes ( @xmath51 , @xmath140 ) where the filling factor is 0.26 .
thus , a much greater portion of the flow is consistent with collections of rigid bodies for the large@xmath1 simulation.,title="fig:",width=338 ] + for lans@xmath1 and @xmath137 for dns . for all black pixels ,
the cubed longitudinal increment is less than @xmath138 ( approximately consistent with rigid bodies ) . on the top
is the large-@xmath6 simulation ( @xmath98 , @xmath139 , @xmath97 ) where the filling factor ( computed over the entire 3d domain ) is 0.67 . on the bottom
is a dns of navier - stokes ( @xmath51 , @xmath140 ) where the filling factor is 0.26 .
thus , a much greater portion of the flow is consistent with collections of rigid bodies for the large@xmath1 simulation.,title="fig:",width=338 ] we further explore the validity of this picture by examining the spatial variation of the cubed longitudinal increment , @xmath141 in dns , and @xmath142 in lans@xmath1 for @xmath143 , which in each case is proportional to the energy flux across a fixed scale @xmath103 .
( the presence of the hypothesized `` rigid bodies '' should be evident as significant portions of the flow where there is no energy flux . ) in fig .
[ fig : rigid_rotate ] we show visualizations of these quantities corresponding to @xmath144 ( @xmath145 ) for both the large-@xmath6 lans@xmath1 simulation and a highly turbulent dns ( @xmath51 , @xmath140 ) . the scale ( @xmath145 ) is chosen as it is in the inertial ranges of both flows .
we note that for lans@xmath1 , a significant portion of the flow is not contributing to the flux of energy to smaller scales ( the filling factor for @xmath146 is 0.67 as compared to 0.26 for the navier - stokes case ) .
these regions can be identified as `` polymerized '' or `` rigid bodies '' in @xmath95 and their locations are found to be robust when the @xmath103 used for @xmath142 is varied over a factor of 2 .
moreover , this is highlighted in the probability distribution functions ( _ pdf_s ) , see fig .
[ fig : rigid_pdfs ] , where we see the lans@xmath1 _ pdf _ is more strongly concentrated around zero than the dns .
this is consistent with the idea that the internal _ dof _ of large portions of the flow ( at scales smaller than @xmath6 ) are frozen .
we point out that this comparison is not a les validation , but , rather , a comparison between the dynamics of two different fluids at similar reynolds numbers .
one flow is a well - resolved numerical solution of the navier - stokes equations , and the other is a well - resolved solution of the lans@xmath1 equations with large @xmath6 .
for this reason a reduced resolution ( @xmath147 ) representation for the dns ( for which @xmath148 ) is not depicted in fig . [
fig : rigid_rotate ] .
we have performed such a down - sampling , however , and find the filling factor is reduced even more , to 0.14 , and the tails of the _ pdf _ increase over the full - resolution analysis ( dotted line in fig .
[ fig : rigid_pdfs ] ) .
no inverse helmholtz filtering , @xmath149 is applied to the dns data . note that this would amount to computing @xmath150 in the dns , which has no meaning in the dynamics of the navier - stokes equations ( the energy flux is proportional to @xmath141 ) .
we end this section with further evidence of coexistent energy spectra , @xmath5 and @xmath7 , in separate spatial portions of the flow .
we mask out all portions of the flow that we identify with rigid bodies ( @xmath151 , a 2d slice of which is shown in fig .
[ fig : rigid_rotate ] ) .
the energy spectrum of the remaining portion of the flow is shown in fig .
[ fig : two_spectra ] as a dashed line to be compared with the spectrum of the entire flow shown as a solid line .
the operation of spatially filtering the flow before computing the spectrum serves to `` smear out '' the energy spectrum by convolving it with the spectrum of the filter .
deconvolution in 3d with @xmath147 is intractable and we are , therefore , unable to remove this `` smearing '' of the energy spectrum of the cascading portions of the flow . nonetheless , after conducting what tests we could with the filtering process ( not shown here )
, we conclude that the power law of the energy spectrum of these portions is negative and , thus , distinctly different from that of the rigid bodies .
it is useful to make a distinction between the quality of a subgrid model and effects arising from nonlinear interactions with discretization errors at marginal spatial resolutions ( which are more characteristic of the discretization employed than of the subgrid model ) @xcite . before doing this
, we require an estimate for the total degrees of freedom for the lans@xmath1 attractor which as we show , unlike for the 2d case ( see @xcite ) , for the 3d case is reduced compared to navier - stokes .
the subdominant @xmath4 scaling is associated with the flux of energy to small scales and thus must be used to estimate the degrees of freedom of the lans@xmath1 attractor , @xmath152 . for dissipation the large wavenumbers dominate and , therefore , combining the lans@xmath1 energy balance , eq .
( [ eq : lansa_balance ] ) , with its sub - filter scale energy spectrum , eq .
( [ eq : lansa_spectrumsc ] ) , allows us to implicitly specify its dissipation wavenumber , @xmath153 , by @xmath154 then we have , @xmath155 using that the linear numerical resolution , @xmath156 , must be proportional to the dissipation wavenumber ( @xmath157 ) and that @xmath158 , we arrive at @xmath159 or , equivalently , @xmath160 where @xmath161 is an unknown constant ( for further details see @xcite ) .
we verify this prediction and determine the constant @xmath161 through the use of a database stemming from studies in which both the free parameter , @xmath6 ( or , equivalently , @xmath162 ) and the linear resolution , @xmath156 , for a set of dns flows with @xmath163 , @xmath164 , @xmath165 , and @xmath166 are varied . in so doing , we establish the necessary numerical resolution for convergence to a grid - independent solution . convergence to the grid - independent solution is determined by comparison of the energy spectrum , @xmath60 , between runs with a constant filter and varying resolution . in fig . [
fig : scale_test_256](a ) , we make such a comparison for @xmath167 ( @xmath147 for dns ) and @xmath168 ( @xmath169 , @xmath170 , @xmath171 , @xmath172 , and @xmath173 for lans@xmath1 ) .
we plot energy spectra compensated by @xmath174 so that a k41 @xmath175 spectrum would be flat .
we see , based on comparing the energy spectra at wavenumbers smaller than @xmath162 to the @xmath176 lans@xmath1 spectrum , that simulations at resolutions of @xmath177 and less are not converged while the one at @xmath178 is .
that is , except for the very small scales at the end of the dissipative range , there is very little difference between the spectra at @xmath178 and at @xmath176 ( i.e. , the solution is `` grid - independent '' ) .
meanwhile , for resolutions of @xmath177 and less the spectra vary greatly with resolution ( i.e. , they are `` unresolved '' ) . in fig .
[ fig : scale_test_256](b ) , we collect all the results of similar studies ( @xmath167 ) in a plot of resolution , @xmath156 , versus inverse filter width , @xmath162 .
( we change @xmath156 for a given @xmath6 , then change @xmath6 and iterate . )
pluses correspond to grid - independent solutions , x s to under - resolved solutions , and squares to `` undecided '' runs ( i.e. , that are neither clearly resolved nor clearly under - resolved ) .
the dashed lines represent eq .
( [ eq : theory ] ) with the minimal and maximal choice of @xmath179 ( where @xmath180 ) , that agrees with our results ( i.e. , @xmath181 ) . in fig .
[ fig : scale_test_512 ] we conduct similar studies for @xmath182 .
we find @xmath183 and again validate the predictive power of eq .
( [ eq : theory ] ) for the necessary numerical resolution for grid - independent solutions .
the greatest utility of the prediction , however , is due to the single constant @xmath161 which is independent of reynolds number .
a determination of this constant can cheaply be achieved repeating this process for several runs for low and moderate @xmath9 , and determines the resolution requirement for the highest @xmath9 attainable .
the ranges of acceptable constants , @xmath184 , for the four reynolds number flows studied are plotted versus @xmath9 in fig .
[ fig : scale_confirm ] .
a power law @xmath185 fits our data with @xmath186 demonstrating the final validation of the prediction , @xmath187 , eq .
( [ eq : theory ] ) .
the value of the constant is found to be @xmath188 .
we made one study for the maximally - helical abc forcing at @xmath189 and @xmath190 .
it is consistent with a value of @xmath191 .
we therefore conclude that the constant @xmath161 is not a strong function of the forcing employed or of the scale at which the system is forced . as a result , and unlike in 2d lans@xmath1 @xcite , we verify that the size of the attractor in 3d lans@xmath1 is smaller than that in navier - stokes , which is a promising result if the lans@xmath1 equation is going to be used as an les . however , before doing this , an assessment of the truncation errors introduced in discretized systems ( as used to solve the equations numerically ) and a study of the optimal choice for @xmath6 to capture the properties of a dns is needed .
we consider these problems in the following section .
in this section , we consider the lans@xmath1 equations as a means to an end , and consider the solutions to their discretized equations as approximations to the navier - stokes solutions .
we seek numerical approximations of lans@xmath1 that minimize the difference to a fully resolved or direct numerical solution ( dns ) of navier - stokes ( i.e. , we analyze the behavior of lans@xmath1 solutions in the les framework , and call here the model a `` lans@xmath1 les '' , or in short `` @xmath6-les '' ) . in the les framework , lans@xmath1 s turbulent stress tensor , @xmath192 , is given by ( see , e.g. , @xcite ) @xmath193 previous studies have not made the distinction between grid - independent lans@xmath1 and lans@xmath1 les , though one did study convergence to grid - independent solutions at moderate @xmath9 @xcite .
we find , however , a definite difference between the two approaches .
we show in this section that , in fact , lans@xmath1 combined with truncation error yields a better fit to dns than grid - independent lans@xmath1 .
the resolution that yields an optimal @xmath6-les ( a terminology to be defined below ) is also found to follow eq .
( [ eq : theory ] ) . in the section [ les_quality ]
, we then address the quality and usability of the predictions of the lans@xmath1 model viewed as an les
. a remark about nomenclature may be in order at this point . traditionally , and for good reasons ,
les attempt at capturing the large - scale properties of a flow with a huge reynolds number , as found , e.g. , in the atmosphere . in that case , the wavenumber at which the dns is truncated is , at best , in the inertial range and it might even be in the energy - containing range , as for the atmospheric boundary layer with a taylor reynolds number @xmath194 . of a different nature
are the modeling methods sometimes called quasi - dns . here , the idea is to model a flow at a given , moderate reynolds number but with an expense in computing resources lesser than if performing a dns .
under - resolved dns fall in that category ; in that case , the large - scales are presumably well reproduced but the small scales are noisy .
it is in that spirit that we now examine the properties of the lans@xmath1 model .
we thus qualify a model as optimal in the sense of being optimal for the class of lans@xmath1 models examined herein ; in order to avoid repetition , we also use the terminology of alpha - optimal . in fig .
[ fig:512_les ] with @xmath195 , we plot the @xmath182 dns spectrum ( solid black line ) and lans@xmath1 spectra at three different resolutions .
we observe that , while the @xmath196 solution is not converged , it is a better approximation to the dns than the grid - independent lans@xmath1 solution . for all simulations we studied ,
the grid - independent lans@xmath1 solution is not the best approximation to the dns .
another example is given in fig .
[ fig : les_err ] where we plot the mean square spectral error normalized to make fair comparisons between large and small @xmath162 results , @xmath197 where @xmath198 is the wavenumber for the forcing scale , @xmath199 is the dns spectrum ( in the @xmath200 norm ) , @xmath60 is the lans@xmath1 spectrum ( in the @xmath56 norm ) , and @xmath201 is the number of terms in the sum .
these errors are calculated for spectra averaged over turbulent steady - state solutions : @xmath202 $ ] for @xmath22 .
we see that for a given filter or a given simulation resolution , there is a local minimum in the error .
this minimum is a balance between truncation errors and the approximation error due to using lans@xmath1 instead of the full navier - stokes equations . due to these errors being , in some sense
, in opposition , the optimal @xmath6-les solution is found at a lower resolution than the grid - independent solution .
indeed , we see by examining fig .
[ fig : les_err ] ( a ) that for a given filter the combination of truncation error and the lans@xmath1 solution is a better approximation to the dns . for fixed resolution , fig .
[ fig : les_err ] ( b ) , the optimal value for @xmath6 is not zero but has some finite value .
this local minimum error shown in the figure keeps @xmath6 from going to zero ( @xmath203 ) in dynamical models @xcite .
we note , also , that the error is low for a finite range of @xmath156 and @xmath162 near the minimum , indicating that an @xmath6-les solution may perform well for a range of parameters near the optimal ones .
we find the resolution for an optimal @xmath6-les is also predicted by eq .
( [ eq : theory ] ) ( with @xmath204 for @xmath182 , or @xmath205 ) .
that is , optimal @xmath6-les resolution is just below that for grid - independent lans@xmath1 solutions .
having demonstrated the predictability of the resolution for grid - independent lans@xmath1 and of lans@xmath1 les given a reynolds number and a filter , in the following section we seek to determine sufficient conditions on the free parameter @xmath6 for lans@xmath1 to be a successful les . in this section
, we make an analysis of the les potential of lans@xmath1 by considering only the grid - independent lans@xmath1 solutions identified using eq .
( [ eq : theory ] ) .
note that from the results discussed in the previous section , we expect lans@xmath1 optimal grid - dependent @xmath0les approximations to have better performance . in the limit of @xmath6 going to zero , lans@xmath1 eq .
( [ eq : lanssc ] ) recovers the navier - stokes equations , eqs .
( [ eq : navier - stokes ] ) , but the question we address now is how small must @xmath6 be for lans@xmath1 solutions to be good approximations to navier - stokes solutions .
there are several length scales that @xmath6 could be related to : the forcing scale @xmath206 , the integral scale @xmath207 , the taylor microscale @xmath208 , or the kolmogorov dissipation scale @xmath209 .
plots of the mean square spectral errors to dns ( see eq .
( [ eq : err_2sc ] ) ) versus these scales are shown in fig .
[ fig : error1 ] . while the general trend of errors decreasing with @xmath6 is apparent in all cases , in fig .
[ fig : error1](a ) we see a large difference between errors at varying reynolds numbers and similar ratios of @xmath6 to the forcing scale , @xmath206 . for a linear least - squares fit , the goodness - of - fit , @xmath210 , was found to be @xmath211 .
the errors for @xmath31 are much larger than for the same ratio @xmath212 as results at both @xmath167 and @xmath182 .
this is also the case for the integral scale .
however , the quality of the @xmath6-les appears to be closely tied to the ratio of @xmath6 to the kolmogorov dissipation scale . in fig .
[ fig : error1](b ) the errors are plotted versus the ratio of the dissipation scale , @xmath209 , to @xmath6 .
we see a very strong dependence ( @xmath213 ) between errors for several runs with four different reynolds numbers indicating that the quality of the lans@xmath1 les approximation is a function of the ratio of @xmath6 to the dissipative scale . finally , in fig .
[ fig : error1](c ) the errors are plotted versus the ratio of the taylor scale , @xmath208 , to @xmath6 .
we find @xmath214 for a linear least - squares fit .
we note that a single experiment conducted at @xmath215 ( the asterisks ) confirms that the maximal value of @xmath6 is tied to the dissipation scale and not the taylor scale .
this is more clearly demonstrated in fig .
[ fig : error2 ] where we plot compensated energy spectra for a nearly constant ratio @xmath216 at three reynolds numbers .
we see that the maximum deviation from the dns spectrum increases with @xmath9 .
as @xmath216 is the same in all cases , the optimal @xmath6 is not dependent on the taylor scale .
these findings were not accessible at lower reynolds numbers due to inadequate separation of scales .
for example , we give in fig . [
fig : flux](a ) spectral flux for dns at @xmath19 , @xmath164 , and @xmath166 respectively .
we define the kinetic energy transfer function , @xmath217 , in fourier space as @xmath218 , where @xmath219 represents the fourier transform .
for lans@xmath1 we have @xmath220 where @xmath221 .
the flux is defined as usual from the transfer function as @xmath222 only @xmath31 ( and @xmath215 not pictured here ) demonstrates a range of nearly constant flux ( a well - defined inertial range ) before the dissipation scales . following the scaling arguments in ref .
@xcite , one effect of the @xmath0model is to increase the time scale for the cascade of energy to small scales .
this reduces the flux as @xmath6 increases ( @xmath162 decreases ) as do the hypothesized `` rigid bodies ; '' this can be seen in fig .
[ fig : flux](b ) .
( note that in dns at high resolution , 80% of the flux is from local interactions which is strongly suppressed at scales smaller than @xmath6 @xcite . ) as dissipation dominates the flux for low and moderate reynolds number , the reduced flux of the @xmath0model has little consequence for these simulations . with a substantial inertial range , however , this reduced flux results in a pile - up of energy for scales larger than the dissipative scale and the spectrum approaches the @xmath130 spectrum discussed in section [ scaling ] . as a consequence of the integral conservation of energy ( @xmath223 )
there is a corresponding decrease of energy at large scales .
consequently , as the inertial range increases , @xmath6 must be moved to smaller and smaller scales in order for lans@xmath1 not to alter scales larger than @xmath6 . in summary ,
the @xmath0model s reduced flux of energy to small scales is more crucial when the dissipation scale is farther away from @xmath6 .
if @xmath6 must be directly proportional to the kolmogorov dissipation scale , we can estimate the les computational savings of the lans@xmath1 model .
for the navier - stokes equations we have @xmath224 and , as we verified in section [ gridindep ] , for lans@xmath1 we have @xmath225 .
if @xmath162 is directly proportional to the navier - stokes dissipation wavenumber , @xmath226 , we arrive at @xmath227 and , consequently , @xmath228 note that for free @xmath6 , @xmath152 ( _ dof _ of lans@xmath1 ) is much smaller than @xmath229 .
but , to obtain an optimal les , @xmath6 is tied to @xmath226 ; then the resolution requirements ( @xmath230 ) are different and the decrease in necessary computational resolution from employing lans@xmath1 is fixed .
in fact , for the forcing and boundary conditions employed , we find @xmath231 we note that eq .
( [ eq : nogain ] ) is consistent with theoretical predictions given in ref .
other les such as the similarity model @xcite and the nonlinear ( or gradient ) model @xcite have also exhibited the characteristic that they achieve only moderate reductions in resolution and are , therefore , frequently used in mixed models with a smagorinsky term ( see , e.g. , @xcite ) .
that such additional terms will be required for lans@xmath1 to reproduce the energy spectrum of high @xmath9 flows , may not be a significant factor in its usability .
note that the usual addition of extra dissipative subgrid - stress terms ( as in the smagorinsky model ) also introduces a stronger dependence of the system of equations with the spatial resolution , since the filter width in such models is often associated to the maximum wavenumber in the box , @xmath232 . in that case
, it can make more sense to use grid - dependent solutions of lans@xmath1 ( discussed at the beginning of section [ les ] ) which give an optimal lans@xmath1 les , and can as a result give an extra gain in the computational costs .
we also conclude that , with the scale @xmath6 being tied to the dissipation scale @xmath209 , the model lans@xmath1 behaves more like a quasi - dns by opposition to a traditional les .
note however that a factor of @xmath233 in resolution gain translates into a factor 27 in cpu and a factor 12 in memory savings , still a substantial gain .
in this section , we compare and contrast lans@xmath1 and navier - stokes solutions at high reynolds number . using results of previous sections for optimal resolution and the necessary value of @xmath6 to approximate dns
, we now evaluate both grid - independent lans@xmath1 solutions and a single lans@xmath1 les for a highly turbulent flow ( @xmath31 , @xmath28 ) .
we calculate grid - independent solutions for @xmath234 ( @xmath235 ) , for @xmath236 ( @xmath235 ) , and for @xmath237 ( @xmath238 ) .
a lans@xmath1 les solution is computed for @xmath236 ( @xmath238 ) .
averaged compensated energy spectra are shown in fig .
[ fig : bad_alpha_70 ] .
we see that the optimal lans@xmath1 les is a better approximation of the dns spectra than the grid - independent lans@xmath1 for the same value of @xmath6 ( @xmath239 ) .
we also see that if @xmath6 is increased further , the energy spectrum approaches the @xmath7 spectrum discussed in section [ subdominant ] .
fig [ fig : vapor_tube ] is a perspective volume rendering of the enstrophy density @xmath240 ( @xmath241 for lans@xmath1 ) for the dns , @xmath242 lans@xmath1 , and @xmath243 lans@xmath1 . due to the late time depicted here ( @xmath244 , longer than a lyapunov time )
there can be no point - by - point comparison between the simulations .
however , we note that the helical structure of the vortex tubes is preserved by the @xmath0model but that the tubes themselves are shorter and somewhat thicker for large values of @xmath6 .
as was noted for moderate reynolds numbers , this is due to lans@xmath1 suppressing vortex stretching dynamics without changing its qualitative features @xcite .
this is in contrast to 2d lans@xmath1 where the vorticity structures are seen to get thinner as @xmath6 increases @xcite .
this could also be related to the scaling differences between 2d and 3d lans@xmath1 .
it has been claimed that the development of helical structures in turbulent flows can lead to the depletion of nonlinearity and the quenching of local interactions @xcite .
the depletion of energy transfer due to local interactions at some cutoff in wavenumber is also believed to bring about the bottleneck effect @xcite .
consistent with these results , in 2d lans@xmath1 ( where the vorticity structures are more fine than navier - stokes ) the spectrum is steeper and in 3d lans@xmath1 ( where the vorticity structures are shorter but fatter than navier - stokes ) the spectrum is shallower .
figure [ fig : struct3 ] shows the third - order ( mixed ) structure functions corresponding to the krmn - howarth theorems versus length @xmath103 . for the dns
, we show @xmath245 and @xmath246 for lans@xmath1 .
the dotted vertical lines indicate the various @xmath6 s .
a small inertial range for the dns near @xmath247 is reproduced by all lans@xmath1 results .
the largest @xmath6 ( @xmath248 ) exhibits a second inertial range at scales just smaller than @xmath6 ( @xmath249 is consistent with eq .
( [ eq : lcubelans ] ) ) .
we note this is the first demonstration of third - order structure functions in lans@xmath1 consistent with a k41 inertial range followed by an @xmath6 inertial range and finally a dissipative range .
next , we observe the scaling of the longitudinal structure functions , @xmath250 where we again replace the @xmath251 norm for the @xmath252 norm in the case of lans@xmath1 , @xmath253 we utilize the extended self - similarity ( ess ) hypothesis @xcite which proposes the scaling @xmath254 or , for lans@xmath1 , @xmath255 we display our results in fig . [
fig : scaling ] .
we note that for lans@xmath1 , the third - order exponent is not equal to unity , contrary to the navier - stokes case .
the krmn - howarth theorem implies @xmath256 , not @xmath257 .
we measured the deviation from linearity for each experiment and found that lans@xmath1 becomes more intermittent as @xmath6 increases ( @xmath243 is slightly more intermittent than the dns ) .
as artificially dropping local small - scale interactions gives enhanced intermittency @xcite , this increased intermittency is the expected result of lans@xmath1 reducing interactions at scales smaller than @xmath6 .
we note , however , that even with such a large filter , lans@xmath1 is a good approximation to the intermittency properties of the dns .
this is surprising given its energy spectrum and reduced flux in the inertial range .
it is probably linked to the fact that lans@xmath1 preserves global properties ( in an @xmath258 sense ) of the navier - stokes equations and that these properties are important to the dynamics of small scales as measured by high - order structure functions .
we computed solutions of the lagrangian - averaged navier - stokes @xmath0model ( lans@xmath1 ) in three dimensions for significantly higher reynolds numbers ( up to @xmath2 ) than have previously been accomplished and performed numerous forced turbulence simulations of lans@xmath1 to study their equilibrium states .
the results were compared to dns for @xmath259 @xmath166 , and @xmath260 , the last performed on a grid of @xmath29 points .
we note that there are two ways to view the lans@xmath1 simulations : as converged or `` grid - independent '' solutions of the lans@xmath1 equations or as large - eddy simulations ( @xmath0les ) which include grid effects .
we found a definite difference between the two approaches in that the fully - converged grid - independent lans@xmath1 is not the best approximation to a dns of navier - stokes .
instead , the minimum error is a balance between truncation errors and the approximation error due to using lans@xmath1 instead of the full navier - stokes equations . due to these errors being , in some sense
, in opposition , the optimal @xmath6-les solution was found at a lower resolution than the grid - independent solution ( the error was low for a finite range of @xmath156 and @xmath6 near the minimum , indicating that a lans@xmath1 viewed as an les solution may perform well for a range of parameters ) . unlike the 2d case @xcite
, 3d lans@xmath1 has been shown to be a subgrid model ( i.e. , it reduces the resolution requirements of a given computation ) .
this difference between 2d and 3d lans@xmath1 indicates that other @xmath0models ( as the lamhd@xmath1 eqs .
@xcite or the bv@xmath1 eqs .
@xcite ) may behave differently and studies of these systems at high resolution may be required .
we confirm the presence of the theoretically predicted @xmath4 scaling of the third - order structure function ( corresponding to a @xmath5 scaling of the energy spectrum ) @xcite through its bound on the number of degrees of freedom for lans@xmath1 @xcite , in the structure functions of the smoothed velocity in simulations with large @xmath6 , and in the spectrum of specific spatial portions of the flow . in so doing
, we have validated the predictive power of the bound @xmath261 , for the numerical resolution for grid - independent lans@xmath1 solutions and for optimal lans@xmath1 les ( with a separate constant of proportionality ) .
the great utility of the prediction is that the single constant can cheaply be determined at low and moderate reynolds number and predicts the resolution requirement for the highest reynolds numbers attainable .
we further found no great change in this single constant when employing the non - helical taylor - green or the maximally - helical abc forcings .
however , the small scale ( @xmath128 ) lans@xmath1 spectrum was observed to be @xmath136 .
we attribute this to the frozen - in - turbulence closure employed in deriving the @xmath0model . for scales smaller than @xmath6
, portions of the smoothed flow @xmath95 are locked into `` rigid bodies . '' by `` rigid bodies , '' we mean the internal degrees of freedom are frozen and these portions give no contribution to the energy cascade .
this is consistent both with the observed @xmath136 spectrum and with field increments @xmath117 being observed to be approximately zero over a large portion ( compared to navier - stokes ) of the flow .
the turbulent energy cascade occurs in the space between these `` rigid '' portions .
while the @xmath5 portions are subdominant to the @xmath136 portions in the energy spectrum , they prevail in the cascade and hence both the structure functions and the degrees of freedom of the lans@xmath1 attractor .
we find that both of these scalings ( @xmath136 and @xmath5 ) contribute to a reduction of flux at constant energy ( i.e. , the dissipation is reduced as has previously been observed in 2d calculations @xcite ) .
this leads to a shallower ( or even growing ) energy spectrum as @xmath6 increases .
thus , for lans@xmath1 viewed as an les to reproduce the navier - stokes energy spectrum it is necessary that @xmath6 be not much larger than the dissipation scale ( @xmath262 independent of reynolds number ) ; in that sense , it can be considered as a quasi - dns as opposed to a traditional les , substantially larger reynolds numbers being modeled in the latter case , leading to substantially larger gain in resolution . as a consequence ,
the computational savings of lans@xmath1 is fixed and not a function of reynolds number .
( however , and unlike the 2d case , the 3d @xmath0model does give a computational saving when used as a les . )
this result _ was not accessible at lower reynolds numbers due to inadequate separation of scales_. however , in one previous study for decaying turbulence with energy initially mostly at low wavenumbers ( @xmath263 ) , it was evident that as time evolved and energy moved to smaller scales , the resolution requirements of lans@xmath1 increased @xcite .
other les such as the similarity model @xcite and the nonlinear ( or gradient ) model @xcite have also exhibited the characteristic that resolution may be decreased only modestly and are , therefore , frequently used in mixed models with a smagorinsky term ( see e.g. , @xcite ) .
that such additional terms will be required for lans@xmath1 to reproduce the energy spectrum of high @xmath9 flows , may not be a significant factor in its usability .
we compared and contrasted lans@xmath1 to a dns at @xmath31 considering both structures and high - order statistics such as the longitudinal structure functions which are related with intermittency . with an appropriate choice of @xmath6
we were able to observe a navier - stokes inertial range followed by lans@xmath1 inertial range at scales smaller than @xmath6 .
for this second inertial range we again observed a @xmath136 energy spectrum .
as @xmath6 increased , we noted a change in the aspect ratio of vortex tubes ( they became shorter and fatter ) . this can be related to quenching of local small - scale interactions at scales smaller than @xmath6 and , thus , to the shallower spectrum for 3d lans@xmath1 @xcite .
therefore , in 2d lans@xmath1 ( where the vorticity structures are more fine than navier - stokes ) the spectrum is steeper @xcite and in 3d lans@xmath1 ( where the vorticity structures are shorter but fatter than navier - stokes ) the spectrum is shallower . finally , an examination of the longitudinal structure functions indicate that intermittency is increased as the parameter @xmath6 is increased consistent with the suppression of local small - scale interactions at scales smaller than @xmath6 @xcite .
the elimination of the faster and faster interactions among smaller and smaller scales through the modified nonlinearity in lans@xmath1 ( together with the discrepancy between its solutions and navier - stokes solutions ) highlights the importance of these interactions down to scales only slightly larger than the dissipative scale .
that is , by removing these interactions anywhere in the inertial range ( e.g. , @xmath264 ) , the resulting energy spectrum was found to differ from the dns at scales larger than @xmath6 .
the intermittency properties of the dns , however , were well reproduced even with large filters . noting this , if lans@xmath1 s @xmath7 energy spectrum is not important for a given application , much greater reductions in resolution can be achieved .
future work should address whether this may be remedied in a lans@xmath1 les by the inclusion of another ( dissipative ) model for these interactions , or ( in the case of magneto - hydrodynamics @xcite whether this problem is less significant because of the presence of greater spectral nonlocality @xcite .
the effect of lans@xmath1 on the detailed scale - by - scale energy transfer should also be investigated as our results indicate that a model for local small - scale interactions would improve the @xmath0model .
another direction of future research is to explore other reduced lans@xmath1 models , clark@xmath1 and leray-@xmath6 , which break the frozen - in - turbulence closure and , also , the conservation of circulation . finally , note that because of its greater mathematical tractability , lans@xmath1 possibly allows for a better understanding of multi - scale interactions in turbulent flows thus modeled ; therefore , detailed studies such as the one presented here may , _ in fine _ , allow for a better understanding of turbulence itself .
computer time was provided by ncar and by the national science foundation terascale computing system at the pittsburgh supercomputing center .
the nsf grant no .
cmg-0327888 at ncar supported this work in part and is gratefully acknowledged .
three - dimensional visualizations of the flows were done using vapor .
the authors would like to express their gratitude for valuable discussions with bob kerr . | we compute solutions of the lagrangian - averaged navier - stokes @xmath0model ( lans@xmath1 ) for significantly higher reynolds numbers ( up to @xmath2 ) than have previously been accomplished .
this allows sufficient separation of scales to observe a navier - stokes inertial range followed by a second inertial range specific to lans@xmath1 .
both fully helical and non - helical flows are examined , up to reynolds numbers of @xmath3 .
the analysis of the third - order structure function scaling supports the predicted @xmath4 scaling ; it corresponds to a @xmath5 scaling of the energy spectrum for scales smaller than @xmath6 .
the energy spectrum itself shows a different scaling which goes as @xmath7 .
this latter spectrum is consistent with the absence of stretching in the sub - filter scales due to the taylor frozen - in hypothesis employed as a closure in the derivation of lans@xmath1 .
these two scalings are conjectured to coexist in different spatial portions of the flow .
the @xmath4 ( @xmath8 ) scaling is subdominant to @xmath7 in the energy spectrum , but the @xmath4 scaling is responsible for the direct energy cascade , as no cascade can result from motions with no internal degrees of freedom . we demonstrate verification of the prediction for the size of the lans@xmath1 attractor resulting from this scaling . from this
, we give a methodology either for arriving at grid - independent solutions for lans@xmath1 , or for obtaining a formulation of a large eddy simulation ( les ) optimal in the context of the alpha models .
the fully - converged grid - independent lans@xmath1 may not the best approximation to a direct numerical simulation of the navier - stokes equations since the minimum error is a balance between truncation errors and the approximation error due to using lans@xmath1 instead of the primitive equations .
furthermore , the small - scale behavior of lans@xmath1 contributes to a reduction of flux at constant energy , leading to a shallower energy spectrum for large @xmath6 .
these small - scale features , however , do not preclude lans@xmath1 to reproduce correctly the intermittency properties of the high reynolds number flow . |
few - electron gate - defined quantum dots@xcite are exploited for single spin manipulation that allows for realization of single - qubit quantum gates.@xcite while the desired spin rotation involves a single spin as a carrier of quantum information , multielectron systems provide a feasible environment for readout of the controlled spin .
strong spin - orbit ( so ) coupling that is present in insb and inas nanowires @xcite allows for electrical spin rotations@xcite that are performed by electric dipole spin resonance ( edsr)@xcite which excludes the need for introduction of local magnetic field gradients@xcite or usage of hyperfine interaction.@xcite the readout of the spin is realized via spin to charge conversion that relies on the pauli spin blockade.@xcite the single electron current @xmath1 [ the numbers in the brackets correspond to the number of electrons in a particular dot ] is blocked at the transition from @xmath2 triplet to @xmath3 singlet .
rotation of one of the spins of electrons constituting the @xmath2 triplet unblocks the current which serves as a proof for the coherent spin control . on the other hand
strong so coupling leads to unavoidable spin relaxation which results in a spontaneous lifting of the pauli blockade when one of the @xmath2 triplets is close in energy to the @xmath3 singlet.@xcite edsr lifting of the current blockade is observed already for two electrons bound in the double dot , which indeed is the case for many of the experiments.@xcite however some of the experimentally studied devices consist an even number @xmath4 of electrons greater than two.@xcite in this case the system is biased such the pauli blockade is between @xmath5 and @xmath6 states .
it is assumed that such configuration is equivalent to the two - electron system.@xcite this approximation resembles the well established concept in chemistry , that the valence electrons are responsible for creation of bonds and the rest in the deep levels can be treated as the frozen core.@xcite this assumption seems questionable for quantum dots in which the single - electron shells are separated by much smaller energies than for the coulomb potential , nevertheless this problem has not been discussed by a theoretical study .
the present work addresses this issue .
we find that for a system with @xmath7 all but two electrons form closed singlet shells .
this is in accordance with predictions of hubbard model,@xcite that appear as a consequence of mattis - lieb theorem,@xcite and which states that the lowest energy states posses the lowest spin ( s=0 ) . as a consequence , in general ,
the low - energy spectra of multielectron double quantum dots in the @xmath5 configuration resemble the spectra of quantum dots in @xmath2 configuration and the states have similar total spins .
we find that the @xmath5 spectra can be well recreated by a configuration interaction calculation in which one excludes the single - electron orbitals that form the singlet shells .
this is analogous to the frozen core approximation regardless of the fact that there are no orbital shells in quantum - dots .
the main finding of the work is that though a general resemblance of @xmath7 and @xmath8 spectra is found , the occupation of excited single - electron orbitals in the @xmath7 case leads to lifting of the degeneracy of spin - zero states .
this in turn is translated to different effective @xmath0-factors in the dots .
such differences have been observed in recent edsr experiments on nanowire quantum dots@xcite and have been related to the differences in the confinement as predicted by the study on self - organized quantum dots.@xcite here we strictly connect the effective @xmath0-factors with the number of electrons in the system and the length of the dots .
we find that unequal effective @xmath0-factors for @xmath8 appear only for an asymmetric system but for @xmath7 they are observed already for the dots of the same lengths .
in the present work we follow the common approach@xcite that treats the nanowire quantum dots as quasi - one dimensional .
the n - electron system is described by the hamiltonian , @xmath9 .
\label{hne}\ ] ] the form of the coulomb interaction term results from the assumption that the electrons are localized in the ground state of the lateral quantization along the nanowire cross section with the wavefunction of a gaussian shape with @xmath10 .
the integration of three - dimensional coulomb interaction term @xmath11 leads@xcite to the operator including @xmath12 which is the exponentially scaled complementary error function.@xcite to obtain n - electron spin - orbitals we diagonalize the hamiltonian ( [ hne ] ) in a basis of slater determinants consisting of single - electron spin - orbitals .
@xmath13 where @xmath14 corresponds to the orbital and spin coordinates , @xmath15 is the antisymmetrization operator and @xmath16 is obtained by the diagonalization .
we use @xmath17 single - electron spin - orbitals which provides accuracy better than @xmath18ev .
the single - electron orbitals @xmath19 are described by the hamiltonian , @xmath20 where @xmath21 corresponds to rashba so coupling@xcite resulting from @xmath22 hamiltonian averaged in the @xmath23-direction .
@xmath24 describes the potential profile of the double dot , @xmath25 where @xmath26 and @xmath27 determine the length of each dot , @xmath28 is the interdot barrier width , @xmath29 is the barrier height and @xmath30 is the bias potential applied to the bottom of the left dot .
we assume a @xmath31 nm thin and @xmath32 mev high interdot barrier .
the computational box ends at the edges of the defined potential and the magnetic field is applied along the nanowire axis .
the single - electron eigenstates are obtained by exact diagonalization of hamiltonian ( [ 1eham ] ) on a mesh of 201 points with @xmath33 nm .
we adopt parameters corresponding to insb nanowires , i.e. @xmath34 , @xmath35 , @xmath36 and @xmath37 mevnm which corresponds to spin - orbit length @xmath38 nm comparable to the value measured experimentally in ref . .
we take @xmath39 nm .
let us first consider a symmetric system of two quantum dots of lengths @xmath40 nm . we set the bottom of the left dot to @xmath41 mev .
the bias results in the energy level configuration such that @xmath3 singlet@xcite is the ground state and the lowest - energy excited states are @xmath2 states with different spin polarizations .
this configuration is necessary for observation of spin pauli blockade .
the inset to fig .
[ spectr2e](a ) shows the lowest part of the energy spectrum .
the ground state singlet of ( 2,0 ) occupation has mean value of @xmath42 operator 0.12 @xmath43 .
figure [ spectr2e](a ) presents energy levels of ( 1,1 ) states .
the two zeeman split energy levels correspond to a spin - positive triplet @xmath44 ( @xmath45 ) with spins oriented approximately along the magnetic field and to a spin - negative triplet @xmath46 ( @xmath47 ) with spins oriented against the magnetic field .
the horizontal curve corresponds to a degenerate energy level of a singlet ( s ) and a triplet ( @xmath48 ) states with zero spin projection along the direction of the magnetic field .
the degeneracy results from the negligible overlap between the adjacent electrons and hence nearly zero exchange interaction .
the mean values of @xmath42 operator for these states are : 1.04 , 1.01 @xmath49 $ ] . in edsr experiments
the spin rotations are performed from one of the non - zero spin triplets : @xmath44 or @xmath46.@xcite when a resonance to a state with zero spin component along the magnetic field occurs the blockade is lifted .
the experimentally measured resonances exhibit a linear dependence of the driving frequency on the magnetic field , equal to ( considering @xmath44 as the initial state ) @xmath50/\hbar$ ] .
the corresponding energy @xmath51 is plotted in fig .
[ 2ediffen ] with the red - dashed curve .
let us now consider the case in which the dots are of unequal length
@xmath52 nm and @xmath53 nm but we keep the @xmath0-factor constant along the structure .
energy levels of states in which electrons occupy adjacent dots are presented in fig .
[ spectr2e](b ) . to keep the energy separation between ( 2,0 )
singlet and four ( 1,1 ) states 1.5 mev at @xmath54 as in the case of symmetric system we set @xmath55 mev in the left dot .
the striking difference between the spectra of fig .
[ spectr2e](b ) as compared to the symmetric case of fig .
[ spectr2e](a ) is that for the non - zero magnetic field the degeneracy of horizontal energy levels is lifted .
the spin densities of the states that correspond to these energy levels calculated as @xmath56 are depicted in the insets to fig .
[ spectr2e](b ) .
we observe that in the state with lower energy the spin in the left dot is oriented against the magnetic field , while in the right dot it is oriented along the magnetic field .
further on we will address to this state as to @xmath57 .
the state @xmath58 with an increasing energy in @xmath59 has an opposite spin configuration .
the @xmath57 , @xmath58 states have zero total spins along the @xmath60-direction therefore the edsr transitions to this states lift the pauli blockade and such transition are visible as a resonance lines in edsr spectra .
we calculate corresponding energy differences @xmath61 $ ] , @xmath62 $ ] and plot them in fig .
[ 2ediffen ] with black solid lines .
we note that this arrangement of resonance lines is present in every edsr map registered experimentally [ see refs . ] and attributed to different @xmath0-factors in the dots . here
it is obtained for a constant @xmath0 along the structure .
the slopes of the curves in fig .
[ 2ediffen ] are connected to _ effective _ @xmath0-factor . for symmetric system
we calculate @xmath63 equal to @xmath64 for @xmath65 mt . in the case of asymmetric dots the effective @xmath0-factors are @xmath66 , @xmath67 . to explain the impact of the dots width on the energy spectra and effective @xmath0-factors let us inspect the single electron spin - orbitals that constitute the two - electron orbitals . figures [ spectr1e ] ( a , e ) show the charge densities of the single - electron states .
the densities correspond to the ground states of orbital quantization of each dot .
the black curves correspond to @xmath68 and @xmath69 states [ the main letter denotes the orbital excitation , @xmath70 denotes the dot in which the electron is localized and the arrows correspond to the average spin polarization direction ] .
the red - dashed curve shows the charge densities of higher energy states @xmath71 and @xmath72 in which the electron occupies the right dot .
we extract the squared absolute values of coefficients
@xmath73 for each of the slater determinant that is used in the configuration interaction approach . for the symmetric case we get 0.806 for the determinant consisting of @xmath74 single - electron orbitals and 0.194 for consisting @xmath75 orbitals for one of the states from the degenerate pair of spin zero two - electron states ( the coefficients for the second state are reversed ) . for the asymmetric case we get 0.992 for @xmath75 for the @xmath57 state and 0.996 for @xmath74 for the @xmath58 state .
the lack of the admixture of other slater determinants is due to small size of the dots which results in a considerable kinetic energy separation of the single - electron orbitals .
the energy spectra displayed in fig .
[ spectr1e](b , f ) show the zeeman splittings of the single - electron energy levels . if we overlay the energy levels of the states in which the electron is localized in the left and right dot ( solid black and red - dashed curves in fig .
[ spectr1e](b , f ) respectively ) we observe that they are exactly the same [ fig .
[ spectr1e](c ) ] when the dots are of identical length but they differ when the dots are of unequal length [ fig .
[ spectr1e](g ) ] .
now we sum the single - electron energies accordingly to the way the states enter the configuration - interaction calculation , i.e. the @xmath57 state corresponds to the occupation of @xmath69 , @xmath71 single - electron orbitals with energies @xmath76 and @xmath77 .
the @xmath58 corresponds to the occupation of @xmath68 , @xmath72 single - electron orbitals with energies @xmath78 and @xmath79 .
the obtained sums are plotted in fig .
[ spectr1e](d ) for the symmetric and in fig .
[ spectr1e](h ) for the asymmetric case .
we find that the degeneracy is lifted due to different zeeman splittings of single - electron energy levels of electrons confined in dots of different length . in a single one - dimensional quantum dot so interaction impacts the zeeman splittings according to @xmath80 where@xcite @xmath81 due to a high interdot barrier we can effectively treat the considered system as two separate dots . for a quantum dot in a form of an infinite quantum
well the term @xmath82 that controls the strength of the zeeman splitting is @xmath83 for @xmath84 in a form of @xmath85-like orbital .
@xmath82 changes from 1 for narrow quantum dots to 0 in the limit of infinite dot length .
accordingly , the zeeman splittings are the strongest ( as strong as in the absence of so coupling ) for a narrow quantum dot and become weaker if the length of the dot is increased .
the @xmath86-factors calculated from @xmath61 $ ] and @xmath62 $ ] depend on the zeeman splitting of the single electron energy levels as follows [ taking @xmath87 as the orbital excitation energy of @xmath57,@xmath58 and @xmath44 states ] : @xmath88 and @xmath89 .
therefore @xmath90 and @xmath91 .
we plot the ratio @xmath92 in fig .
[ g1g2](a ) . for @xmath52 and @xmath53
we obtain @xmath93 which matches well the value obtained in the exact calculation of fig . [
2ediffen ] , @xmath94 it should be noted here that the effect of the so interaction on the strength of the zeeman splittings is influenced also by the orientation of the magnetic field.@xcite for the magnetic field vector forming a @xmath95 angle with the nanowire axis the splitting becomes @xmath96 , i.e. the @xmath86 values obtained for the magnetic field oriented perpendicular to the nanowire axis approach the bulk @xmath0-factor value .
figure [ spectr4e](a ) with black solid curves presents energy levels of four - electron symmetric system with @xmath97 mev .
the levels correspond to the states with ( 3,1 ) occupation .
the plot omits the ground state with ( 4,0 ) occupation that is @xmath98 mev lower in energy with respect to presented energy levels for @xmath54 .
the mean value of @xmath42 operator for the following ( 3,1 ) states are : 1.97 , 1.04 , 1.03 , 1.97 @xmath49 $ ] .
these values are close to the ones obtained for the two electron system .
the absence of total spins of two electrons shows that two electrons form a singlet state with zero total spin .
let us extract the coefficients for each slater determinant that is used to create configuration interaction basis .
for the subsequent states whose energy levels are depicted in fig .
[ spectr4e](a ) the only non - zero ( and nearly equal to unity , the other coefficients are less than 0.006 ) are coefficients for slater determinants consisting of following single electron orbitals : @xmath99 , @xmath100 , @xmath101 and @xmath102 respectively .
the corresponding orbitals are depicted in fig .
[ 1efor4e](a , b , c ) .
let us assume now that two electrons of the four - electron system form a singlet closed - shell that does not impact the spin properties of the two remaining electrons and thus they can be separated away : we exclude from the configuration interaction basis the @xmath103 orbitals and limit number of electrons in the calculation to two .
the obtained energy levels are depicted with the red - dashed curves in fig .
[ spectr4e](a ) .
besides the shift between the energy levels obtained in full four - electron and two - electron calculation with restricted basis the spectra perfectly match . the total spins and the zeeman splittings in the four - electron energy spectra
resemble the ones obtained for two electrons . however , the striking feature of spectrum of fig .
[ spectr4e](a ) is that the two horizontal energy levels become separate in non - zero magnetic field already for a symmetric system .
let us invoke the two - electron approximation with the restricted basis to explain this observation .
the @xmath103 orbitals are occupied by two spin - opposite electrons that form the singlet state and are separated from the basis .
the two energy levels that slightly split in the magnetic field are constructed from an @xmath104-like orbital of electron localized in the left dot ( @xmath105 ) [ see fig . [ 1efor4e](b ) ] and an @xmath85-like orbital formed by an electron localized in the right dot ( @xmath106 ) [ see fig . [ 1efor4e](a ) ] .
the single - electron energy levels are depicted in fig . [ 1efor4e](d ) .
we observe that the zeeman splittings between energy levels of @xmath85-states differ from the ones for @xmath104-orbitals : 0.282 mev compared to 0.294 mev . here
the dots are symmetric so it its the shape of @xmath107 that is changed .
we integrate eq .
[ integral ] for an @xmath104-like orbital and obtain , @xmath108 the effective @xmath0-factors are obtained from the energy splittings analogically as in the two - electron case : @xmath109 and @xmath110 . as a result
the two states constructed from @xmath111 and @xmath111 single - electron orbitals have different energies at @xmath112 for @xmath113 . figure [ g1g2](b ) presents @xmath114 .
the plot suggests that the @xmath0-factor ratio can be altered significantly as compared to the two - electron case for an asymmetric system .
namely , if one makes the dot that is occupied by three electrons longer one can amplify the ratio of the @xmath0-factors in the dots greater than elongating the dot with a single electron .
the energy spectra for asymmetric system with @xmath52 nm and @xmath53 nm are presented in fig .
[ spectr4e](b ) .
the splitting between the central lines is visibly increased as compared to the symmetric case of fig .
[ spectr4e](a ) .
we calculate @xmath115 which is close to the value obtained in analytical calculation from fig .
[ g1g2](b ) equal to 0.910 .
figure [ spectr6e](a ) presents the energy spectrum of ( 5,1 ) states for six - electron double dot system for the bias @xmath116 mev .
the energy level structure resembles the spectrum of four - electron system depicted in fig .
[ spectr4e](a ) .
we again obtain the splitting of the central lines already for a symmetric system . for @xmath65 mt
is 7.2 @xmath117ev for four electrons while for six electrons we get 8.6@xmath117ev .
also the total spins of ( 5,1 ) states are similar : 1.97 , 1.04 , 1.03 , 1.97 @xmath49 $ ] . the coefficients for slater determinants extracted from the configuration interaction calculation show that mainly a single determinant ( with the square of absolute value equal to 0.987 ) describes each of the discussed six - electron state .
the single electron states that constitute the determinants are : @xmath118 , @xmath119 , @xmath120 , @xmath121 for the following ( 5,1 ) states from fig .
[ spectr6e](a ) .
the single - electron energy levels are depicted in fig .
[ spectr6e](b ) .
the determinants correspond to the occupation of single - electron orbitals in which two pairs of electrons occupy closed singlet shells : two electrons occupy spin opposite @xmath85-like orbitals and the next pair occupies two spin opposite @xmath104-like orbitals .
the two remaining electrons occupy an @xmath122-like orbital in the left dot and an @xmath85-like orbital in the right dot . the calculated spectrum for the basis excluding the four single electron states that form the two singlet shells is presented in fig .
[ spectr6e](a ) with the red - dashed curves .
the spectra obtained in the exact calculation and in the restricted basis agree . for six electrons
we calculate the ratio @xmath123 where @xmath124 is determined from integration of eq .
[ integral ] with an @xmath122-like orbital and plot it in fig .
[ g1g2](c ) .
we see that it is similar to the four - electron case of fig .
[ g1g2](b ) and is strongly altered as compared to the @xmath8 case .
our work shows that increasing the number of electrons results in an amplification of the difference between effective @xmath0-factors in the dots .
table i shows @xmath125 values taken from the experimental works .
it is clearly seen that the studies that considered @xmath7 electrons indeed measured ratios that deviate more from 1 as compared to the @xmath8 cases .
the actual experimental values could be affected by a number of effects omitted in the present modeling : the detailed structure of the confinement potential , or they can be impacted by non - zero exchange interaction . @xcite nevertheless the tendency drawn by these data is clear and agrees with the result of the present study .
.@xmath125 ratio given in the experimental works versus the number of electrons .
[ cols="^,^,^ " , ] [ t1 ]
we investigated nanowire double quantum dots occupied by an even number of electrons tuned to the pauli spin blockade regime . by the exact configuration interaction study we found that in a system with an even , larger than two , number of electrons all but two electrons form closed singlet shells .
this allows to obtain the properties of these structures by configuration interaction calculation where the number of electrons is limited to two and where @xmath126 lowest in energy single - electron orbitals forming singlet shells are excluded .
despite the fact that for @xmath7 the properties of the system are controlled by only two electrons the dots with such occupation can not be treated as an exact equivalent of two - electron systems .
we found that the occupation of excited single - electron orbitals by the valence electrons results in different effective @xmath0-factors in the adjacent dots .
for @xmath7 the difference is obtained already for a symmetric system while for two - electrons it results from the dots asymmetry .
the differences of effective @xmath0-factors present in our results are observed in recent edsr experimental studies on double - quantum dots .
this work was supported by the funds of ministry of science and higher education for 2014 and by pl - grid infrastructure .
00 r. hanson , l. p. kouwenhoven , j. r. petta , s. tarucha , and l. m. k. vandersypen , rev .
mod . phys . * 79 * , 1217 ( 2007 ) .
d. loss and d. p. divincenzo , phys .
a * 57 * , 120 ( 1998 ) . c. fasth , a. fuhrer , l. samuelson , v. n. golovach , and d. loss , phys .
rev . lett . * 98 * , 266801 ( 2007 ) .
a. pfund , i. shorubalko , k. ensslin , and r. leturcq , phys .
b * 76 * , 161308(r ) ( 2007 ) . k. c. nowack , f. h. l. koppens , yu . v. nazarov , and l. m. k. vandersypen , science * 318 * , 1430 ( 2007 ) . v. n. golovach , m. borhani , and d. loss , phys .
b * 74 * , 165319 ( 2006 ) .
e. a. laird , c. barthel , e. i. rashba , c. m. marcus , m. p. hanson and a. c. gossard , semicond .
* 24 * , 064004 ( 2009 ) . f. h. l. koppens , c. buizert , k. j. tielrooij , i. t. vink , k. c. nowack , t. meunier , l. p. kouwenhoven and l. m. k. vandersypen , nature * 442 * , 766 ( 2006 ) .
e. n. osika , b. szafran , and m. p. nowak , phys .
b * 88 * , 165302 ( 2013 ) .
s. nadj - perge , s. m. frolov , e. p. a. m. bakkers and l. p. kouwenhoven , nature ( london ) * 468 * , 1084 ( 2010 ) .
m. d. schroer , k. d. petersson , m. jung , and j. r. petta , phys .
lett . * 107 * , 176811 ( 2011 ) .
s. nadj - perge , v. s. pribiag , j. w. g. van den berg , k. zuo , s. r. plissard , e. p. a. m. bakkers , s. m. frolov , l. p. kouwenhoven , phys .
lett . * 108 * , 166801 ( 2012 ) .
s. m. frolov , j. danon , s. nadj - perge , k. zuo , j. w. w. van tilburg , v. s. pribiag , j. w. g. van den berg , e. p. a. m. bakkers , and l. p. kouwenhoven , phys . rev . lett . * 109 * , 236805 ( 2012 ) . j. w. g. van den berg , s. nadj - perge , v. s. pribiag , s. r. plissard , e. p. a. m. bakkers , s. m. frolov , and l. p. kouwenhoven , phys . rev
. lett . * 110 * , 066806 ( 2013 ) .
k. d. petersson , l. w. mcfaul , m. d. schroer , m. jung , j. m. taylor , a. a. houck , and j. r. petta , nature ( london ) * 490 * , 380 ( 2012 ) .
h. tasaki , j. phys . : condens .
matter * 10 * , 4353 ( 1998 ) .
e. lieb and d. mattis , phys . rev . *
125 * , 164 ( 1962 ) . c. e. pryor , and m. e. flatte , phys
lett . * 96 * , 026804 ( 2006 ) .
y. v. pershin , j. a. nesteroff , and v. privman , phys .
b * 69 * , 121306(r ) ( 2004 ) ; c. flindt , a. s. sorensen , and k. flensberg , phys . rev . lett . * 97 * , 240501 ( 2006 ) ; j. phys . :
. ser . * 61 * , 302 ( 2007 ) .
y. a. bychkov and e. i. rashba , j. phys .
c * 17 * , 6039 ( 1984 ) .
m. p. nowak and b. szafran , phys .
b * 87 * , 205436 ( 2013 ) .
s. bednarek , b. szafran , t. chwiej , and j. adamowski , phys .
b * 68 * , 045328 ( 2003 ) .
y. l. luke , _ the special functions and their approximations _
( academic , new york , 1969 ) , vol . 1 . | nanowire double quantum dots occupied by an even number of electrons are investigated in the context of energy level structure revealed by electric dipole spin resonance measurements .
we use numerically exact configuration interaction approach up to 6 electrons for systems tuned to pauli spin blockade regime .
we point out the differences between the spectra of systems with two and a greater number of electrons . for two - electrons
the unequal length of the dots results in a different effective @xmath0-factors in the dots as observed by the recent experiments . for an increased number of electrons the @xmath0-factor difference between the dots appears already for symmetric systems and
it is greatly amplified when the dots are of unequal length .
we find that the energy splitting defining the resonant electric dipole spin frequency can be quite precisely described by the two electrons involved in the pauli blockade with the lower - energy occupied states forming a frozen core . |
most microscopic studies of heavier nuclei are based on mean - field methods such as density functional theory .
however , important correlation effects beyond the mean field can be missed . the configuration - interaction ( ci )
shell model approach accounts for correlations and shell effects but conventional diagonalization methods are limited by the size of the many - particle model space .
the auxiliary - field monte carlo ( afmc ) method enables calculation of thermal observables and ground state properties in model spaces that are many orders of magnitude larger than model spaces that can be treated by conventional diagonalization methods . in nuclear physics this method is also known as the shell model monte carlo ( smmc ) method @xcite .
smmc has proven to be a powerful method in calculating statistical and collective properties of nuclei @xcite . here
we review recent advances in the development and application of smmc to nuclei .
in particular , we describe a method to circumvent the sign problem that arises from the projection on an odd number of particles and its application to calculate pairing gaps in mid - mass nuclei and level densities of odd - mass isotopes .
we also discuss the application of smmc to describe the microscopic emergence of collectivity in heavy rare - earth nuclei and present results for the collective enhancement factors of level densities in such nuclei .
the afmc method utilizes the hubbard - stratonovich transformation @xcite , in which the imaginary - time propagator @xmath0 of a nucleus that is described by a hamiltonian @xmath1 at inverse temperature @xmath2 is represented by a functional integral of one - body propagators of non - interacting nucleons over external auxiliary fields @xmath3
g_\sigma u_\sigma \;.\ ] ] here @xmath4 is a gaussian weight and @xmath5 is the propagator of non - interacting nucleons moving in time - dependent auxiliary fields @xmath6 .
the thermal expectation value of an observable @xmath7 is @xmath8 w_\sigma \phi_\sigma \langle o \rangle_\sigma \over \int d[\sigma ] w_\sigma \phi_\sigma } \;,\ ] ] where @xmath9 is a positive - definite weight function and @xmath10 is the monte carlo sign function .
the quantity @xmath11 is the thermal expectation value of the observable for a given configuration @xmath12 of the auxiliary fields . in smmc , we choose samples @xmath13 distributed according to @xmath14 and estimate the observables from @xmath15 .
afmc often suffers from a sign problem at low temperatures .
however , there are classes of good - sign interactions that are free of the sign problem and are composed of the dominant collective components of effective nuclear interactions @xcite .
small bad - sign components can be treated by the extrapolation method of ref .
applications of smmc to odd - even and odd - odd nuclei have been hampered by a sign problem that originates from the projection on a odd number of particles .
such a sign problem occurs at low temperatures even for good - sign interactions and makes it difficult to estimate accurately the ground - state energy of an odd - particle system .
a breakthrough was a method we introduced to calculate accurately the ground - state energy of an odd - particle system that circumvents this sign problem @xcite .
it is based on the scalar imaginary - time single - particle green s functions @xmath16 here @xmath17 denotes time ordering , @xmath18 denotes the radial quantum number @xmath19 , orbital angular momentum @xmath20 and total spin @xmath21 of the single - particle levels , and @xmath22 is the magnetic quantum number of @xmath21 . in an asymptotic regime of sufficiently large @xmath23 but @xmath24 , the green s functions of an even - even nucleus @xmath25 have the form @xmath26|\tau|}$ ] , where @xmath27 is the lowest energy of spin @xmath28 in the neighboring even - odd nuclei @xmath29 and @xmath30 is the ground - state energy of the nucleus @xmath25 .
thus @xmath27 can be determined from the slope of the calculated @xmath31 and the ground - state energy of the even - even nucleus .
the ground - state energy of the even - odd nucleus is then found by minimizing @xmath32 over @xmath21 .
we used this method to calculate neutron pairing gaps in mid - mass nuclei from the second - order energy difference in neutron number @xmath33/2 $ ] , where @xmath34 is the ground - state energy of the nucleus with @xmath35 protons and @xmath36 neutrons .
[ pairing - gaps ] shows neutron pairing gaps for mid - mass isotopes .
the calculations were carried out in the complete @xmath37 shell with the interaction of ref .
the smmc results ( solid circles ) are in overall good agreement with the experimental values ( open circles ) .
smmc has been successful in calculating nuclear state densities @xmath38 versus excitation energy @xmath39 in the presence of correlations .
the density calculated in smmc is the _ state _ density , in which each level with spin @xmath40 is counted @xmath41 times .
however , often the density measured in the experiments is the _ level _ density , in which each level with spin @xmath40 is counted once .
we recently introduced a method to calculate the level density directly in smmc @xcite . denoting by @xmath42 the level density at given value @xmath43 of the angular momentum component @xmath44 , the level density @xmath45
is given by @xmath46 for even - mass nuclei and by @xmath47 for odd - mass nuclei .
@xmath44 projection can be carried out exactly in smmc as discussed in ref .
@xcite .
level densities were recently measured by proton evaporation spectra in a family of nickel isotopes @xmath48ni @xcite .
we calculated these level densities @xcite within the @xmath37 shell using the hamiltonian of ref .
the calculation of the ground - state energies of the odd - mass nickel isotopes , and hence their excitation energies , was made possible by the method discussed in sec .
[ odd - particle ] .
[ ni - level ] shows the level densities of @xmath48ni .
the smmc results are in close agreement with the level densities determined from various experimental data sets .
heavy nuclei are known to exhibit various types of collectivity .
typically , nuclei near shell closure are spherical and display vibrational collectivity , while mid - shell nuclei are strongly deformed and display rotational collectivity .
different types of collectivity are well described by empirical models but much less is understood microscopically .
here we use smmc to describe the emergence of collectivity in heavy nuclei within the framework of the ci shell model approach .
a given type of collectivity is usually identified by its associated spectrum .
while smmc is a powerful method for calculating thermal and ground - state properties in very large model spaces , it does not provide detailed spectroscopic information . to overcome this difficulty ,
we have identified a thermal observable whose low - temperature behavior is sensitive to the type of collectivity .
such an observable is @xmath49 , where @xmath50 is the total angular momentum . assuming an even - even nucleus with a rotational or a vibrational ground - state band and an excitation energy @xmath51 of the first @xmath52 state , we have @xcite @xmath53 we carried out calculations for rare - earth nuclei using the @xmath54 shell plus @xmath55 for protons , and the @xmath56 shell plus @xmath57 for neutrons . the hamiltonians we used are given in refs .
the left panel of fig .
[ j2 ] shows @xmath49 for the deformed rare - earth nucleus @xmath58dy .
the smmc results are compared with fits to both the rotational and vibrational models .
the agreement of the smmc results with the rotational model confirms that we are able to describe the rotational character of @xmath58dy in the framework of a truncated spherical shell model approach .
the fit to the rotational formula gives @xmath59 kev , in agreement with the experimental result of @xmath60 kev .
the right panel of fig .
[ j2 ] shows @xmath49 for the spherical rare - earth nucleus @xmath61sm . here
the smmc results follow closely the vibrational formula fit with @xmath62 kev , in agreement with the experimental value of @xmath63 kev .
heavy nuclei are known to exhibit a crossover from vibrational to rotational character as the number of neutrons increases from shell closure towards mid - shell . in a mean - field approximation , this is described by a phase transition from spherical to deformed nuclei .
we have studied two such families of rare - earth isotopes , @xmath64sm and and @xmath65nd @xcite .
[ nd - j2 ] shows @xmath49 for a family of even neodymium isotopes .
the open circles are the smmc results .
we observe a crossover from a soft response to temperature in @xmath66nd , typical of a vibrational nucleus , to a rigid response in @xmath67nd , typical of a rotational nucleus .
the solid lines are extracted from experimental results using @xmath68 where @xmath69 is the corresponding experimental partition function . here
@xmath70 are the measured low - lying energy levels with spin @xmath71 , which we assume to be complete up to to certain threshold energy @xmath72 . above the threshold energy ,
we use a backshifted bethe formula for the level density @xmath38 whose parameters are determined from a fit to level counting data at low energy and neutron resonance data at the neutron resonance threshold .
we use in eq .
( [ j2-exp ] ) the spin cutoff model estimate @xmath73 , where @xmath74 is calculated assuming a rigid - body moment of inertia @xmath75 .
we observe an overall good agreement between the smmc values of @xmath49 and the experimentally determined values .
smmc state densities in rare - earth nuclei were found to be in good agreement with experimental data @xcite .
collective states in these nuclei lead to enhancement of their state density , which is described by a collective enhancement factor @xmath76 .
the decay of @xmath76 with excitation energy is one of the least understood topics in the modeling of level densities @xcite , and it is often parameterized by empirical formulas .
we define a collective enhancement factor by @xmath77 , where @xmath78 is the total smmc state density and @xmath79 is the level density calculated in the finite - temperature hartree - fock - bogoliubov ( hfb ) approximation .
the latter describes the density of intrinsic states and thus the above ratio is a measure of the enhancement in the number of states due to collective states that are built on top of the intrinsic states .
[ nd - enh ] shows the calculated enhancement factor @xmath76 as a function of excitation energy @xmath39 for the even neodymium isotopes @xmath65nd .
we also show by arrows the energies of the various thermal phase transitions that occur in the hfb approximation .
the thin arrows correspond to the pairing transitions ( for both protons and neutrons ) , while the thick arrows describe the shape transitions .
the spherical nuclei @xmath66nd do not support rotational bands , so the enhancement factor @xmath76 is due to vibrational states alone .
we observe that this vibrational enhancement factor decays to @xmath80 at excitation energies above the proton and neutron pairing transitions .
the heavier neodymium isotopes @xmath81nd are deformed in their ground state .
we observe that @xmath76 has a local minimum above the pairing transitions , which we attribute to the interplay between the decay of vibrational collectivity and the enhancement due to rotational collectivity .
the plateau that follows the local minimum describes the enhancement from rotational states .
this rotational enhancement decays in the vicinity of the shape phase transition .
the smmc is a powerful method for calculating statistical and collective properties of nuclei in the framework of the ci shell model in very large model spaces .
we reviewed a recent method to circumvent the odd - particle sign problem and its application to the calculation of pairing gaps and level densities of odd - mass nuclei .
we discussed recent applications of smmc to heavy rare - earth nuclei and in particular the microscopic emergence of collectivity in such nuclei .
it is a pleasure to dedicate this article to aldo covello in recognition of his important contributions to nuclear structure physics .
this work was supported in part by the doe grant no .
de - fg-0291-er-40608 , by the jsps grant - in - aid for scientific research ( c ) no .
25400245 , and by the turkish science and research council ( tbitak ) grant no .
ardeb-1001 - 112t973 .
computational cycles were provided by the nersc high performance computing facility and by the high performance computing center at yale university .
lang g h , johnson c w , koonin s e and ormand w e 1993 _ phys .
_ c * 48 * 1518 alhassid y , dean d j , koonin s e , lang g h , and ormand w e 1994 _ phys . rev .
lett . _ * 72 * 613 nakada h and alhassid y 1997 _ phys . rev . lett . _
* 79 * 2939 alhassid y , liu s and nakada h 1999 _ phys .
lett . _ * 83 * 4265 alhassid y , liu s and nakada h 2007 _ phys
lett . _ * 99 * 162504 alhassid y , fang l and nakada h 2008 _ phys .
lett . _ * 101 * 082501 zen c , alhassid y and nakada h 2013 _ phys . rev
_ * 110 * 042502 hubbard j 1959 _ phys .
lett . _ * 3 * 77 ; stratonovich r l 1957 _ dokl . akad
s.s.s.r . _ * 115 * 1097 dufour m and zuker a p 1996 _ phys . rev .
_ c * 54 * 1641 mukherjee a and alhassid y 2012 _ phys .
lett . _ * 109 * 032503 alhassid y , bonett - matiz m , liu s and nakada h arxiv:1304.7258 voinov a v , grimes s m , brune c r , massey t and schiller a 2012 _ epj web of conferences _ * 21 * 05002 ; voinov a v ( private communication ) bonett - matiz m , mukherjee a and alhassid y 2013 _ phys
_ c rapid comm .
* 88 * 011302 alhassid y , zen c and nakada h 2014 _ nuclear data sheets _ * 118 * 233 capote r _
_ 2009 _ nuclear data sheets _ * 110 * 3107 , and references therein | the shell model monte carlo ( smmc ) method is a powerful technique for calculating the statistical and collective properties of nuclei in the presence of correlations in model spaces that are many orders of magnitude larger than those that can be treated by conventional diagonalization methods .
we review recent advances in the development and application of smmc to mid - mass and heavy nuclei . |
the past decade has seen a rapid increase in the number and quality of observations of transiting planet atmospheres . since the first observations of light transmitted charb02 and emitted charb05,deming05b by hot jupiters ,
the realm of exoplanet characterization has expanded into a larger sample size of hot jupiters seager10 as well as into smaller planets .
two benchmark smaller planets are the 22.4 @xmath5 warm neptune " gj 436b butler04,gillon07 and the `` super - earth mass '' 6.5 @xmath5 gj 1214b charb09 .
characterizing these atmospheres is a stepping stone towards one day characterizing the relatively thin atmospheres of rocky planets .
in addition , results for nasa s _ kepler mission _ borucki10,borucki11,batalha13 have now shown that 2 - 3 @xmath6 planets , perhaps at the boundary between super - earths with thin outgassed atmospheres and sub - neptunes with accreted envelopes , appear to be an extremely common type of planet howard12,youdin11 .
more recent work suggests that this planet size bin could be the _ most _ common size of planet around sun - like stars , at least on orbits less than 85 days fressin13 , and these planets are also common around m stars dressing13 .
gj 1214b is our nearby example of a 2 - 3 @xmath6 planet , and the question of ``
what kind of planet is it ? ''
is an incredibly interesting one .
it is already known that this question can not be answered from its mass and radius alone . solutions that involve mixtures of rock / iron , water , and h / he gas range from models with a rock / iron core and h / he envelope , to those with massive amounts of water , no h / he , and a small rocky core rogers10 . nettelmann11 and
valencia13 have suggested that water - rich models with a steam atmosphere and no h / he envelope are unlikely on cosmogonic grounds . in models with a h / he envelope , the water ( or water plus other volatiles ) mixing ratio in the envelope or in the core are extremely poorly constrained nettelmann11,valencia13 . as was suggested by millerricci09 and kempton10 ,
probing the outer atmosphere of the low - mass low - density ( lmld ) planets can in principle constrain the composition of the gaseous envelope . this in turn helps to constrain the bulk composition of the planet .
tremendous recent effort has gone into characterizing gj 1214b in particular , since it is the lowest mass transiting planet with an atmosphere than can currently be characterized .
bean10 were the first to obtain a transmission spectrum , which showed a transit radius that was essentially flat at red optical wavelengths .
many additional observations from the ground bean11,carter11,croll11,crossfield11,demooij12,narita12,teske13,narita13 and from space desert11,berta12,fraine13 have for the most part confirmed this view of a relatively featureless spectrum . the relatively flat spectrum from the blue to the mid infrared has been suggested as being due to either a high mean molecular weight ( @xmath7 ) atmosphere , which shrinks the scale height and transmission spectrum features millerricci09 , or due to some kind of gray atmospheric condensate , which would obscure molecular absorption features .
the viability of equilibrium condensates was first briefly explored in kempton12 .
the viability of non - equilibrium condensates ( such as created by the photolysis of methane and perhaps other molecules ) has also been investigated .
howe12 explored a number of photochemical hazes as a way to flatten the model spectrum .
more recently morley13 found that a range of viable equilibrium condensates , including kcl and zns , and methane - derived hazes , using output from kempton12 models , could readily flatten the planet s spectrum , even down to solar metallicity compositions . at this point
it is still fair to say that the flat transmission spectrum is degenerate with regard to its explanation .
the discovery and atmospheric characterization of additional lmld planets is one powerful way to break this degeneracy . in this paper
we explore the phase space of lmld planets from @xmath8 @xmath5 as a function of atmospheric heavy element enrichment and planetary atmospheric temperatures .
hotter versions of gj 1214b ( which has a maximum @xmath3@xmath9 k ) may be warm enough that carbon chemistry will be dominated by co , rather than ch@xmath10 .
this will eliminate a pathway to haze formation and could allow molecular absorption features to be seen , if a haze is what currently obscures them in gj 1214b . furthermore , with the aid of state of the art population synthesis planet formation models mordasini12a ,
mordasini12b we also investigate the expected abundance of metals ( or heavy elements " ) in the atmospheres of these lmld planets , and explore how a wide range in the mass fraction of metals in the envelope and atmosphere ( @xmath11 ) will lead to changes in the transmission spectrum , as a function of planet mass .
we show that a high @xmath11 , and corresponding high atmospheric mean molecular weight ( @xmath7 ) appears to be a natural outcome of the planet formation process of small planets .
as planet masses increase beyond that of gj 436b , the predicted metal enrichment of the atmospheres decreases , thereby decreasing the mean molecular weight , which would also allow molecular absorption features to be better seen .
the goal of the paper is to provide a pathway for current and future planetary characterization over a wide phase space of planetary temperature , mass , and envelope composition .
while previous works , such as spiegel10 have investigated the structure and spectra of atmospheres in the neptune mass range , here we are interested in a wider mass range , with a strong focus on composition and diversity . as the sample size of characterized low - mass low - density planets grows we will be able to determine the roles ( if any ) of atmospheric clouds / hazes and high @xmath11 in making transmission spectra of lmld planets challenging .
this is a diverse population .
figure [ tempmass ] shows the range of planet masses vs. planet equilibrium temperatures for transiting and radial velocity planets below 30 @xmath5 .
we have computed models of the atmosphere of gj 1214b and similar planets using a collection of model atmosphere methods that have been widely used to model the atmospheres of exoplanets , brown dwarfs , and solar system planets .
the fully non - gray atmosphere code has been used to derive model pressure - temperature ( _ p t _ ) profiles for the atmosphere of titan mckay89 , uranus mm99 , brown dwarfs marley96,burrows97,saumon08 , and irradiated exoplanets fortney05,fortney08a , cahoy10,morley13 .
the radiative transfer methods are described in toon89 .
we use the opacity database of freedman08 , the `` correlated - k '' method of opacity tabulation described in goody89 , the equilibrium chemical abundances tables of k. lodders and collaborators lodders02,lodders06,lodders09 , and the solar system abundances of lodders03 .
the code we use to model transmission spectra is fully described in hubbard01 , fortney10b , and shabram11 .
in shabram11 the code was tested against the analytical relations of lecavelier08a .
the transmission spectrum code uses wavelength - by - wavelength opacities from the freedman08 database . to showcase baseline models ,
we have computed transmission spectra using the planet - wide average one - dimensional solar metallicity _ p
profile from kempton10 .
we use this model , and also a version of it where the mixing ratio of h@xmath12o is increased to 0.4 , and a version where the water mixing ratio is 1.0 ( pure steam ) .
these models are shown in figure [ full ] . as already noted by several authors , the gj 1214b model with a solar metallicity cloud - free atmosphere
is ruled out .
berta12 , fitting to the highest s / n ground - based and space - based observations , suggested that water mixing ratio greater than 0.4 best matched their _ hst _ spectrum and all other available data , and we plot such a model , along with a pure water ( steam ) model for comparison .
additional observations since berta12 have not materially changed their constraints .
the plot shows the effect of increasing values of @xmath7 , which we will discuss later in the paper .
we wish to examine at what planetary temperatures ( and hence , at what level of incident flux ) pathways towards haze formation may be most likely . zahnle09b and kempton12 have shown that the photochemical destruction of ch@xmath10 in atmospheres just cooler than the hot jupiters ( @xmath13@xmath14 k or less ) can lead to the creation of higher order hydrocarbons , which may progress to the eventual formation of `` soot '' or `` haze '' particles that could obscure the atmosphere howe12,morley13 .
such haze is seen in jupiter s atmosphere , particularly at the poles [ e.g.,][]rages99 . we have calculated _ p t _
profiles for mature gj 1214b - like planets at a range of incident flux levels , from 0.3@xmath15 to 30@xmath15 that currently experienced by gj 1214b .
these profiles utilize a metallicity that is 50@xmath15 solar and the gj 1214a parent star .
these models are shown in , compared to the chemical equilibrium equal abundance curve of ch@xmath10 and co. atmospheres that are everywhere to the right of this curve would be expected to be co - dominated , with a relatively small ch@xmath10 mixing ratio .
we extend the non - equilibrium chemistry calculations of kempton12 and morley13 here for versions of gj 1214b at higher and lower incident fluxes .
the models were run using 50@xmath15 solar metallicity and two values of the eddy diffusion coefficient , @xmath16 .
our aim is to understand better at what incident flux level the atmospheric carbon chemistry changes from being ch@xmath10-dominated to co - dominated .
as described in zahnle09b and morley13 , the photolysis of ch@xmath10 leads to the creation of c@xmath17h@xmath18 , higher order hydrocarbons , which are precursors to the creation of upper atmosphere aerosols , which may be described as `` hazes '' or soots . "
since it has been demonstrated that such hazes can blanket the transmission spectrum howe12,morley13 it is important to understand the planetary temperature beyond which this regime is passed .
figure [ c2h2 ] shows the mixing ratio of acetylene ( c@xmath12h@xmath12 ) , the simplest higher order hydrocarbon .
we can examine its abundance as a function of @xmath3 as a proxy for the abundance of all higher order hydrocarbons .
these calculations use two values of the eddy diffusion coefficient , @xmath16 , which parameterizes the vigor of vertical mixing .
higher values of @xmath16 here lead to higher mixing ratios of ch@xmath10 transported from depth to the upper atmosphere , which can then be destroyed by uv flux .
therefore any boundary in @xmath3 between potentially haze - poor and haze - rich models depends on the unknown @xmath16 value . for the high @xmath19 @xmath20 s@xmath21 case
, the warm versions of gj 1214b show abundant acetylene up until the model with 10@xmath15 higher incident flux , beyond that boundary the acetylene abundance in the upper atmosphere drops off precipitously . for the lower @xmath22 @xmath20s@xmath21 case ,
the boundary is at lower incident fluxes , between the 3@xmath15 and 10@xmath15 flux models . taken together , these models suggests a boundary in @xmath3 at @xmath4 k where the potential effects of a haze layer could wane .
this echoes the results of zahnle09b , who suggested a boundary of 1200 k , based on isothermal _ p t _ profiles .
we therefore suggest that the hottest lmld planets may not posses haze layers that could blanket their transmission spectrum , although further work on these complex photochemical problems is needed .
in particular , extremely metal - rich atmospheres may be ch@xmath10 poor even at low temperatures below 800 k moses12 .
also a wider variety of parent stars could be considered , although planets around red dwarfs are the most amenable to characterization .
the planet formation process is chaotic and complex , as planetesimals merge to become embryos , and then planets .
the composition of the planets themselves can change as smaller objects are accreted before the gas and planetesimal disk is cleared away .
our understanding of the accretion of planets is still developing , but we are at the point now where it is possible to directly compare models of planet formation to the distribution and composition of detected planets
. what we are interested in is understanding what @xmath11 may be for gj 1214b , how this value may vary for other planets of similar mass , and how @xmath11 will vary as a function of planet mass .
we have conducted population synthesis calculations to obtain theoretical predictions for the heavy element fraction in h / he envelopes .
the population synthesis calculations are very similar to the ones presented in mordasini12b .
the planets are synthesized using an upgraded version mordasini12a of the planet formation model originally presented in alibert04 .
the model combines several simple , standard prescriptions for the growth of the solid core and the gaseous envelope pollack96 , the evolution of the protoplanetary disk lynden74,papaloizou99 , and disk migration lin86,paardekooper10 .
the reader is referred to alibert05 and mordasini12a for details . here in particular
we focus on the accretion of solid planetesimals by the protoplanetary atmosphere . during each calculation of the formation of one synthetic planet
, we simulate at every numerical timestep a collision between the protoplanet and a planetesimal of a given size .
this is done using the core mass and envelope structure of the protoplanet at this moment in time .
the collision is simulated with a semi - analytical model with a basic structure that is inspired by the work of podolak88 , but takes into account the results from studying the impact of sl9 with jupiter .
this event illustrated both the importance of mechanical disruption for massive impactors [ e.g.,][]zahnle94 , boslough94 , and the lower efficiency of thermal ablation for such large bodies field95,svetsov95 .
a short overview of the model is given in alibert05 , mordasini06 , and below .
a more detailed description will be given in future work . to obtain the distribution of where mass and energy is deposited into the envelope
, we integrate the 3d trajectory of the impacting planetesimal under the actions of gravity ( in the two body approximation ) , gas drag , thermal ablation , and mechanical disruption ( flattening and fragmentation ) . the planetesimal is described by its position , velocity , mass , and radius .
these quantities are evolved with a coupled set of differential equations . for the thermal ablation
, we have developed a prescription for the heat transfer coefficient as a function of the flow regime .
we solve the normal shock wave jump conditions as in chevalier94 to get the shock wave radiation .
this controls the efficiency of thermal ablation during hypersonic flight in the denser parts of the atmosphere , which is the decisive regime for large impactors opik58,zahnle92 . for the mechanical disruption
, we use the `` pancake '' model zahnle92 , but couple it to a model for the growth of rayleigh - taylor instabilities that develop on the front side of the impactor that behaves approximately as a fluid at dynamic pressures that exceed significantly the impactor s tensile strength or self - gravity [ e.g.,][]roulston97,korycansky00 .
the model was tested by comparing it to simulations of different impacts , namely of the lost city meteorite revelle79 , the tunguska event chyba93 , impacts of km sized impactors into venus zahnle92 , and most importantly , the sl9 collision with jupiter [ e.g.,][]zahnle94 , boslough94 .
the calculation of the trajectories of the planetesimal yields the radial mass deposition profile .
we thus know the fraction of the accreted solids that is deposited in the envelope while the rest reaches the core directly .
this is done during the entire formation of the planet , giving at the end of a simulation the final core mass , the total mass deposited in the envelope , and the amount of gas h / he that was accreted .
two populations of synthetic planets were calculated which only differ in the assumed size of the planetesimals .
the size of the planetesimals is fixed during one synthesis and always equal to either 1 or 100 km .
this is clearly an idealization , since in reality a size spectrum of different planetesimals will be accreted .
the size spectrum will be a function of time and the distance from the star .
if a protoplanet accretes beyond the iceline , the planetesimals are assumed to have an icy composition , otherwise they are made of rocks .
this influences their density , tensile strength , heat of vaporization , and some other material parameters opik58 , podolak88 , boslough94 .
the impact geometry is head - on .
the initial velocity of the planetesimals relative to the planet is equal to the local escape speed , i.e. , we assume that the random velocities are small . diversity in the formation conditions of the planets , which are manifest in the final envelope enrichment of the planets , comes from a variety of factors , as described in mordasini12a .
these include diversity in : the dust to gas ratio in the protoplanetary disk , according to observed [ fe / h ] distributions in the solar neighborhood , the protoplanetary disk gas mass , according to observed disk mass distributions in rho ophiuchi , the external photoevaporation rate which sets together with viscous evolution the disk lifetimes , chosen in a way that the disks have a lifetime distribution similar as observed , and finally the initial position of the embryo , uniform in the logarithm of the semimajor axis . for the resulting envelope enrichment
we make a number of assumptions ( besides the general model assumptions , like no mass loss or accretion by the planets after the protoplanetary gas disk is gone , see mordasini et al .
first we assume that the matter that is ablated during the passage of the planetesimal does not later on sink to the core .
the actual fate of the solids will depend on the relative timescales of settling and mixing , which are controlled by the solubility of heavy elements in the h / he wilson12 and the state of the envelope ( radiative vs. convective layers ) .
a sinking of a part of the solids initially deposited in the envelope would reduce the envelope metallicity .
alternatively it is also possible that the solids initially ( during the formation phase ) settle deep into the envelope , but at a later stage get mixed homogeneously through the envelope iaroslavitz07 .
second , we assume that the solid core ( formed by impactor material that can directly penetrate through the entire envelope ) does not subsequently dissolve [ see][]guillot04,wilson12 . this process would in contrast increase the envelope metallicity . finally , we caution that the calculations are currently not self - consistent since our planetary structure code is not yet able to handle a compositionally varying equation of state and opacity . while we keep track of the amount of solids deposited in the envelope , we nevertheless add all solids to the ( computational ) core mass .
this set - up corresponds to a similar simplification made in pollack96 .
thus the formation and evolution of the planets is in fact the one of planets with all solids in the ( computational ) core and a pure h / he envelope which is modeled with the eos of scvh , and the solar composition opacities of bell97 and freedman08 .
this must be critically kept in mind , since it has been shown hori11 that the enrichment of the envelope lowers the critical core masses and speeds up gas accretion [ see also][]stevenson82b .
this means that the envelope composition feeds back onto the formation process .
one should further note that compositional gradients can potentially have important implications also for the evolution of planets by inhibiting efficient , large - scale convection stevenson85 , leconte12 .
we finally mention that the formation model , including the part used to simulate the impact , are strongly simplified descriptions of the actual processes .
this means that the resulting envelope enrichments must be regarded as first , rough estimates .
in particular the link between the initial deposition in the envelope , and the atmospheric enrichment after several gyrs of evolution needs to be explored . in figure
[ zenv ] , the blue dots show the envelope enrichment using 100 km planetesimals as a function of planet mass , while the red dots assume 1 km planetesimals . for 100 km
planetesimals , and the lowest planetary masses ( less than a few earth masses ) , @xmath11 is typically relatively low .
this is a consequence of the following : these very low mass planets can only accrete tenuous h / he envelopes since the gas accretion timescale nonlinearly increases for lower core mass [ e.g.,][]ikoma00 .
the 100 km planetesimals can fly through these thin atmospheres without losing much mass , but instead deposit their entire mass directly in the core , leading to low enrichments . while this basic mechanism is plausible ( cf .
the radar - dark `` shadows '' on venus indicating that the planet is opaque " to planetesimals smaller than a certain critical size , zahnle 1992 ) , this is not necessarily a realistic result : if there are additionally smaller planetesimals which contribute important amounts of solids , they would get destroyed already in the tenuous envelopes .
this is indeed the case for the 1 km planetesimals . in the simulations with these smaller impactors ,
the very low mass planets are found to have high primordial enrichments .
for the 100 km planetesimals , @xmath11 then grows with mass to a local maximum of approximately 0.7 at a planet mass of @xmath07 @xmath5 .
this is due to the fact that these planets can accrete a sufficient amount of gas so that the planetesimals get efficiently destroyed in the envelope , but their gas envelope is still not very massive , leading to the high values . for even more massive planets
, @xmath11 starts to decrease again , since now the mass of h / he that can be accreted increases .
giant planets have low @xmath11 , as expected , because gas runaway accretion strongly dilutes the metals in the envelope .
the results for the 1 km planetesimals are qualitatively similar , with the difference that the @xmath11 are usually higher , as expected .
an important common feature of the two populations is that at all masses , there is a wide spread in possible @xmath11 .
if this can be preserved and has an imprint on the atmospheric composition we can observe today , we expect to find a large compositional diversity , too .
this diversity would overlay the general trend of a decreasing envelope enrichment with planet mass characterizing .
a key question that one can examine from the population synthesis output is how @xmath11 changes with planet mass .
clearly figure [ zenv ] shows a tremendous range of planetary properties . this is a clear example of the planetary diversity that should at a
minimum be expected .
the quantity of interest when trying to determine the suitability of a planet for characterization by transmission spectroscopy is a comparison of the area of the atmosphere s annulus to the cross - sectional area of the parent star . in figure
[ temparea ] we examine this area ratio for all known transiting low mass planets .
we have arbitrary used an atmospheric temperature equal to @xmath3 and a height of five scale heights to plot the ratio of the area of the atmosphere annulus to the stellar area .
open circles use the @xmath7 of a solar mix of 2.35 , while filled circles use a simply estimated higher @xmath7 , as a function of mass , based on figure [ zenv ] .
we can crudely estimate a mean @xmath23 as a function of planet mass , further assume it is uniform in the entire envelope , and then convert @xmath23 to @xmath7 .
we do this by assuming that the heavy element component is water , to find the average @xmath7 as a function of planet mass .
the value of @xmath7 for some planets reaches 7 , dramatically shrinking the scale height .
the arrows show the shift for individual planets , which are labeled .
the ratio for gj 1214b falls by a factor of @xmath0 3.5 , although we note that it is still the most promising target .
we note that the apparent magnitude of the parent star is also a particularly important factor in the suitability of characterizing particular planets .
of course figure [ temparea ] is just one realization of the wide range of possible values for @xmath7 for these planets .
we have taken the population synthesis output of figure [ zenv ] and binned the output to find the range of @xmath7 values for different planetary mass bins .
figure [ histo ] shows our results .
several trends are worth noting . for 100 km planetesimals ( in blue ) the lowest - mass bin ( 3 to 5 @xmath5 , upper left ) values of @xmath7 are close to solar , since these planetesimals readily pass through the envelope without ablating .
only for the larger planetary mass bins does @xmath7 begin to increase , due to the greater likelihood of appreciable deposition of the heavy elements into the h / he envelope .
for the 1 km planetesimals ( in red ) values of @xmath7 of @xmath010 are common from 3 to 20 @xmath5 , with a modest trend to lower @xmath7 for the highest mass bin . in nearly all cases and for nearly all envelopes ,
the 1 km planetesimals dissolve into the h / he envelopes , enriching these envelopes .
one should clearly expect a tremendous amount of diversity in @xmath7 for lmld planets , in particular below 15 @xmath5 .
put another way , at this point there is no reason to expect that all planets with the mass of gj 1214b should show a similar atmospheric metallicity enrichment .
the size of accreted planetesimals , and their composition , could vary greatly based on planetary @xmath3 , stellar type , formation time , the mass and location of nearby planets , and other parameters . within the solar system
the planets uranus and neptune are the closest to lmld planets in log(mass ) . while it could be an overstatement to suggest that lmld planets are all smaller version of these planets , it is worthwhile to review our current understanding of the structure and atmospheres of these planets .
a window into interior structure that is only available in the solar system is the gravity field of the planets , which yields constraints on planetary density as a function of radius .
recently , nettelmann13 have investigated constraints on the interiors structure , including uncertainties in the gravity fields , shapes , and rotation rates of the planets .
`` standard '' three - layer models were calculated , including a rock / metal core , a middle layer composed mostly of water , and an outer layer .
the water mass fraction of the outer envelope ( @xmath24 ) of up to 0.7 are acceptable for neptune , and 0.2 for uranus .
it is tempting to compare these numbers directly to the model productions of @xmath11 from the population synthesis models .
however , this must be done with great care .
such three layer models are likely artificial , and if the interior structure is more complex ( see [ evap ] ) it is no longer clear what @xmath11 for these planets should be .
one can also compare the molecular abundances in the atmosphere of uranus and neptune to the elemental abundances in the sun .
since the planets are quite cold , most volatile elements are condensed into clouds far below the visible atmosphere .
methane is only partly condensed , so it is possible to derive its abundance in the troposphere from infrared spectroscopy .
data from baines95 , along with the revised solar abundances [ e.g.,][]lodders03 , suggest an enrichment of carbon of @xmath25 solar for neptune fletcher10 and similar number of @xmath26 solar for uranus .
furthermore , the detection of co in the atmosphere of neptune , brought up from deeper layers by vertical mixing , suggests an even larger mixing ratio of water .
past and current modeling work fegley94b , luszcz13 suggest that the atmospheric oxygen abundance of neptune could be as large as @xmath27 solar . assuming a @xmath7 for the heavy elements of 18 amu ( correct for water ) @xmath28 solar would lead to @xmath29 and @xmath30 solar is @xmath31 [ see , for instance figure 6 of][and our appendix]nettelmann11 .
these very high values of @xmath32 in the h / he atmosphere of neptune are similar to what was found from the gravity fields of the planet [ @xmath32 values up to 0.7 are allowed][]nettelmann13 .
this suggests that s it is reasonable for gj 1214b and lmld planets as a class to have a similar @xmath32 in their h / he envelopes , or perhaps even higher .
there are several lines of observational evidence that h@xmath12 dominated exoplanet atmospheres are not stable against atmospheric evaporation .
most prominent are the observations that hot jupiters lose atmospheric mass vidal03,linsky10,lecavelier10,lecavelier12 , as has been observed in the ultra - violet with the _ hubble space telescope_. a more subtle but emerging trend is the incident flux vs. planet density plane , which shows a lack of lmld planets at high incident fluxes lopez12,lopez13 . at the time of the discovery of gj 1214b , it was clear that atmospheric evaporation could be important in shaping the planet that we see today charb09,rogers10,nettelmann11 .
since it is likely that lmld planets lose mass , and that a high @xmath7 atmosphere could explain the gj 1214b transmission observations , it is natural to wonder if _ differential _ mass loss can occur . if hydrogen is lost preferentially , the atmosphere could become enriched in heavy elements with time .
however , we find that this is not possible for a hydrogen - dominated atmosphere . [ see , e.g.][for a review.]hunten73 .
we show this by examining the diffusion - limited flux for hydrogen with respect to other gases .
the maximum escape flux of h@xmath12 ( component @xmath33 , for atmosphere ) through a non - escaping minor gas ( component @xmath34 ) is : @xmath35 @xmath36 where @xmath37 is the h@xmath12 flux , @xmath38 ( cm@xmath21s@xmath21 ) is the binary diffusion coefficient between h@xmath12 and the minor gas , and @xmath40 and @xmath41 are the masses of h@xmath12 and the minor gas , respectively . for h@xmath12 and either co or ch@xmath10 , typical minor gases ,
the binary diffusion coefficient at relevant temperatures is @xmath42cm@xmath21 marrero72 .
the time scale for diffusion - limited h@xmath12 escape of a giant planet is @xmath43 = @xmath44 . with @xmath45 ,
@xmath43 becomes : @xmath46 which has been evaluated for co. it would be longer still for ch@xmath10 .
equation [ mdot ] depends only weakly on temperature , and is independent of other planetary parameters .
the long time scale in equation [ mdot ] means that selective escape of hydrogen from other gases is effectively impossible when h@xmath12 atmospheres evaporate .
of course this statement does not apply to selective escape of gases with respect to condensates , which can form droplets that rain down through the wind .
while we have so far assumed h / he envelopes with uniform heavy element mass fractions , this may well be too simple .
atmospheric evaporation can expose deeper atmospheric layers which could be enhanced in metals due to a composition gradient .
is has long been suggested that uranus and neptune may feature a composition gradient , with the fluid planetary `` ice '' becoming gradually more abundant with depth [ e.g.][]hubbard91 .
such interior structures with composition gradients may suppress convection and could lead to thermal evolution models that better match uranus s current @xmath13 , compared to the three - layer models in [ unsec ] . whether or not there are well - defined layers , it is certainly true that the h / he mixing ratio decreases with depth in both planets .
it is worth exploring a small range of models for uranus and neptune with a composition gradient to assess how quickly @xmath11 may rise with depth inside a @xmath015 @xmath5 planet .
we can use the methods of @xcite to suggest possible interior structures .
figure [ un2 ] shows viable models ( meaning all observational constraints are satisfied ) of the structure of uranus and neptune can be calculated that have a gradient in @xmath11 .
these use a solar @xmath32 in the visible atmosphere , but a gradient in log ( pressure ) below that . instead of a sharp change in @xmath32 at the interface between layers , the value of @xmath32 can rise sharply as a function of depth
. if such models are close to reality , and if a planet with a structure like this were placed into a relatively close - in orbit , some amount of the outer layers would be lost , exposing a higher @xmath11 as time goes on .
while this is not a point that we can quantify for exoplanets at this time , it seems likely that mass loss from sub - neptune and neptune - class planets could lead to a potentially dramatic increase in @xmath11 with time .
a natural question that arises is whether such high values of @xmath32 is the envelope of lmld planets is consistent with their measured masses and radii .
this question has really only be extensively investigated for gj 436b nettelmann10 and gj 1214b rogers10,nettelmann11,valencia13 .
in particular , only nettelmann11 and valencia13 have investigated detailed thermal evolution including different values of @xmath11 .
thermal evolution models are generally important when comparing to observations since the equation of state of the h - he component is quite temperature sensitive leading to significant contraction of the h - he envelope with time [ e.g.][]lopez13 . calculations of thermal evolution models of all of the current lmld planets , with a variety of @xmath11 values is beyond the scope of this work .
however , we can perform simpler calculations to yield plausibility to our suggestion of metal - rich atmospheres , in particular for the lower mass planets . in figure
[ mr ] we plot model mass - radius curves for planets with rock - iron cores and envelopes made of h - he and water .
the core mass fraction @xmath47 is varied from 0.9 to 0.7 , while the @xmath11 is varied from 0.5 to 0.8 , using a water eos french09 .
we have arbitrarily set the radiative - convective boundary at 100 bars , in the absence of a lack of understanding of how interior cooling may slowed by these extremely high opacity atmospheres .
previous work at solar metallicity for planets of similar mass suggest boundaries from several hundred bars to 1 kbar at gyr ages fortney07a .
figure [ mr ] shows that high values of @xmath47 and @xmath11 readily match the lmld planets in the lower left , which was already known for gj 1214b , which can be modeled as an entirely water planet [ e.g.][]nettelmann11 , valencia13 .
as either @xmath47 or @xmath11 are decreased , this leads to a larger mass fraction of h - he gas , with leads to larger radii . for the population of known planets , generally lower values of @xmath47 and/or @xmath11
are needed as planet mass increases .
this may indicate that more massive envelopes are being accreted , or that these envelopes are less metal - rich . beyond @xmath015 @xmath5
there is tremendous diversity in planetary properties .
kepler-30d , at 23@xmath5 and nearly 9 @xmath6sanchis12 , is perhaps 70% h - he by mass batygin13 .
we might naturally expect that planets like gj 436b at 4.3 @xmath6 are metal - rich , while planets like kepler-30d are relatively metal - poor .
such a scenario is certainly observationally testable .
the 22 @xmath5 planet gj 436b has been observed extensively to understand better its atmosphere . _
spitzer _ at occultation ( secondary eclipse ) has probed the dayside spectrum stevenson10,beaulieu11 .
a variety of observations have been made of the transmission spectrum ballard10,pont09,alonso08,caceres09,knutson11 .
we have seen in figure [ histo ] that values of @xmath7 from @xmath0 4 to 10 are quite possible .
model interpretation of the dayside data at the time of occultation has focused on the apparent absence of ch@xmath10 and abundance of co and co@xmath12 madhu11a , moses12 .
in particular , preliminary calculations by moses12 suggest that an atmosphere many hundreds of times solar ( or higher ) is most consistent with the derived carbon chemistry mixing ratios .
our suggestion of high atmospheric metal mass fractions for lmld planets certainly concurs with this work .
we can briefly examine model transmission spectra of the planet and compare it to observational data .
we will predominately focus on data from _ epoxi _ and _ spitzer _ , as it has the smallest error bars . in figure
[ 436 ] , shown in green , is a transmission model with a 50@xmath15 solar metallicity atmosphere , and a corresponding @xmath48 .
the optical transit depth is badly mis - matched by the model , while it reasonably matches the _
spitzer _ data .
clearly a relatively flat spectrum would be a better match to the observations .
this immediately suggests higher @xmath7 atmospheres or those with cloud opacity .
the larger radius at 4.5 @xmath7 m compared to 3.6 @xmath7 m also suggests strong co and co@xmath12 opacity knutson11 , as was found on the dayside .
we examine two models that generally match the data better than the 50@xmath15 solar model . in orange
is a model with a high @xmath7=8.9 due to a 40% water mixing ratio , similar to a lower limit suggested by berta12 for gj 1214b .
this flattened transmission spectrum is a reasonable match from the optical to the mid ir . in blue
is a model with an additional 10% mixing ratio of h@xmath12o , co , and co@xmath12 , evenly distributed between the three .
the optical fit is not quite as good , but the 3.6/4.5 @xmath7 m ratio is better reproduced .
we strongly concur with the findings of knutson11 that models with strongly enhanced co / co@xmath12 , and depleted ch@xmath10 ( which would lead to a shallower model transit at 3.6 @xmath7 m ) lead to the best _ spitzer _
it appears that high @xmath7 atmospheres fit the available data reasonably well .
the atmosphere of gj 1214b has generally been modeled as being of solar composition with enhancements of water vapor , or heavy inert molecules such as n@xmath12 howe12 .
the enrichment of the atmosphere of gj 1214b , as well as our solar systems giant planets , by the accretion and dissolution of planetesimals is thought to be a generic outcome of the planet formation process .
the exact composition of such planetesimals may depend strongly on the abundances of the planet - forming disk and the location of the planets within the disk owen99,gautier01a , oberg11,madhu11b .
structure models of uranus and neptune often use an icy mixture of c - h - o - n atoms called synthetic uranus " nellis97 since more than just water is expected to dissolve into the atmosphere . as a suggestion of what might be seen with more precise _ hst _ data for gj 1214b , we have computed transmission spectrum models in figure [ berta ] where the additional , high @xmath7 `` icy '' volatile that we have mixed in are either water , ammonia , methane , or carbon dioxide . with the current data , we do not feel it is wise to try and put limits on possible mixing ratios of molecules , and we compute these models merely as a guide to show where absorptions features are located .
however , j. bean is leading a 60 orbit _ hst _ campaign that will achieve @xmath49 higher signal - to - noise than berta12 .
the goal of the program is to detect transmission absorption features even from a pure steam atmosphere kreidberg13 . clearly with a higher quality spectrum
we may be able to detect molecular features besides those of water , which may provide insight on the details of the planet s atmosphere .
determining the relative abundances of volatiles would be an important constraint on the planet s formation , such that heavy element enrichment should not just viewed simply as a way to decrease the atmospheric scale height . the _ james webb space telescope _
( _ jwst _ ) will be the next observatory to provide transit and eclipse spectroscopic capabilities significantly beyond _ hst _ , and we have simulated transit observations of our model atmospheres with its nirspec @xcite and miri instruments @xcite .
this work is similar to our previous detailed transmission spectrum modeling of _ jwst _ observations for gj 436b shabram11 .
we computed the number of photons detected using a model of the host star , transmission models of the planet , and estimates of the total efficiency ( detected electrons per incident photon ) at each wavelength for the _ jwst _ nirspec @xmath50 @xmath7 m and miri low resolution spectrograph ( lrs ) @xmath51 spectroscopic modes .
photon noise was also computed , and we added it in quadrature to an assumed observational noise floor that is consistent with current knowledge of instrument performance .
the chosen models include three from figure [ berta ] : the 100% water model ( blue ) , the 1@xmath15solar+75% ch@xmath10 model ( green ) , and the 1@xmath15solar+20% co@xmath12 model ( red ) .
we also include a hazy model with 0.25@xmath7m - size soot particles ( magenta ) from morley13 ( their figures 12 and 13 ) . simulated transit depths and 1 @xmath52 error bars for observations of 4 different models are shown in figure [ g1 ] .
a total integration time of 5 hours was used in each simulation , divided into a total of 6000 s on the star before the transit , 6000 s during transit , and 6000 s on the star after transit .
this corresponds to observing 4 transits if the integration duty cycle is 50% .
a model of an m5v star by @xcite was used for gj 1214 , and the error bars are spaced according to the modeled spectroscopic resolution of the 2 instrument modes .
the systematic noise floor was set to 35 parts per million ( ppm ) for nirspec because this was recently achieved by @xcite using the new scanning mode of the g141 grism on _ hst _ wfc3 ( which uses a similar hgcdte detector ) . due to the relatively coarse nirspec spatial sampling below 3 @xmath7 m wavelength ,
a similar scanning mode will have to be implemented on _ jwst _ to obtain a similarly low systematic noise value in the presence of known detector intrapixel response variations and observatory pointing errors .
however , the slitless niriss spectrograph @xcite will sample the _ jwst _ point spread function much more finely and will likely provide low systematic errors at these wavelengths even without a scanning mode .
a systematic noise floor value of 50 ppm was adopted for the miri lrs simulations , shown in figure [ g2 ] , the same value used by @xcite in similar simulations .
more details of the simulation code can be found in @xcite .
it is interesting to note that the two flattest spectra in the wfc3 bandpass , the pure water and hazy models , only begin to significantly diverge in the mid - infrared , when the mie scattering opacity of the small soot particles begins to wane at long wavelengths .
it certainly appears that characterizing relatively high @xmath7 atmospheres of lmld planets with _
jwst _ will be challenging , and we strongly encourage future simulations of exoplanet characterization with all 4 instruments
. that will give the community a better understanding of the time involved to characterize a large sample of planets .
we are just now starting an era that could be termed the rise of the lmld planets .
the ubiquity of these planets has now been shown from kepler data howard12,youdin11,fressin13,dressing13,morton13 .
gj 1214b and gj 436b have recently been joined by gj 3470b bonfils12 and hd 97658b dragomir13 as small transiting planets around bright stars .
it is likely that soon other planets around bright stars will be found and they will be targets for detailed characterization .
formation models are now aiming to understand the frequency of such planets , their formation timescale , and how often they end up in compact orbital configurations hansen12,rogers11,ikoma12,chiang13 .
our experience from gj 1214b shows that such planets may not give up their secrets easily .
we find that this behavior could be a generic feature of the atmospheres of these planets .
our population synthesis models , as well as experience for uranus and neptune , show that large enhancements in metals may well be common for lmld planets .
simulations presented here show values of @xmath11 , the metals mass fraction in the h / he envelope are commonly 0.6 to 0.9 .
this will shrink atmospheric scale heights for planets in the 2 - 4 @xmath6 range .
however , there is incredible diversity within this class , far more than the diversity for planets of tens of earth masses or higher , so some planets should have smaller @xmath11 values .
if lmld planets have @xmath11 composition gradients , as has been suggested for uranus , the gradual loss of the outer regions of the h / he atmosphere could lead to the excavation of atmosphere with progressively larger values of @xmath11 . we also find that increasing the atmospheric @xmath7 via the differential escape of hydrogen from heavy elements is unlikely due to extremely long diffusion times between h@xmath12 and heavier molecules .
extremely high mixing ratios of heavy elements will lead to vast reservoirs of material to form thick and opaque clouds morley13 .
condensate clouds are ubiquitous in planetary atmospheres across wide temperature ranges . morley13
have shown that equilibrium condensate clouds such as zns and kcl may be found in the atmosphere of gj 1214b , as well as hazes derived from methane photo - destruction .
equilibrium clouds could easily move in or out of the visible atmosphere of gj 1214b analogs at hotter and cooler temperatures .
however if photochemical hazes are present , they could obscure the atmosphere across a wide range of temperatures .
a break in behavior could occur for the warmest lmld planets , as we find that planets above @xmath13 of 800 - 1100 k will no longer be methane dominated , but rather co - dominated , which would end the pathway to haze formation . in terms of future theoretical work
, we make several suggestions .
the first is an exploration of the equilibrium and non - equilibrium chemistry of atmospheres strongly enriched in metals as @xmath11 values from 0.6 to 0.9 suggest atmospheres of 100@xmath15 to 400 + @xmath15 solar abundances .
chemistry calculations will be especially important in understanding the thermal emission spectra of the planets , since dayside spectra will not suffer for high @xmath7 values diminishing the atmospheric signatures . if planetesimal accretion includes rocky objects that do not bring hydrogen ( as the ices do ) even higher metallicities are possible , if these rocky planetesimals also ablate within the atmosphere .
we have shown that high atmospheric metallicities are a generic outcome of the mordasini et al
. formation models .
it may be that across the range of lmld planets to warm neptune class objects like gj 436b , that exceptionally metal rich atmospheres are very common , and should be studied in much more detail [ e.g.,][]moses12 .
atmospheric mass loss of h / he envelopes in the extreme metal - rich atmosphere regime should also be investigated .
the cooling provided by abundant metals , in the form of molecules , atoms , and ions , should be quantified .
the high atmospheric @xmath7 together with efficient radiative cooling , could decrease atmospheric mass loss relates below those observed and modeled for h / he dominated hot jupiter atmospheres . with a much better understanding of planet formation and migration
, one could in principle consider characterizing planetary atmospheric composition to understand better a planet s formation location and history .
such work is now happening for gas giants konopacky13 .
one could imagine a gradient in accreted planetesimals that depended on distance from the parent star .
in addition , planets that formed in situ inside the snow line would accrete relatively volatile - deficient planetesimals , while those that formed beyond and migrated closer in would accrete a mix of volatile - rich and -poor planetesimals that changed with time .
exoplanetary science will continue to be an observation driven field .
now that we know the lmld planets are ubiquitous , to better characterize these planets more nearby examples are needed to provide targets for jwst and other platforms .
with the rise of specialized ground - based transit survey for m dwarfs nutzman08,berta12b , giacobbe12,sozzetti13 , and the transiting exoplanet survey satellite [ tess,][]ricker10 , the prospects are extremely good
. we should expect to be rewarded with a rich diversity of planetary properties .
+ jjf acknowledges support from nasa awards nnx09ac22 g , nnx12ai43a , and hst - go-12251.05 .
we had many helpful conversations with mark marley , caroline morley , and eric lopez .
we thank brad hansen and johanna teske for comments on an earlier draft .
we compute the o : h ratio and the mean molecular weight of hydrogen - helium - water mixtures with given mass fractions @xmath53 of hydrogen , @xmath54 of helium , and @xmath32 of heavy elements ( here , water ) .
further given are the molecular weights of each of these components , which we call @xmath55 , @xmath56 , and @xmath57 .
the molecular weights in g mole@xmath21 are atomic h ( @xmath58 ) , helium ( @xmath59 ) , and water ( @xmath60 ) . by definition ,
@xmath61 where @xmath62 , @xmath63 , @xmath64 are the numbers of particles of these species in a given volume @xmath65 with total mass @xmath66 and total number of particles @xmath67 , given by @xmath68 from ( [ xyz ] ) we see the general relations @xmath69 turning from the general to specific relations , we next assume a mixture of hydrogen , helium , and water ( h@xmath12o ) and derive the o : h ratio for two cases , case @xmath70 for atomic hydrogen , where @xmath71 , case @xmath72 for molecular hydrogen , where @xmath73 . in case @xmath70 we have @xmath74 and in case @xmath72 we have @xmath75 so we can write down the o : h ratios as @xmath76 } \\ \mbox{(case $ ii$)}\quad{\rm o : h } = \frac{n_{\rm o}}{n_{\rm h } } & = & \frac{n_z}{2\,n_x+2\,n_z } = \frac{n_z / n_x}{2 + 2\,n_z / n_x } = \frac{\frac{z}{x } ( \mu_x / \mu_z)}{2 + 2[\frac{z}{x } ( \mu_x / \mu_z)]}\:,\end{aligned}\ ] ] in which the @xmath77 ratios from ( [ nxyznxyz ] ) have been inserted to obtain the o : h ratio in terms of @xmath53 and @xmath32 at the far right . for reference , in figure 6 in nettelmann11 , we assumed case ( ii ) because in the h - he envelope of gj1214b and lmld atmospheres generally , hydrogen would be molecular .
the mean molecular weight reads @xmath78 + \mu_z[\frac{z}{x}(\mu_x / \mu_z ) ] } { 1 + [ \frac{y}{x}(\mu_x / \mu_y ) ] + [ \frac{z}{x}(\mu_x / \mu_z)]}\:,\ ] ] where again the ratios from ( [ nxyznxyz ] ) were inserted to obtain @xmath79 as a function of @xmath53 , @xmath54 , and @xmath32 , with @xmath55 , @xmath56 , @xmath57 , chosen according to the current situation , e.g. , @xmath80 ( for molecular hydrogen ) and @xmath81 . for a given composition @xmath83 of atomic heavy elements with respective atomic weights @xmath84 we can write down the corresponding metals mass fraction @xmath32 according to its definition , @xmath85
often , the composition is provided in terms particle abundance ratios @xmath86:h , thus we have @xmath87 of course , the value of @xmath32 must not change whether the abundance ratios refer to atomic hydrogen ( @xmath86:h ) or molecular hydrogen ( @xmath86:h@xmath12 ) , where @xmath86:h@xmath12 = @xmath88:h .
we show this explicitly : @xmath89 we can invert eq . (
[ eq : zdef ] ) to express the mean molecular weight of the heavy elements , @xmath90 in terms of @xmath32 , @xmath91 when using equation ( [ eq : muz ] ) one has to make sure that the heavy element enrichment factors @xmath92:h / @xmath86:h@xmath93 over the solar system abundances @xmath86:h@xmath93 are the same as the ones used to compute @xmath32 . however , the total mean molecular weight of an atmosphere , @xmath7 does depend on whether hydrogen is in molecular or atomic form . if atomic ( _ case i _ ) , we have @xmath94 and if molecular ( _ case ii _ ) , we have @xmath95 both forms are consistent with eq .
( [ eq : barmu ] ) . | we perform modeling investigations to aid in understanding the atmospheres and composition of small planets of @xmath02 - 4 earth radii , which are now known to be common in our galaxy .
gj 1214b is a well studied example whose atmospheric transmission spectrum has been observed by many investigators .
here we take a step back from gj 1214b to investigate the role that planetary mass , composition , and temperature play in impacting the transmission spectra of these low - mass low - density ( lmld ) planets . under the assumption that these planets accrete modest hydrogen - dominated atmospheres and planetesimals , we use population synthesis models to show that predicted metal enrichments of the h / he envelope are high , with metal mass fraction @xmath1 values commonly 0.6 to 0.9 , or @xmath0100 to 400 + times solar . the high mean molecular weight of such atmospheres ( @xmath2 ) would naturally help to flatten the transmission spectrum of most lmld planets
. the high metal abundance would also provide significant condensible material for cloud formation .
it is known that the h / he abundance in uranus and neptune decreases with depth , and we show that atmospheric evaporation of lmld planets could expose atmospheric layers with gradually higher @xmath1 .
however , values of @xmath1 close to solar composition can also arise , so diversity should be expected .
photochemically produced hazes , potentially due to methane photolysis , are another possibility for obscuring transmission spectra .
such hazes may not form above @xmath3 of @xmath4 k , which is testable if such warm , otherwise low mean molecular weight atmospheres are stable against atmospheric evaporation .
we find that available transmission data are consistent with relatively high mean molecular weight atmospheres for gj 1214b and warm neptune " gj 436b .
we examine future prospects for characterizing gj 1214b with hubble and the james webb space telescope . |
in a @xmath0cdm ( the cosmological constant @xmath0 + cold dark matter ) universe , a mildly overdense site which accretes masses from the surrounding with growing comoving radius will never encounter the turn - around moment since its gravity will be incapable of winning over the accelerating expansion of spacetime driven by such inert dark energy as @xmath0 .
therefore , only those overdense sites that have already formed bound objects until the present epoch or at least turned around before the onset of the accelerating phase of spacetime can possess finite turn - around radii , defined as the effective radii of the sites at the moment when their gravitational force begins to hold down the hubble expansion @xcite . in other words , the standard @xmath0cdm cosmology generically predicts the existence of a finite upper limit on the turn - around radii of the bound objects observed in the local universe . pointing out that the turn - around radius of a bound object , @xmath2 , is a well defined linear quantity , @xcite found an analytic expression for the upper limit , @xmath3 , on the possible values of @xmath2 under one simplified assumption that the gravitational collapse after the turn - around moment proceeds isotropically : @xmath4 strictly speaking , @xmath5 in the rhs of equation ( [ eqn : rt_u ] ) denotes the mass enclosed by the turn - around radius @xmath2 the ratio of which to virial mass @xmath6 depends on mass scales and redshifts .
notwithstanding , @xcite suggested that the upper limit of the turn - around radius be expressed in terms of a readily measurable mass like the virial mass since the mass enclosed by @xmath2 can not be independently measured in practice .
see @xcite for the detailed explanations . throughout this paper ,
we approximate @xmath5 in the rhs of equation ( [ eqn : rt_u ] ) by @xmath6 .
equation ( [ eqn : rt_u ] ) predicts that the average turn - around radius on the mass scale @xmath6 stays much lower than @xmath3 .
if a substantially large number of the objects in the local universe should be found to exhibit a bound violation having larger turn - around radii than @xmath3 given in equation ( [ eqn : rt_u ] ) , then it would challenge the @xmath0cdm model . as mentioned in @xcite ,
the upside of using the turn - around radii of the bound objects as a test of the @xmath0cdm model lies in the fact that it requires only local observables whose systematics would be better controlled unlike the standard global diagnostics which usually require to measure the growth and/or expansion history of the whole universe ( see * ? ? ?
* ; * ? ? ?
* and references therein ) . in the conventional approach , the values of @xmath2 of the galaxy groups and clusters observed in the local universe have been estimated by directly measuring the distances to the surfaces of the zero radial velocities of the bound - zone galaxies ( e.g. , see * ? ? ? * and references therein ) .
however , the estimates of @xmath2 based on this conventional method has been known to suffer from low accuracy that is imputable to notoriously large uncertainties in the measurements of the peculiar velocities of the bound - zone galaxies around the groups and clusters ( see * ? ? ?
very recently , @xcite proposed a new methodology to overcome the low accuracy in the estimation of @xmath2 and applied it to a nearby galaxy group ngc 5353/4 @xcite .
their methodology turned largely to the numerical claim of @xcite that the peculiar velocity profile in the bound zone abound a massive group takes on a universal form and that the profile can be readily reconstructed from the two - dimensional spatial distributions of the bound - zone galaxies along the cosmic web .
the turn - around radius of ngc5353/4 estimated by @xcite with their new method was found to commit a bound violation .
that is , the value of @xmath2 of the ngc 5353/4 group has been found to exceed the upper limit @xmath3 given in equation ( [ eqn : rt_u ] ) .
being cautious and conservative , @xcite have suggested that the robustness of the new method as well as the validity of the universal formula for the peculiar velocity profile proposed by @xcite should be thoroughly examined before putting a proper interpretation of their result . as clearly mentioned in @xcite ,
the turn - around radius of an object is a stochastic quantity , varying in a broad range .
equation ( [ eqn : rt_u ] ) puts the upper limit on the _ average _ turn - around radius on the mass scale @xmath6 but not on all turn - around radii of individual objects with virial mass @xmath6 . in other words
, it might not be forbidden even in the @xmath0cdm cosmology to observe an individual object with mass @xmath6 whose turn - around radius exceeds the spherical bound - limit .
what equation ( [ eqn : rt_u ] ) truly implies is that it would be quite rare to witness a bound violation in the @xmath0cdm universe . besides , what has not been properly taken into account in the analytic derivation of @xmath3 by @xcite is the fact that the true gravitational collapse proceeds quite anisotropically @xcite which is responsible for the eventual formation of the cosmic web in the universe @xcite .
although @xcite used the same analytic formula proposed by @xcite to make an estimate of the turn - around radius of ngc 5353/4 along the filament , it has yet to be tested whether or not the peculiar velocity profile constructed along the cosmic web would still be validly approximated by the same formula that was obtained by taking the isotropic average over all directions .
thus , the remaining critical questions are ( 1 ) how rare occasion a bound violation is in a @xmath0cdm universe ; and ( 2 ) how the true gravitational collapse that tends to proceed anisotropically along the cosmic web affects the rarity of the occurrence of bound violation , both of which we attemp to answer in this work . the contents of this paper can be summarized as follows . in section [ sec : review ] a brief review of the methodology of @xcite and its key assumptions are presented . in section [ sec : web ] a comprehensive numerical test of the methodology is presented and the effect of the web environment on the construction of the peculiar velocity profiles is explored . in section [ sec : iso ]
the estimates of the turn - around radii along the cosmic web on various mass scales are presented . in section
[ sec : con ] the final results are summarized and the physical implications are discussed .
let us consider a spatial point at a distance , @xmath7 , from the center of a massive group with virial radius , @xmath8 , in a space expanding at a global rate , @xmath9 .
from here on , the radial component of the peculiar velocity field in the direction from the halo center to the spatial point is called _ the peculiar velocity at @xmath7 _ for short . if the distance @xmath7 is included in the bound zone range of @xmath10 where the subspace expands at a lower rate than @xmath9 due to the net gravitational effect of the main group , the spherically averaged value of of the peculiar velocity , @xmath11 , at @xmath7 is well depicted by the following universal formula with two parameters @xmath12 and @xmath13 @xcite : @xmath14 where @xmath15 is the magnitude of the circular velocity of the main group at its virial radius given as @xmath16 with virial mass @xmath6 .
fitting equation ( [ eqn : vpr ] ) to the numerical results from a high - resolution @xmath17-body simulation , @xcite found @xmath18 and @xmath19 and speculated that these best - fit values of the two parameters would be insensitive to the redshifts as well as to the mass scales . inspired by the existence of the universal peculiar velocity profile in the bound zone , @xcite suggested that the turn - around radius , @xmath2 , of a main group be readily obtained just as the value at which the following relation is satisfied : @xmath20 where the left - hand side ( lhs ) represents the global expansion speed at @xmath2 and the right - hand side ( rhs ) is the peculiar velocity at the same radius predicted by equation ( [ eqn : vpr ] ) .
provided that the virial radius ( or equivalently , virial mass ) of a massive group is already known as a prior and that the two parameters in the rhs , @xmath12 and @xmath13 , have truely constant values of @xmath21 and @xmath22 as speculated by @xcite , solving equation ( [ eqn : rt ] ) for @xmath2 should allow one to estimate the turn - around radius of a given group without necessitating an accurate measurement of the peculiar velocity field in its bound zone .
although @xcite showed that equation ( [ eqn : vpr ] ) with the constant values of @xmath12 and @xmath13 matched well the numerically obtained peculiar velocity profiles of three different cluster - size halos , it is inconclusive whether or not the bound - zone peculiar velocity profile is truly universal since the variations of its shape with mass and redshift were reported by other numerical works ( e.g. , * ? ? ?
* ; * ? ? ?
* ) . to account for the possibility that the bound - zone peculiar velocity profile may not be universal
, it is unavoidable to regard @xmath12 and @xmath13 in equation ( [ eqn : vpr ] ) as two _ varying parameters _ and to determine their best - fit values by fitting equation ( [ eqn : vpr ] ) to the bound - zone peculiar velocity profile constructed from observational data before estimating @xmath2 .
the practical feasibility of constructing the bound - zone peculiar velocity profiles was improved by @xcite who showed that if the average of the peculiar velocity field was taken over the spatial points distributed along the cosmic web ( filaments and sheets ) , @xmath23 can be expressed in terms of directly measurable quantities : @xmath24 here @xmath25 represents the difference in redshifts between the main group and a bound - zone object , @xmath26 is the projected distance of the bound - zone object from the group center in the plane of sky , and @xmath27 is the inclination angle between the direction of the cosmic web and the line - of - sight , which are all directly measurable without making any specific assumption about the background cosmology .
@xcite determined the values of @xmath12 and @xmath13 via adjusting the rhs to the lhs in equation ( [ eqn:2dvpr ] ) that can be also directly computed from the observables . applying this method to the bound zone of the ngc 5353/4 group around which a narrow filamentary structure was detected ( s.kim in private communication ) , @xcite pulled it off to estimate its turn - around radius .
the halo catalog from the small multidark planck simulation ( smdpl ) at @xmath28 is utilized to perform our numerical investigation . as a part of multidark and bolshoi project @xcite , smdpl
was conducted with @xmath29 particles of individual mass @xmath30 in a periodic box of a volume @xmath31 @xcite for a planck cosmology , i.e. , a flat @xmath0cdm cosmology with @xmath32 @xcite .
the smdpl halo catalog at @xmath28 provides information on the comoving position ( @xmath33 ) , peculiar velocity ( @xmath34 ) , virial radius and mass ( @xmath8 and @xmath6 ) of each dark matter halo identified by applying the rockstar halo finder to the phase space distributions of the particles @xcite .
the massive rockstar halos with virial mass @xmath35 are selected from the smdpl catalog as the _
main groups _ whose bound - zone peculiar velocity profile and turn - around radii are to be estimated in the current work . the lower - mass halos @xmath36 are excluded from the sample of the main groups since the number of the neighbor halos in their bound zones is not large enough to construct the peculiar velocity profiles .
a total of @xmath37 halos in the smdpl catalog at @xmath28 are found to be above the imposed mass cut - off and thus included in our sample of the main groups . throughout this analysis
we focus only on the present epoch ( @xmath28 ) to match the real observations since the prior information in the virial radius and mass of the main groups required for our analysis is available with high accuracy only for the case of the objects located in the nearby universe .
the region inside a spherical shell surrounding each main group with inner ( outer ) radius of @xmath38 ( @xmath39 ) from the group center is designated as its bound zone . to construct the peculiar velocity profile from the halos located in the bound zone around each main group , we consider only those well resolved halos composed of @xmath40 or more particles , which should help us avoid numerical flukes .
basically , we search for those halos with @xmath41 in the bound zone of each main group and then measure their separation distance vectors , @xmath42 , from the group center . the relative peculiar velocity of each bound - zone halo , @xmath11 is determined as @xmath43 where @xmath44 and @xmath45 are the peculiar velocities of a bound - zone halo and a main group , respectively . to construct the peculiar velocity profile from the bound zone halos distributed along all directions
, we first rescale the separation distance and the peculiar velocity of each bound - zone halo as @xmath46 and @xmath47 , respectively , where @xmath8 and @xmath15 are the virial radius and circular velocity of its main group , respectively . dividing the range of @xmath48 into ten short intervals , we calculate the mean value of the rescaled peculiar velocity at each @xmath48-interval .
taking the ensemble average of the mean peculiar velocity at each @xmath48-interval over all of the selected main groups , we numerically construct @xmath49 .
the errors associated with @xmath49 is also calculated as one standard deviation in the measurement of the ensemble average at each @xmath48-interval .
we finally fit the resulting numerical data points , @xmath50 , to equation ( [ eqn : vpr ] ) by adjusting the values of the two parameters , @xmath12 and @xmath13 , with the help of the maximum - likelihood method @xcite .
the best - fit values are determined to be @xmath51 and @xmath52 where the marginalized error in the determination of each parameter is obtained by treating the other parameter as a nuisance one . in figure
[ fig : vpr_z0 ] , the black filled circles represent the numerical result of @xmath53 with errors while the red solid line displays the analytic formula , equation ( [ eqn : vpr ] ) , with the best - fit values @xmath54 and @xmath55 . the blue dotted line is the analytic formula with the two parameters set at the original values of @xmath56 and @xmath57 given by @xcite .
figure [ fig : vpr_z0 ] clearly reveals that although the best - fit values of @xmath12 and @xmath13 are determined to be different from the original values given in @xcite , equation ( [ eqn : vpr ] ) with the best - fit parameters successfully describes the behavior of the bound - zone peculiar velocity profile , which is consistent with the result of @xcite .
now , we want to explore how the best - fit values of @xmath12 and @xmath13 change when the construction of the mean peculiar velocity profile in the bound zone is done along the cosmic web .
three dimensional spatial distributions of the bound - zone halos around a main group randomly selected from our sample is illustrated in figure [ fig : slice3d ] , which gives a glimpse of a specific direction along which the number density of the bound - zone halos is conspicuously high , corresponding to the direction of a bound - zone filament .
as expected , the bound - zone halos are not distributed isotropically . to construct the peculiar velocity profile of each main group along the bound - zone sheet , we first measure the cosines of the polar angles of the position vectors , @xmath42 , of the bound - zone objects from the center of each main group .
then , we divide the volume into ten slices through binning the values of @xmath58 and count the number of the bound - zone objects whose values of @xmath58 belong to each bin . locating the slice to which the largest number of the bound - zone objects belong and designating it as the bound - zone sheet ,
we include only those objects belonging to the bound - zone sheet to calculate the mean rescaled peculiar velocity profile for each main group .
taking the ensemble average of the profiles over all of the main groups , we finally construct the peculiar velocity profile from the bound - zone halos distributed along the bound - zone sheet , which is shown as the black filled circles in figure [ fig : vpr_sheet ] .
as can be seen , in the large @xmath48-section ( @xmath59 ) , the magnitude of @xmath50 does not monotonically decrease with @xmath48 unlike the case that the construction of @xmath50 is performed from the bound - zone halos distributed along all directions ( see figure [ fig : vpr_z0 ] ) .
rather , it has a minimum value at @xmath60 and becomes almost constant in the range of @xmath61 .
fitting equation ( [ eqn : vpr ] ) only to those numerical data points in the range of @xmath62 yields @xmath63 and @xmath64 .
the red solid line in figure [ fig : vpr_sheet ] corresponds to the analytic model with these best - fit parameters . in the restricted range of @xmath62 ,
the best - fit formula still works well , as can be seen . for the construction of @xmath53 along the bound - zone filament
, we also measure the azimuthal angles ( @xmath65 ) of the position vectors , @xmath42 , of the bound - zone objects around each main group .
the bound - zone volume of each main group is first divided into @xmath66 conic subvolumes through simultaneously binning the values of @xmath67 and @xmath65 . counting the number of the bound - zone halos belonging to each subvolume
, we locate the conic subvolume which contains the maximum number of the bound - zone halos and designate it as the bound - zone filament . using only those bound - zone halos belonging to the bound - zone filaments , we construct the average peculiar velocity profile , which is shown as black filled circles in figure [ fig : vpr_fil ] .
note that the magnitude of @xmath53 touches it minimum value at @xmath60 and changes its trend , increasing with @xmath48 in the range of @xmath68 .
equation ( [ eqn : vpr ] ) fitted only to those numerical data points in the range of @xmath62 is shown as red solid line in figure [ fig : vpr_fil ] , where the two parameters are set at their best - fit values found to be @xmath69 and @xmath70 .
the distinct behaviors of @xmath53 in the range of @xmath68 revealed by figures [ fig : vpr_sheet ] and [ fig : vpr_fil ] clearly demonstrate that equation ( [ eqn : vpr ] ) with constant values of @xmath12 and @xmath13 fails to describe the peculiar velocity profiles constructed along the bound - zone sheets and filaments .
our explanation for this result is as follows . in the bound - zone around a main group
, the halos would be affected not only by the gravity of the main group but also by that of the neighbor groups located outside the bound zone .
when the peculiar velocity profile is constructed by taking the average over all directions , the gravitational forces from the neighbor groups would be cancelled out and thus have no net effect on the behaviors of the bound - zone peculiar velocity profile .
however , when the construction of the peculiar velocity profile is made along the sheets and filaments , the gravitational forces from the neighbor halos no longer average out to zero but have non - zero net effect of flattening or increasing the magnitude of the peculiar velocities of the bound - zone halos at the distances larger than some threshold distance .
the comparison between the two results in figures [ fig : vpr_sheet ] and [ fig : vpr_fil ] indicates that when the bound - zone peculiar velocity profile is constructed along one dimensional filament , the profile in the range of @xmath68 becomes more severely contaminated by the gravitational influences of the neighbor groups outside the bound zone .
this result is in line with the recent finding of @xcite that the gravitational forces of the surrounding large - scale structures quantified by the two halo terms in the density correlation function @xcite have the strongest effect on the peculiar velocities of the galaxies that fall into the clusters along the filaments . to see the trend of @xmath53 with the variation of the range of @xmath6
, we also divide the values of @xmath6 into three intervals and reconstruct @xmath53 by using only those main groups whose masses belong to each @xmath6-interval .
figure [ fig : vpr_mweb ] shows how differently the peculiar velocity profiles constructed along all directions and along the cosmic web behave for three cases of the @xmath6 bins .
the best - fit values of @xmath12 and @xmath13 as well as the numbers of the main groups for each case are listed in table [ tab : all ] . as can be seen , the best - fit amplitude and slope decrease as the mass scales of the main groups decrease ( see ,
also * ? ? ?
for the case of the most massive groups with @xmath71 , the amplitude and slope of the bound - zone peculiar velocity profile constructed along all directions have the best - fit values of @xmath72 and @xmath73 , which agree well with the original values determined by @xcite . whereas , the other two cases of the lower - mass groups yield substantially lower slopes and amplitudes than the original values , which implies that the lower masses are less central since the energy consideration requires the slope to be close to @xmath74 for the central masses .
figure [ fig : vpr_mweb ] also displays in the middle and left panels how the shapes of the peculiar velocity profiles constructed along the cosmic web differ between the @xmath6-bins and table [ tab : all ] lists the best - fit values of @xmath12 and @xmath13 for each case that the construction of @xmath53 are made along the cosmic web .
note that the less massive a main group is , the more rapidly the magnitude of its bound - zone peculiar velocity profile constructed along filaments increases with distance in the section of @xmath59 .
the same logic provided in the above can also explain this result : the gravitational forces of the neighbor groups outside the bound zone should have stronger influences on the bound zones around the lower mass groups .
it is worth noting that the threshold value of the rescaled distances , say @xmath75 , at which the monotonic decreases of the magnitudes of the peculiar velocity profiles constructed along the cosmic web come to a halt seem to be constant @xmath76 .
the threshold value of @xmath75 represents the maximum rescaled distance up to which the gravitational effect of the neighbor groups outside the bound zone on @xmath53 is negligible .
we make a serendipitous discovery that the threshold rescaled distance , @xmath77 , coincides with the value of the following ratio : @xmath78^{1/3 } \approx 5.43 \end{aligned}\ ] ] here , we employ the relation of @xmath79 where @xmath80 is the critical density at the present epoch and @xmath81 is the ratio of the mass density inside @xmath8 to @xmath82 @xcite .
the dotted blue lines shown in figures [ fig : vpr_sheet]-[fig : vpr_mweb ] correspond to this ratio , @xmath83 .
it may be worth discussing here why we did not follow the conventional scheme to identify the sheets and filaments in the bound zones around the main groups .
the lagrangian perturbation theory describes that the sheets ( filaments ) form when the gravitational collapse proceeds anisotropically along the major ( major and intermediate ) principal axes of the deformation tensors calculated as the second derivaties of the perturvation potentials @xcite .
accordingly , the sheets ( filaments ) have been conventionally identified as the regions where the local tidal shear tensors have one ( two ) positive and two ( one ) negative eigenvalues @xcite . if we identified the sheets and filaments by using the conventional scheme , however , equation ( [ eqn:2dvpr ] ) would not be applicable since the success of equation ( [ eqn:2dvpr ] ) requires all of the bound - zone objects belonging to a sheet ( or a filament ) to possess the same inclination angles or at least to have the angles constrained in the small interval @xmath84 $ ] . in other words ,
equation ( [ eqn:2dvpr ] ) is applicable only for the case that a bound - zone sheet ( filament ) possesses a flat plane - like ( straight line - like ) shape .
our scheme is deliberately devised to ensure that the identified sheets and filament have the required shapes .
in this section , we are going to estimate the turn - around radii of the main groups with the help of the methodology of @xcite based on equation ( [ eqn : rt ] ) .
but , before embarking on the estimates of the turn - around radii , let us appreciate the implication of the results obtained in section [ sec : web ] that the bound - zone peculiar velocity profiles constructed along the cosmic web can be contaminated by the gravitational influence of its neighbor groups with comparable masses located outside the bound zone .
the bottom line of section [ sec : web ] is that equations ( [ eqn : vpr ] ) and ( [ eqn:2dvpr ] ) fail to describe the bound - zone peculiar velocity profiles constructed along the cosmic web . given that the validity of equations ( [ eqn : vpr ] ) and ( [ eqn:2dvpr ] ) is a key to the practical success of the methodology of @xcite , it should be necessary to sort out those main groups whose peculiar velocity profiles are not severely contaminated by the gravitational field outside the bound zone .
we suspect that for the case of _ isolated _ groups which have no neighbor groups with comparable masses in the regions near to the bound zone , the contamination of the peculiar velocity profiles due to the gravitational field outside the bound - zone may be attenuated to the negligible level . for each selected main group with @xmath1 , we search for its nearest neighbor group and measured the distance , @xmath85 , between them .
then , those groups which satisfy @xmath86 are identified as the isolated main groups .
following the same procedures described in section [ sec : web ] , we construct the bound - zone peculiar velocity profiles averaged over a total of @xmath87 isolated main groups .
figure [ fig : vpr_iso ] shows the same as figure [ fig : vpr_z0 ] but for the case of the isolated main groups .
the best - fit parameters for this case are found to be @xmath88 and @xmath89 . as can be seen , the magnitude of the bound - zone peculiar velocity profile averaged over the isolated main groups decreases much more rapidly with @xmath48 than that averaged over all main groups shown in figure [ fig : vpr_z0 ] .
figures [ fig : vpr_sheetiso ] and [ fig : vpr_filiso ] show the same as figures [ fig : vpr_sheet ] and [ fig : vpr_fil ] but for the cases of the isolated main groups , respectively .
the best - fit values are found to be @xmath90 and @xmath91 ( @xmath92 and @xmath93 ) for the case of @xmath50 constructed along the bound - zone sheets ( filaments ) around the isolated main groups .
it is interesting to note that for the case of the isolated main groups , the magnitudes of @xmath53 constructed along the cosmic web decrease monotonically in the whole bound zone range @xmath10 just as @xmath53 constructed via isotropic average .
this result indicates that the contamination caused by the gravitational influences of the neighbor groups on the bound - zone peculiar velocity profiles vanishes for the cases of the isolated main groups , as we have anticipated .
we have tested the robustness of this result against the variation of @xmath94 and that only when @xmath95 , the profile @xmath96 shows monotonic decrease with @xmath48 in the whole range of @xmath48 .
figure [ fig : vpr_mwebiso ] and table [ tab : iso ] show the same as figure [ fig : vpr_mweb ] and table [ tab : all ] but for the case of the isolated main groups , respectively .
it is important to note that for the cases of the isolated main groups the bound - zone peculiar velocity profiles do not show strong dependence on the mass scale @xmath6 , although the best - fit values of @xmath12 and @xmath13 suffer from large errors due to the small - number statistics ( especially in the large-@xmath6 bin ) .
the results shown in figures [ fig : vpr_iso]-[fig : vpr_mwebiso ] indicate that the analytic formula , equation ( [ eqn : vpr ] ) , with fixed values of @xmath12 and @xmath13 can reliably describe the peculiar velocity profile in the whole bound zone range only when the condition of isolation is imposed on the selection of the main groups .
now , we are ready to estimate the turn - around radii of the isolated main groups by following the algorithm of @xcite .
first , we construct the peculiar velocity profile of each individual isolated main group and determine the best - fit values of @xmath12 and @xmath13 by fitting equation ( [ eqn : vpr ] ) to each individually constructed peculiar velocity profile . putting the best - values of @xmath12 and @xmath13 into equation ( [ eqn : rt ] )
, we estimate the turn radius @xmath2 of each individual isolated main group .
then , we divide the range of @xmath6 into @xmath97 bins and take the ensemble average of @xmath2 over the isolated main groups whose virial masses belong to each @xmath6-bin .
the one standard deviation scatter around the average is also calculated at each @xmath6-bin as @xmath98 where @xmath99 is the number of the isolated main groups belonging to each @xmath6-bin and @xmath100 represents the difference of the turn - around radius of the @xmath101-th main group from the ensemble average .
the errors associated with the calculation of the ensemble average is also computed as @xmath102 .
the turn - around radii @xmath2 , estimated by the algorithm of @xcite is plotted versus the virial masses @xmath6 of the isolated main groups as the black filled circles with errors in figure [ fig : ta_iso ] .
the red solid line is the spherical bound limit , @xmath3 , given in equation ( [ eqn : rt_u ] ) predicted by the planck cosmology .
the blue dotted line is the one standard deviation scatter @xmath103 around the average values of @xmath2 .
this result clearly reveals that the spherical bound limit ( red solid line ) is much higher than the estimated turn - around radii in the whole mass range , implying that a bound- violation ( i.e. , an event of @xmath104 ) seldom occurs in the planck universe .
note that a bound violation is relatively less rare on the low mass scale ( @xmath105 ) , which can be attributed to the fact that the spherical symmetry of the gravitational collapse process breaks down more severely on the low mass scales @xcite .
the green solid line in figure [ fig : ta_iso ] represents the theoretical turn - around radius as a function of @xmath6 evaluated by solving equation ( [ eqn : rt ] ) whose two parameters are set at the best - fit values of @xmath106 and @xmath107 determined from the average ( not individual ) peculiar velocity profile ( see figure [ fig : ta_iso ] ) . while the black filled circles with errors have been obtained by applying the method of @xcite to the bound - zone peculiar velocity profiles of the individual isolated groups , the green solid line has been theoretically evaluated by applying the method of @xcite to the mean bound - zone peculiar velocity profile averaged over all of the isolated groups .
the good agreement between the black filled circles and the green solid line indicates implies the robustness and solidity of the algorithm of @xcite for the estimation of the turn - around radii .
now , we would like to estimate @xmath2 of the isolated main groups along the cosmic web . basically , we repeat the same procedure for the estimation of @xmath108 as described in the above but by using the individual peculiar velocity profiles @xmath53 constructed along the sheets and the filaments .
figures [ fig : ta_sheetiso ] and [ fig : ta_filiso ] show the same as figure [ fig : ta_iso ] but for the case that the turn - around radii of the individual isolated main groups are estimated along the sheets and the filaments , respectively . to plot the green solid line in figure [ fig : ta_sheetiso ] ( figure [ fig : ta_filiso ] )
the method of @xcite has been applied to the mean peculiar velocity profile shown in figure [ fig : vpr_sheetiso ] ( figure [ fig : vpr_filiso ] ) . as can be seen , the ensemble averages of the turn - around radii estimated along the cosmic web are still lower than the spherical upper limit in the whole mass range .
the comparison of figures [ fig : ta_sheetiso ] and [ fig : ta_filiso ] with figure [ fig : ta_iso ] reveals that the values of @xmath2 estimated along the cosmic web vary in a wider range and that the spherical bound limit ( red solid line ) is placed within one standard deviation scatter from the ensemble average . in other words , if the turn - around radii of the individual isolated main groups are estimated along the cosmic web , it will be less rare to witness the bound violations even in a @xmath0cdm universe .
hence , an observation of a single bound violation of an individual object would not challenge the standard model . only if the spherical bound limit is found to be violated by the _
average _ value of the turn - around radius rather than by the individual ones , it should be regarded as an anomaly .
we have numerically estimated the turn - around radii @xmath2 of the massive groups with virial mass @xmath109 identified in the smdpl @xcite by constructing the peculiar velocity profiles of the bound - zone halos with @xmath110 around the groups . the shapes and behaviors of the peculiar velocity profiles
have been found to vary with the directions along which the constructions were made .
when the bound - zone halos distributed along all directions are used for the constructions , the magnitudes of the resulting peculiar velocity profiles monotonically decrease with distance , being well approximated by the simple formula of @xcite with two characteristic parameters .
however , when only those bound - zone halos distributed along some confined directions around the groups are used for the constructions , the resulting peculiar velocity profiles exhibit distinct features : instead of monotonically decreasing with distance , the magnitudes of the profiles reach the minimum values at some critical distances @xmath111 and remain constant ( show increment ) with distance after @xmath111 when the constructions of the profiles are made from those bound - zone halos distributed along the sheets ( filaments ) . explaining that this distinctly unique behaviors of the bound - zone peculiar velocity profiles constructed along the sheets and the filaments should originate from the external gravitational effect of the neighbor massive groups , we have imposed the condition of isolation on the selection of the main groups to avoid any contamination in the estimates of their turn - around radii due to the external gravity .
as anticipated , it has been shown that the bound - zone peculiar velocity profiles around the isolated main groups exhibit monotonic decrement with distance , even when the profiles are constructed along the sheets and the filaments .
finally , the average turn - around radii of the isolated main groups have been estimated from the bound - zone peculiar velocity profiles with the help of the algorithm developed by @xcite and found to be well below the spherical bound limit predicted by the @xmath0cdm cosmology in the whole mass range .
we have also quantitatively shown that it is quite rare but not forbidden for individual groups to violate the bound limit especially on the relatively low - mass scale @xmath112 and that the rarity of the bound - violations diminishes when the turn - around radii are estimated along the cosmic web .
given our result , it has been suspected that the bound - violation of ngc 5353/4 group reported by @xcite may not be interpreted as a counter - evidence against the @xmath0cdm model at the moment .
it is worth discussing our serendipitous discovery of the coincidence between @xmath111 and @xmath3 for the case of the non - isolated main groups .
when the bound - zone peculiar velocity profiles are constructed along the filaments around the _ non - isolated _ main groups , the threshold distances , @xmath111 , at which the magnitudes of the bound - zone peculiar velocities touch their minimum values coincide with the spherical upper limit , @xmath113 , at all mass scales . given that the spherical upper limit , @xmath113 is a direct indicator of the amount of @xmath0 as shown in equation ( [ eqn : rt_u ] ) , our result hints at the possibility of using @xmath114 as another local test of the @xmath0cdm cosmology . constructing the average bound - zone peculiar velocity profile along the filaments around the non - isolated massive groups from the direct observables with the help of equation ( [ eqn:2dvpr ] ) and locating where the magnitude of the average bound - zone peculiar velocity profile reaches the minimum value
, we can determine @xmath114 .
then , equating it to the theoretical constant given in equation ( [ eqn : rsc1 ] ) , we can estimate the amount of @xmath0 .
it would also allow us to constrain the equation of state of dark energy since the value of the spherical upper limit changes if the equation of state of dark energy deviates from @xmath115 @xcite . to make a robust estimate of @xmath111 from the average bound - zone peculiar velocity profile from observations , it will be definitely necessary to use a large sample of the massive groups observed in the local universe , which is the direction of our future work .
we are very grateful to our anonymous referee for very useful comments which helped us improve the original manuscript .
this work was initiated during the workshop on `` near field cosmology 2016 '' held at the obergurgl university centre in tyrol , austria from march 29 - april 3 , 2016 .
we thank stefan gottlber and the other organizers of the workshop for making stimulating discussions among the participants possible during the workshop .
this study made use of data from the cosmosim database ( http://www.cosmosim.org/ ) , hosted and maintained by the leibniz - institute for astrophysics potsdam ( aip ) .
the multidark - planck ( mdpl ) and the bigmd simulation suite have been performed on the supermuc supercomputer at lrz using time granted by prace . j.l .
acknowledges the support of the basic science research program through the nrf of korea funded by the ministry of education ( no . 2013004372 ) .
was also partially supported by a research grant from the national research foundation ( nrf ) of korea to the center for galaxy evolution research ( no .
2010 - 0027910 ) .
g.y . thanks mineco / feder for financial support under research grants aya2015 - 63810-p and aya2012 - 31101 .
behroozi , p. s. , wechsler , r. h. , & wu , h .- y . 2013 , , 762 , 109 behroozi , p. s. , wechsler , r. h. , wu , h .- y . , et al .
2013 , , 763 , 18 bernardeau , f. 1994 , , 427 , 51 bond , j. r. , & myers , s. t. 1996 , , 103 , 1 bond , j. r. , kofman , l. , & pogosyan , d. 1996 , , 380 , 603 boylan - kolchin , m. , springel , v. , white , s. d. m. , jenkins , a. , & lemson , g. 2009 , , 398 , 1150 cuesta , a. j. , prada , f. , klypin , a. , & moles , m. 2008 , , 389 , 385 falco , m. , hansen , s. h. , wojtak , r. , et al .
2014 , , 442 , 1887 faraoni , v. 2016 , physics of the dark universe , 11 , 11 hahn , o. , porciani , c. , carollo , c. m. , & dekel , a. 2007 , , 375 , 489 hayashi , e. , & white , s. d. m. 2008 , , 388 , 2 karachentsev , i. d. , tully , r. b. , wu , p .- f . ,
shaya , e. j. , & dolphin , a. e. 2014 , , 782 , 4 klypin , a. , yepes , g. , gottlber , s. , prada , f. , & he , s. 2016 , , 457 , 4340 lee , j. , kim , s. , & rey , s .- c .
2015 , , 815 , 43 lee , j. 2016 , arxiv:1603.06672 linder , e. 2003 , , 90 , 091301 linder , e. v. 2005 , , 72 , 043529 ma , c .- p . , & fry , j. n. 2000 , , 543 , 503 pavlidou , v. , tetradis , n. , & tomaras , t. n. 2014 , jcap , 5 , 017 pavlidou , v. , & tomaras , t. n. 2014 , jcap , 9 , 020 planck collaboration , ade , p. a. r. , aghanim , n. , et al .
2014 , , 571 , a16 svensmark , j. , wojtak , r. , & hansen , s. h. 2015 , , 448 , 1644 tully , r. b. , & trentham , n. 2008 , , 135 , 1488 tully , r. b. 2015 , , 149 , 54 wall , j. v. , & jenkins , c. r. 2012 , practical statistics for astronomers , ( cambridge , uk : cambridge university press ) zeldovich , y. b. 1970 , , 5 , 84 at @xmath28 . the black filled circles with errors represent the numerical results from the multi - dark simulations and the red solid line represents the analytic formula with the best - fit parameters .
the blue dotted lines corresponds to the original formula derived by @xcite . ]
ccccc @xmath116 & @xmath117 & all & @xmath118 & @xmath119 + @xmath116 & @xmath117 & sheet & @xmath120 & @xmath121 + @xmath116 & @xmath117 & filament & @xmath122 & @xmath123 + @xmath124 & @xmath125 & all & @xmath126 & @xmath127 + @xmath124 & @xmath125 & sheet & @xmath128 & @xmath129 + @xmath124 & @xmath125 & filament & @xmath130 & @xmath131 + @xmath132 & @xmath133 & all & @xmath134 & @xmath135 + @xmath132 & @xmath133 & sheet & @xmath136 & @xmath137 + @xmath132 & @xmath133 & filament & @xmath138 & @xmath139 + [ tab : all ] ccccc @xmath116 & @xmath140 & all & @xmath141 & @xmath142 + @xmath116 & @xmath140 & sheet & @xmath143 & @xmath144 + @xmath116 & @xmath140 & filament & @xmath145 & @xmath146 + @xmath124 & @xmath147 & all & @xmath148 & @xmath149 + @xmath124 & @xmath147 & sheet & @xmath150 & @xmath151 + @xmath124 & @xmath147 & filament & @xmath152 & @xmath153 + @xmath132 & @xmath154 & all & @xmath155 & @xmath156 + @xmath132 & @xmath154 & sheet & @xmath157 & @xmath158 + @xmath132 & @xmath154 & filament & @xmath159 & @xmath160 + [ tab : iso ] | a bound - violation designates a case that the turn - around radius of a bound object exceeds the upper limit put by the spherical collapse model based on the standard @xmath0cdm paradigm .
given that the turn - around radius of a bound object is a stochastic quantity and that the spherical model overly simplifies the true gravitational collapse which actually proceeds anisotropically along the cosmic web , the rarity of the occurrence of a bound violation may depend on the web environment .
assuming a planck cosmology , we numerically construct the bound - zone peculiar velocity profiles along the cosmic web ( filaments and sheets ) around the isolated groups with virial mass @xmath1 identified in the small multidark planck simulations and determine the radial distances at which their peculiar velocities equal the hubble expansion speed as the turn - around radii of the groups .
it is found that although the average turn - around radii of the isolated groups are well below the spherical bound - limit on all mass scales , the bound violations are not forbidden for individual groups and that the cosmic web has an effect of reducing the rarity of the occurrence of a bound violation .
explaining that the spherical bound limit on the turn - around radius in fact represents the threshold distance up to which the intervention of the external gravitational field in the bound - zone peculiar velocity profiles around the _ non - isolated _ groups stays negligible , we discuss the possibility of using the threshold distance scale to constrain locally the equation of state of dark energy . |
this is the first in a series of two papers in which we present the results of a study with the hubble space telescope ( hst ) of the line - of - sight velocities of stars in the central few arcsec of the globular cluster m15 ( ngc 7078 ) .
the present paper discusses the spectroscopic observations with the hst space telescope imaging spectrograph ( stis ) and the extraction and calibration of the stellar spectra .
it also discusses the construction of a photometric catalog from hst imaging with the second wide field and planetary camera ( wfpc2 ) , which proved essential in the reduction and interpretation of the data . the second paper ( gerssen et al .
2002 , hereafter paper ii ) discusses the determination of line - of - sight velocities from the spectra , and the implications for the dynamics and structure of m15 .
the main motivation for our study is to better constrain the possible presence of an intermediate - mass black hole in the center of m15 .
such a black hole was hinted at by previous work ( e.g. , peterson , seitzer & cudworth 1989 ; gebhardt et al .
2000a ) , but the limited quality of the data precluded very strong conclusions ( see the review in van der marel 2001 ) .
the globular cluster m15 at a distance of 10 kpc has one of the highest central densities of any globular cluster in our galaxy .
the high density has likely affected the stellar population , as evidenced by the presence of two bright x - ray sources ( white & angelini 2001 ) and several millisecond pulsars ( phinney 1993 ) .
a variety of interesting dynamical and structural evolutionary phenomena have been predicted to occur at high stellar densities , including mass segregation and the possible formation of a central black hole ( hut et al .
1992 ; meylan & heggie 1997 ; see the additional discussion in the introductory section of paper ii ) . as a result
, m15 has been one of the globular clusters for which the structure and dynamics have been most intensively studied in the past decade .
studies of the dynamics of m15 have focused primarily on the determination of line - of - sight velocities of individual stars , using either spectroscopy with single apertures , long - slits or fibers ( peterson et al . 1989 ; dubath & meylan 1994 ; dull et al . 1997 ; drukier et al . 1998 ) , or using imaging fabry - perot spectrophotometry ( gebhardt et al . 1994 , 1997 , 2000a ) .
integrated light measurements using single apertures have also been attempted ( peterson et al . 1989 ; dubath , meylan & mayor 1994 ) , but are only of limited use ; integrated light spectra are dominated by the light from only a few bright giants , and as a result , inferred velocity dispersions are dominated by shot noise ( zaggia et al .
1992 , 1993 ; dubath et al .
1994 ) .
line - of - sight velocities are now known for @xmath5 m15 stars , as compiled and analyzed by gebhardt et al .
( 2000a ) .
the projected velocity dispersion profile increases inwards from @xmath6 at @xmath7 arcmin , to @xmath8 @xmath9 .
the velocity dispersion is approximately constant at smaller radii , and is @xmath10 at the innermost available radius @xmath11 .
the rotational properties of m15 are quite peculiar .
the position angle of the projected rotation axis in the central region is @xmath12 different from that at larger radii , and the rotation amplitude increases to @xmath13 for @xmath14 , so that @xmath15 in this region .
a plausible explanation for these properties is still lacking . to improve our understanding of the central mass distribution in m15 , it is of crucial importance to obtain more stellar velocities close to the center .
however , velocity determinations at @xmath16 are very difficult due to severe crowding and the presence of a few bright giants in the central arcsec .
gebhardt et al .
( 2000a ) used an imaging fabry - perot spectrophotometer with adaptive optics on the cfht , and obtained fwhm values as small as @xmath17 .
the strehl ratio was only @xmath18% , so that even in these observations the light from the fainter turnoff and main - sequence stars in the central arcsec was overwhelmed by the psf wings of the nearby giants . as a result
, there are only 5 stars with known velocities within @xmath19 from the cluster center .
to improve upon this situation we initiated the project on which we report here , in which we used hst / stis to map the center of m15 spectroscopically at high spatial resolution . in the future , additional kinematic data on m15 may come available from stellar proper motion measurements with hst , using techniques such as those described by anderson & king ( 2000 ) .
however , at present no such measurements are available .
this paper is organized as follows .
section [ s : imaging ] describes the construction of a stellar catalog from wfpc2 data .
this catalog is essential for a proper interpretation of the m15 spectra .
it is distributed electronically as part of the present paper .
section [ s : calibspec ] describes stis spectra of a bright field star which were obtained to calibrate the stis point spread function ( psf ) and line - spread function ( lsf ) .
section [ s : specm15 ] describes the m15 stis spectroscopy , the observational setup and data reduction , and an analysis of the achieved slit positioning accuracies .
the velocity calibration accuracy of the spectra is the most crucial aspect of the observations , and this issue is discussed in detail .
section [ s : indstars ] discusses the algorithm adopted for the extraction of individual stellar spectra .
these spectra form the end - product of the present paper .
they are used in paper ii to determine line - of - sight velocities and to study the structure and dynamics of m15 .
section [ s : conc ] presents concluding remarks .
for proper planning and interpretation of long - slit spectra of m15 it is necessary to know the positions and magnitudes of the stars in m15 . because m15 is dense , and its central regions are crowded , it is important to determine these quantities from high spatial resolution data .
several imaging studies of m15 were undertaken with the hst in the years 1990 - 1994 ( lauer et al .
1991 ; ferraro & paresce 1993 ; stetson 1994 ; de marchi & paresce 1994 ; yanny et al .
1994 ) , but these suffered from the aberrated hst psf .
imaging studies with the faint object camera ( foc ) on the refurbished hst ( with costar ) were reported by de marchi & paresce ( 1996 ) and sosin & king ( 1997 ) . however , while foc provided a sharp and well - sampled psf , it had only a very small field of view and its data were hard to calibrate accurately .
data from wfpc2 are therefore preferable ; see biretta et al .
( 2001 ) for a detailed description of this instrument .
wfpc2 images of m15 were obtained on april 7 , 1994 in the context of hst program go-5324 ( pi : yanny ) .
results from these observations were reported in a study led by one of us ( guhathakurta et al .
1996 , hereafter g96 ) .
data were obtained using the filters f336w , f439w and f555w , respectively , which correspond roughly to the johnson u , b and v bands .
the g96 paper provides a detailed discussion of the data acquisition and reduction .
their figure 1 shows a large scale black - and - white image of m15 , while their figure 2 shows a ` true - color ' image of the central @xmath20 .
the photometric analysis of g96 yielded a list of stars , and for each star its position on one of the wfpc2 chips and the countrate in each of the filters .
this list is the starting point of the present analysis . while g96 applied both an astrometric calibration ( the transformation from chip position to right ascension ra and declination dec ) and a photometric calibration ( the transformation from count - rate to johnson u , b and v magnitudes ) to their star list , these were not particularly accurate .
the wfpc2 had only just been installed on hst when the g96 data were obtained , and there was little information available on the appropriate astrometric and photometric transformation equations when their paper was written .
this was not a problem for the g96 study , which focused on a determination of the stellar number density profile of m15 .
however , for the present application both astrometric and photometric accuracy are important , and we therefore decided to apply an improved calibration to the g96 star list .
for the astrometric calibration we first ran the iraf / stsdas task uchcoord , as recommended by the hst data handbook ( voit et al . 1997 ) .
this properly sets the image header parameters for the wfpc2 plate scales and chip - to - chip rotations .
the task metric was then used to transform the pixel coordinates of each star to ra and dec in the guide star coordinate system ( gscs ) . for the position of the m15 cluster center we adopted the pixel position calculated by g96 ( which has an estimated @xmath21 uncertainty of @xmath22 in each coordinate ) . for each star the relative position with respect to the cluster center is given by @xmath23 and @xmath24 . in the following it is more convenient to work in a right - handed cartesian coordinate system , so we define @xmath25 and @xmath26 .
the 1@xmath27 absolute accuracy of the gscs is only @xmath28 .
however , the absolute astrometric accuracy can be improved using the star ac211 .
this x - ray source has a radio counterpart for which the position is determined to within @xmath0 accuracy .
kulkarni et al .
( 1990 ) find for ac211 that @xmath29 and @xmath30 .
we find for ac211 that @xmath31 and @xmath32 .
this implies that the m15 cluster center is at @xmath33 and @xmath34 in the radio coordinate system .
this differs by @xmath35 from the position that was calculated in gscs , consistent with the absolute accuracy of the latter . for the photometric calibration we started with the count rates ( i.e. , the integrated counts divided by the exposure time ) determined by g96 .
these had been corrected to an equivalent aperture with a radius of @xmath36 , although the actual photometry was performed with a smaller aperture .
these count rates were first corrected for geometric distortion and then for the wfpc2 charge transfer efficiency ( cte ) problem , as described in the hst data handbook ( voit et al .
for the cte correction we used the formulae given by whitmore & heyer ( 1997 ) .
however , we increased their recommended corrections by a factor 3 , to take into account that the cte was more of a problem before april 24 , 1994 ( i.e. , when the m15 data were taken ) . on this date
the wfpc2 temperature was lowered from @xmath37c to @xmath38c .
the factor of 3 was chosen on the basis of quantitative comparisons of the cte effect at the two different temperatures ( holtzman et al .
we then transformed the count rates to johnson u , b and v magnitudes using the procedures described by holtzman et al .
( 1995b ) .
several iterations were performed to get the color terms correct in the transformation equations .
the filter- and chip - dependent change in throughput that occurred on april 24 , 1994 was taken into account , using the values given in the wfpc2 data handbook ( voit et al .
1997 ) .
the above steps provide an optimal relative photometric calibration for the data
. however , the absolute calibration remains uncertain due to uncertainties in aperture corrections and contamination - related throughput corrections .
this is the case for wfpc2 data in general , but is particularly important for data taken before april 24 , 1994 .
wfpc2 was operated at @xmath37c for only a few months , and the instrument calibrations available for this period are considerably less accurate than those available for subsequent data .
we therefore applied a constant additive correction to all the stellar magnitudes for a given filter+ccd combination .
the corrections were chosen to align the red giant branch ( rgb ) and horizontal branch ( hb ) features in color - magnitude diagrams ( cmds ) with the corresponding features in well - calibrated ground - based data ( durrell & harris 1993 ; stetson 1994 ) .
the corrections ranged from @xmath39@xmath40 mag . with these corrections
we estimate the calibration of the final catalog to be accurate to @xmath41 mag in @xmath42 and @xmath43 and to @xmath44 mag in @xmath45 .
this is sufficient for the purposes of the present project .
the final stellar catalog contains 31,983 stars with their positions and u , b and v magnitudes .
it is distributed electronically as part of this paper .
for reference , the first 25 entries are listed in table [ t : catalog ] .
each star has an i d number .
these numbers were already used for cross - identification purposes by gebhardt et al .
( 2000a ) in their compilation of m15 stars with ground - based line - of - sight velocity determinations . the spatial distribution of the stars in our wfpc2 catalog is shown in figure [ f : catimage ] .
the footprint of the wfpc2 field of view is clearly visible . towards the east the catalog extends as far as @xmath46 arcmin from the cluster center . for comparison ,
the half - mass radius of m15 is @xmath47 arcmin and the tidal radius is @xmath48 arcmin ( harris 1996 ) , so the catalog only covers the central part of m15 .
note from figure [ f : catimage ] that stars are not detected along the ccd chip boundaries .
stars within @xmath49 from the cluster center are all detected on the pc chip , and this inner region of m15 is therefore spatially complete in the catalog .
figure [ f : cmd ] shows a cmd of @xmath43 vs. @xmath50 for the stars in the catalog .
the well - known stellar - evolutionary features of the main sequence , subgiant branch , rgb , hb , and blue stragglers are all easily discernible .
the increased width of the hb between @xmath51 ( the instability strip ) is due to rr lyrae variables .
the implications of the morphology and position of the observed cmd features for the age , metallicity and stellar population of m15 have all been addressed previously ( e.g. , durrell & harris 1993 ; ferraro & paresce 1993 ; stetson 1994 ; de marchi & paresce 1994 ; yanny et al .
1994 ; de marchi & paresce 1996 ; sosin & king 1997 ) and will not be discussed here .
the large color spread of the main sequence at faint magnitudes is due to photometric errors .
the photometric errors and completeness of the catalog , and their dependence on distance from the cluster center , were discussed in detail by g96 .
the catalog is essentially complete for magnitudes brighter than @xmath52 , and the completeness drops steadily towards fainter magnitudes .
no stars are detected fainter than @xmath53 .
we were awarded 25 orbits in the context of hst program go-8262 ( pi : van der marel ) for a spectroscopic study of the central region of m15 .
all observations were taken with stis ( see leitherer et al .
2001 for a detailed description of this instrument ) . for all observations we used the 52x0.1 slit , which has a calibrated width of @xmath54 ( bohlin & hartig 1998 ) .
as dispersive element we used the g430 m grating centered at a wavelength of 5216 .
this yields spectra that cover the region from 50735359 .
this wavelength range has often been used for kinematical studies and it includes the mg b triplet region around 5175 .
the pixel size in the spatial direction is @xmath55 ( bowers & baum 1998 ) .
the pixel size in the spectral direction is 0.276 .
the latter corresponds to @xmath56 at the center of the wavelength range . for comparison , the central velocity dispersion of m15
has previously been measured to be @xmath57 ( gebhardt et al .
2000a ) , so for a proper understanding of the m15 kinematics the doppler shifts must be measured to a fraction of a pixel accuracy .
this requires accurate calibration , and a detailed understanding of the stis point spread function ( psf ) and line - spread function ( lsf ) . for this reason we used one orbit on june 14 , 1999 , for a set of calibration observations of a bright field star . for the calibration observations we chose to observe the star hd 122563 , an extremely metal - poor field giant with magnitude @xmath58 . at the start of the orbit
an acq acquisition procedure was executed to put the star near the center of the slit .
this procedure has a nominal accuracy between @xmath59 ( leitherer et al .
2001 ) and @xmath60 ( voit et al .
it was not necessary to achieve better pointing accuracy than this , so we did not perform an additional ( more accurate ) acq / peak procedure .
a series of small telescope slews were then used to successively position the star at 14 different positions relative to the slit , as shown schematically in figure [ f : hdslit ] . at each position
two spectra were obtained .
these were dithered by an integer number of @xmath61 pixels in the direction parallel to the slit , to allow removal of ccd defects during data reduction .
the exposure times ranged from 2 to 20 seconds per exposure .
the longer exposure times were used when the star was offset perpendicularly from the center of the slit , so as to obtain an approximately similar signal - to - noise ratio ( @xmath3 ) for all exposures .
we used 2-pixel on - chip binning in the spatial direction to obtain an effective pixel size of @xmath62 . the stis mode selection mechanism ( which selects and positions the grating ) has a small mechanical non - repeatability .
this causes an uncertainty ( amounting to a few pixels ) in the relative position of the 2d spectrum with respect to the ccd rows and columns .
in addition , slight drifts in this position ( of the order of @xmath39 pixel ) can occur over the course of an orbit .
for accurate calibration it is therefore important that contemporaneous spectra of an arc lamp are obtained . for the calibration star observations we obtained five exposures of an arc lamp , spread equally in time between the beginning and the end of the orbit .
our data reduction procedures are based on the default stis reduction pipeline . however , we applied a number of modifications and additional calibrations to obtain more accurate results .
we started with the raw data files , and used the iraf / stsdas task getref to ensure that the relevant header keywords point to the most appropriate calibration files .
subsequently , the basic2d task was run to perform the tasks of data quality initialization , bias subtraction , dark current subtraction , and flat - fielding .
for the dark current subtraction we used the ` superdark ' frames that are generated by stsci on a weekly basis .
we experimented with the use of daily darks , but found that they did not improve the results due to their lower @xmath3 .
we fitted and subtracted a spatially linear background from all frames to ensure that source - free regions along the slit were consistent with zero .
the two exposures obtained at each stellar position were then aligned and co - added .
the latter was done with the task ocrrej , which rejects cosmic rays and ccd defects .
bad columns were corrected using linear interpolation .
after these steps a significant number of bad pixels remain in the data .
positive outliers are generally pixels that are ` hot ' in the data frame but normal in the ( non - contemporaneous ) dark frame , while for negative outliers the opposite is generally the case .
all outliers were interpolated over using the task cosmicrays ( to find the negative outliers the science frame was multiplied by minus one ) .
the arc lamp spectra were reduced similarly as the object spectra .
the wavecal task was then used to determine the shift of the spectra with respect to a fiducial spectrum maintained in the hst data archive . these mode selection mechanism ( msm )
shifts were interpolated linearly in time over the course of the observations to determine the msm shift appropriate for each individual exposure of the target .
an additional complication in the velocity calibration arises from the fact that hst moves around the earth with an orbital velocity of @xmath63 .
this can cause velocity shifts between observations taken at different times of up to @xmath64 .
this exceeds the velocity dispersion of m15 , and it is therefore imperative to correct for this .
this is not normally done by the stis / ccd pipeline .
however , the support files that are created and delivered with hst data contain header keywords ( velocstx , velocsty , velocstz ) that contain the three - dimensional velocity vector of hst ( as measured halfway through the exposure time ) .
phil hodge ( priv .
comm . ) kindly provided a routine to calculate the velocity component in the direction of the target , given its ra and dec .
we used this routine to calculate this velocity for each exposure . for each exposure , the hst velocity shift was combined with the msm wavelength shift calculated previously ( using the fact that 1 pixel is @xmath65 at the center of the wavelength range ) .
the shifts calculated for different exposures at the same target position were then averaged , weighted by the individual exposure times .
these shifts were then written to the header keywords in the co - added and reduced target spectra .
the task x2d was then used to read these keywords , and perform a two - dimensional rectification to a linear heliocentric wavelength scale along rows , and a linear spatial scale along columns .
this also performs a flux calibration to units of @xmath66 .
one problem in all stis spectroscopy of point sources is the fact that the spatial pixel size of @xmath55 undersamples the hst psf .
for the calibration star observations we used 2-pixel on - chip binning in the spatial direction to obtain an effective pixel size of @xmath62 , which further increases the undersampling .
it is useful to have an understanding of the influence of this on the inferred spectra . for any spectrograph ,
a spectrum that falls on the ccd is generally slightly titled ( by of order a few degrees ) with respect to the detector rows . for a point source that is well sampled , the 2d spectrum will have the appearance of a smooth tilted line
. however , for a point source that is undersampled , the 2d spectrum will a have the appearance of a stairwell . during data reduction ,
the spectrum is interpolated and rotated to make it align with the rows of the frame .
however , this does not remove any undersampling . in an undersampled situation ,
the rectified spectrum will on average be horizontal , but it will have the zig - zag appearance of an un - tilted stairwell .
if all the rows in the rectified spectrum are summed , one recovers the correct spectral shape .
however , any horizontal cut along an individual row will have strong low - frequency undulations multiplied into the spectrum .
these effects are illustrated in figure [ f : undersamp ] , which shows the results for one of the two - dimensional spectra of the calibration star .
the top curve is the rectified spectrum summed along all rows , which provides a fairly accurate representation of the actual spectrum of the star .
by contrast , the bottom curve shows the spectrum for the brightest row in the rectified image .
it has low - frequency undulations multiplied into the spectrum .
individual rows in all of the rectified spectra obtained for our program look similar to this .
we obtained most of the m15 data without spatial rebinning , but this does not remove the undulations .
it merely decreases their amplitude and increases their frequency by a factor of 2 . at first glance
, one might think that an undulating spectrum such as in figure [ f : undersamp ] is completely useless , unless one can somehow correct for the undersampling . however , it turns out that for the present purposes these undulations are not a problem at all .
we wish to measure radial velocities from the doppler shifts in absorption lines .
the absorption lines are much narrower than the undulations in the spectrum , and their positions are not in any way affected . in cross - correlation techniques one always filters out low frequency components , and we have found that this efficiently removes any influence of the undulations on the inferred velocities .
figure [ f : crosscor ] illustrates this by showing the cross - correlation function of the two spectra shown in figure [ f : undersamp ] , both without and with low - frequency filtering .
the tests described below and in paper ii have convinced us that the undulations do not in any way compromise our ability to infer radial velocities . after the reduction of the calibration star data
, we have 14 different two - dimensional spectra . in each spectrum
there is signal in more than one row , because the stis psf is not a delta function .
all rows that contain signal can be analyzed as separate spectra .
for the analysis it proved useful to also have a single grand - total one - dimensional spectrum for each position of the calibration star .
such spectra were obtained by co - adding all rows in a given two - dimensional frame .
we wish to use the spectra to calibrate the stis psf and lsf . for each spectrum ( either for a given row , or grand - total ) we therefore extracted two numbers : ( a ) the total count rate per second @xmath67 summed over all wavelengths ; and ( b ) the wavelength shift @xmath68 of the spectrum with respect to some fixed standard spectrum .
the standard was chosen to be the grand - total spectrum obtained with the star at the origin of the coordinate system shown in figure [ f : hdslit ] .
the wavelength shifts were determined by cross - correlation , using techniques that are described in detail in paper ii .
figure [ f : calibtot ] shows @xmath67 and @xmath68 as function of the offset @xmath69 of the star perpendicular to the slit ( defined as in figure [ f : hdslit ] ) .
the count - rate @xmath67 was normalized arbitrarily ; its absolute value scales linearly with the brightness of the star , and is of no interest in the present context .
the count - rate is large if the star is centered in the slit , and of course , it drops if the star is moved further from the slit .
however , the count - rate remains positive even if the star is outside the slit , due to the wings of the psf .
the peak of @xmath67 identifies the center of the slit , @xmath70 .
this peak does not occur at @xmath71 , which implies that the star was not perfectly centered in the slit after the target acquisition procedure .
the shift @xmath68 is measured in pixels ( in the wavelength direction ) .
it is not a constant number , but shows a well - defined behavior as function of @xmath69 .
the observed behavior is easily understood qualitatively . at each wavelength
, the grating makes an image of the slit on the detector .
if the light that enters through the slit is not centered in the slit , then the observer perceives a shift in wavelength .
let @xmath72 be the shift ( with respect to the arbitrary standard ) for a source placed in the center of the slit ( @xmath73 ) .
if the calibration star is positioned at @xmath74 , then the intensity - weighted average @xmath69 coordinate of the light that enters through the slit has @xmath75 .
this is perceived as a positive wavelength shift , @xmath76 . alternatively ,
if @xmath77 , then @xmath78 . if a star is moved out of the center of the slit , then initially @xmath79 increases monotonically
. however , this behavior ultimately reverses . for large values of @xmath80 only light in the wings of the psf enters through the slit .
these wings have relatively small intensity gradients , and the slit will be illuminated more or less homogeneously .
this causes @xmath79 to revert to values near zero . it is easy to see that @xmath81 can never be more than half the slit width , i.e. , @xmath82 ( this limit corresponds to a delta function psf ) .
taking into account the anamorphic magnification of the grating , one detector pixel in the wavelength direction corresponds to @xmath83 ( bowers & baum 1998 ) .
thus one always has @xmath84 pixels , or alternatively that the peak - to - peak variation in @xmath68 must be less than @xmath85 pixels .
the observed peak - to - peak variation in @xmath68 is less , but still amounts to @xmath86 pixels ( cf .
figure [ f : calibtot ] ) .
this corresponds to @xmath87 , which is more than the m15 velocity dispersion .
it is therefore important to be able to model this effect , and to correct for it in the analysis of the m15 data . to model the calibration star data we consider the case of observations of a point source through a rectangular aperture .
the aperture size is @xmath88 , where @xmath89 is the slit width and @xmath90 the pixel size .
we adopt a coordinate system with the origin at the center of the aperture .
we assume that the psf can be modeled as a circularly symmetric sum of gaussians , @xmath91 .
\label{psfgauss}\ ] ] the individual @xmath92 may be either positive or negative , but must satisfy @xmath93 .
the associated encircled flux curve @xmath94 is given by @xmath95 .
\label{encircled}\ ] ] we assume that the aperture is perfectly rectangular , with no diffraction off the aperture edges . let the target be a point source positioned at @xmath96 .
the fraction of the light from the point source that is observed through the aperture is then @xmath97 -\mathop{\rm erf}\nolimits \left
[ \fr{x_0-(a_x/2)}{\sqrt{2}\ > \sigma_i } \right ] \right \rbrace \ > \left \lbrace \mathop{\rm erf}\nolimits \left
[ \fr{y_0+(a_y/2)}{\sqrt{2}\ > \sigma_i } \right ] -\mathop{\rm erf}\nolimits \left [ \fr{y_0-(a_y/2)}{\sqrt{2}\ > \sigma_i } \right ] \right \rbrace , \label{psffrac}\ ] ] where @xmath98 is the error function ( van der marel 1995 ) . the normalized intensity distribution in the @xmath99 direction , integrated over the @xmath100-direction , of the light that enters through the aperture , is @xmath101 \
> \left \lbrace
\mathop{\rm erf}\nolimits \left [ { { y_0+(a_y/2 ) } \over { \sqrt{2}\ > \sigma_i } } \right ] -\mathop{\rm erf}\nolimits \left [ { { y_0-(a_y/2 ) } \over { \sqrt{2}\ > \sigma_i } } \right ] \right \rbrace , & \quad \vert x \vert \leq { { a_x}\over2 } ; \\ 0 , & \quad \vert x \vert > { { a_x}\over2 } .
\end{array } \right .
\label{fofx}\ ] ] for our observational setup , the coordinate @xmath99 is related to the wavelength @xmath102 through the equation @xmath103 .
the lsf is the convolution of @xmath104 with the additional broadening functions due to the microroughness of the grating , the natural diffraction properties of light , and the binning into pixels ( van der marel , de zeeuw & rix 1997 ) .
however , the latter are all to lowest order symmetric functions of wavelength that do not cause net wavelength shifts .
the only net wavelength shift therefore comes from the fact that @xmath105 ( the intensity - weighted average @xmath99-coordinate of the light that enters through the aperture ) is generally different from zero ( the x - coordinate of the aperture center ) .
integration of equation ( [ fofx ] ) yields @xmath106 - \exp \left [ - \left ( { { x_0-(a_x/2 ) } \over { \sqrt{2}\ > \sigma_i } } \right ) ^2 \right ] \right \rbrace \nonumber \\ & & \times \left \lbrace \mathop{\rm erf}\nolimits \left [ { { y_0+(a_y/2 ) } \over { \sqrt{2}\ > \sigma_i } } \right ] -\mathop{\rm erf}\nolimits \left [ { { y_0-(a_y/2 ) } \over { \sqrt{2}\ > \sigma_i } } \right ] \right \rbrace
. \label{xmean}\end{aligned}\ ] ] with the above equations we are now in a position to model the observed quantities @xmath67 and @xmath68 .
the count - rate @xmath107 , where @xmath108 is a constant that is proportional to the brightness of the target and the transmission of the telescope+instrument system .
the wavelength shift @xmath68 is @xmath109 , or alternatively in terms of velocity , @xmath110 .
the latter equation is formally valid only at the central wavelength of the grating setting , but the dependence on wavelength can be neglected in the present context . one additional effect needs to be taken into account , namely the pixel scattering function of the ccd . a photon that falls onto a given detector pixel
has a finite probability of being detected in an adjacent pixel .
this effect can be quantitatively described as convolution with a kernel of size @xmath111 pixels .
the tinytim software package ( krist & hook 2001 ) provides an approximation for this kernel at wavelengths of 4000 and 5500 .
these can be interpolated to the central wavelength of the our observations , which yields : @xmath112 .
the convolution in the spectral direction of the ccd corresponds to a symmetric line broadening , which does affect our radial velocity measurements for m15 .
the only important effect in the present context is the spatial broadening along the slit .
this is described by the 3 pixel convolution kernel @xmath113 , which is the spatial projection of the @xmath111 kernel listed above . to obtain predictions for the observed values @xmath67 and @xmath68 at a pixel @xmath114 along the slit , one must first calculate the predictions at pixels @xmath115 , @xmath114 , and @xmath116 , and then apply the convolution kernel : @xmath117 to interpret the calibration star data we need to fit several unknowns .
the most important parameters to be determined are the parameters @xmath92 and @xmath118 that describe the multi - gaussian psf in equation ( [ psfgauss ] ) .
in addition , there are : ( a ) the constant @xmath108 defined through the equation @xmath119 ; ( b ) the position @xmath70 of the slit center , which measures the accuracy of the target acquisition procedure ; ( c ) the shift @xmath120 ( with respect to the arbitrary standard ) for a source placed in the center of the slit ; and ( d ) the offset @xmath121 of the target with respect to the center of a pixel along the slit , at the end of the target acquisition procedure .
we assume that all telescope slews between observations were accurate , as they are believed to be ( to the level of milli - arcsecs ) .
the relative positions of the star for the different observations are thus fixed by the values commanded to the telescope ( which are shown in figure [ f : hdslit ] ) . to constrain all the unknown parameters we used the measurements for @xmath67 and @xmath68 , not only for the grand - total spectra as shown in figure [ f : hdslit ] , but also for each of the rows in the individual two - dimensional spectra . while these constraints by themselves were sufficient to obtain acceptable results , we found that the calibration star data do not very accurately constrain the psf at intermediate and large radii .
to improve this situation we used a model estimate for @xmath122 as an additional constraint in the fit .
for this , the tinytim software was used to calculate the stis psf at the central wavelength of our observations . while this psf is formally appropriate only for the stis imaging mode , no better approximation for spectroscopy is readily available .
the tinytim psf and its associated encircled flux curve are shown in figure [ f : psf ] .
of course , addition of a model psf as a constraint in the fit can make the results worse rather than better , if the model psf is not in fact consistent with the calibration star data .
we have found no evidence for this , and in fact , from a variety of tests we conclude that the two are fully consistent . to determine the complete set of parameters in our multi - gaussian psf model that best fits all the combined constraints , we used a ` downhill simplex ' minimization routine ( press et al .
we found that adequate results were obtained with @xmath123 gaussians in the multi - gaussian psf .
the thin solid curves in figures [ f : calibtot ] and [ f : psf ] shows the predictions of the best fitting model with this @xmath124 .
the parameters of the multi - gaussian psf are listed in table [ t : psf ] . as has been discussed ,
there are many complications in the determination of stellar velocities with hst / stis at the few km / s level : hst orbits the earth as it exposes ; the wavelength scale may drift during an orbit ; and a wavelength shift is perceived for a star that is not perfectly centered in the slit .
the most important result of the calibration star observations is therefore the data - model comparison shown in the bottom panel of figure [ f : calibtot ] .
the model predictions fit the observed wavelength shifts @xmath68 with an rms residual of @xmath125 pixels .
this corresponds to @xmath126 .
so if the position of a star with respect to the slit is known , then the observed line of sight velocity can be calibrated to this accuracy . since the central velocity dispersion of m15 is @xmath127 , this accuracy is sufficient for the purposes of the present program .
the excellent fit in the bottom panel of figure [ f : calibtot ] verifies not only our psf and lsf modeling procedures , but also provides a stringent test of our data acquisition strategy and data reduction procedures .
figure [ f : psf ] shows that our multi - gaussian psf model is essentially a smooth fit to the true psf .
computational convenience is gained , but some of the intricate detail of the psf is lost .
most noticeably , the multi - gaussian psf does not reproduce the first airy ring of the psf ( see in particular the middle panel of figure [ f : psf ] ) .
however , this is no reason for great concern . to appreciate this , note that the psf is itself a purely theoretical construct that is generally unobservable .
what is observed is the convolution of the psf with the response function of an individual pixel , denoted the ` effective psf ' ( epsf ) by anderson & king ( 2000 ) . due to the convolution with the pixel size ,
the airy rings are much less pronounced in the epsf than in the psf . as a result ,
the epsf of our model is much closer to the true epsf , than is the psf of our model to true psf .
in fact , the observable quantities of interest for this program are exactly those shown in figure [ f : calibtot ] , and indeed , those are adequately reproduced by the model .
we obtained spectra of the central region of m15 for a period of 24 hst orbits , separated over 6 different visits . the first visit ( 5 orbits )
was executed on october 21 , 1999 .
based on the analysis of the data from the first visit some minor modifications were made to the observing strategy .
the remaining visits were then planned for the third week of may 2001 .
indeed , the data for visit 2 ( 4 orbits ) were obtained on may 15 , 2001 . however , on may 16 , 2001 , the primary ( side 1 ) set of stis electronics failed , which took the instrument out of commission for two months .
stis was then revived using the redundant ( side 2 ) set .
this did not affect the instrument characteristics in a major way , apart from a minor increase in the detector read - noise and a somewhat different dark count rate .
the remaining visits 35 ( 4 orbits each ) and 6 ( 3 orbits ) executed between october 2226 , 2001 . a complete log of the observations is presented in table [ t : obslog ] .
all observations were done with the same slit width and grating setting as described in section [ s : calibspec ] .
exposures were taken at 18 different positions covering the central region of the cluster .
figure [ f : center ] shows the slit positions as they were commanded to the telescope .
the accuracy with which the intended positioning was achieved is discussed in section [ ss : slitpos ] below .
for all observations the slit was aligned along position angle @xmath128 ( measured from north over east ) .
this pa was chosen to allow the most flexible scheduling of the observations , and is otherwise fairly arbitrary . at each slit position we obtained spectra for a total exposure time that ranged from 46 to 60 minutes .
three or four exposures were obtained per position .
individual exposures were dithered by 4 pixels along the slit , to allow removal of ccd defects during data reduction . in visit 1
we placed the central region of m15 near the center of the ccd . in subsequent visits we placed the central region near row 900 ( out of 1024 ) , to mitigate the loss of signal due to deteriorations in the cte of the stis / ccd . in no case
did we apply on - chip rebinning in the spectral direction , given that the @xmath65 size of a single pixel is already large compared to the m15 velocity dispersion .
however , in the first visit we applied 2-pixel on - chip rebinning in the spatial direction ( as for the calibration star observations discussed in section [ s : calibspec ] ) to minimize the read - noise component in the spectra . in subsequent visits we did not apply such on - chip rebinning , to minimize the effects of psf undersampling in the spatial direction . the large majority of the orbits consisted of the same simple observational sequence : an arc lamp exposure , followed by three target exposures and then another arc lamp exposure . in the end , the spectra from all visits had sufficient quality to achieve the science goals of our program . blind pointings with hst are accurate only at the level of @xmath129 , due to the accuracy of the gscs . to position the stis slit on m15 with the required accuracy it is therefore necessary to perform a target acquisition .
stis target acquisitions work by centering on the brightest object in a @xmath130 field . to ensure that the target acquisition would not acquire the wrong target
, we searched for a bright and relatively isolated star in m15 .
the cross in figure [ f : catimage ] marks the spatial position of the star that was chosen ( # 27604 in the wfpc2 catalog ) .
its position in the @xmath131 coordinate system defined in section [ ss : astrometric ] is @xmath132 .
its position in the m15 cmd shown in figure [ f : cmd ] is also indicated by a cross .
each visit started with an acquisition of this star , consisting of an acq acquisition procedure .
in most visits we then slewed the telescope to the intended position near the center of m15 to take the spectra .
however , in visits 1 and 6 we performed an additional acq / peak acquisition procedure ( leitherer et al .
2001 ) and obtained spectra of the acquisition star , before slewing to the center of m15 . in visit 1
we also repeated this procedure at the end of the visit , to allow an independent check of the repeatability of the velocity measurements for this particular star .
the positional accuracy of the target acquisitions and the slews is discussed in section [ ss : slitpos ] below .
we generally performed a target acquisition only at the beginning of the first orbit of each visit , and not in any of the subsequent orbits .
the data reduction of the m15 spectra was similar to the reduction of the calibration star spectra discussed in section [ ss : calibreduc ] , including the wavelength calibration using arc lamp spectra obtained in the same orbit , and the correction for the hst velocity during the observations .
this yields for each slit position a single two - dimensional frame with a linear heliocentric wavelength scale along rows , and a linear spatial scale along columns . one difference between the observations of m15 and those of the calibration star is that we obtained only two arc - lamp spectra per orbit for m15 , as compared to five for the calibration star .
we did some quantitative analysis to address to what extent this may affect the analysis .
the arc lamp spectra show that there are drifts in the wavelength scale over the course of a visit .
these drifts are generally smoothly varying , and are fairly well modeled as a linear function of time . in the worst case ( visit 2 ) the drift was @xmath133 pixels over the 4 orbits of the visit . because we do in fact take two arc - lamp spectra per orbit , we can correct for most of this . in the reduction of the m15 spectra
we assume that the wavelength scale for a given exposure can be obtained by linear interpolation in time between the results for the nearest arc lamp spectra . to test the accuracy of this
, we applied this same procedure to each of the arc - lamp spectra .
this yields the difference between the actual wavelength scale and the one that is obtained by linear interpolation between the previous and the next arc - lamp exposure in the visit .
the rms difference for all arc lamp spectra was found to be @xmath134 pixels , which corresponds to only @xmath135 .
this accuracy is sufficient for the purposes of the project , especially since other effects already set a lower limit to the velocity accuracy of @xmath126 ( cf .
section [ ss : calibdiscuss ] ) . for accurate velocity determination of an m15 star it is essential to know its position with respect to the slit .
from the wfpc2 imaging we know where all the stars are on the sky ( see section [ s : imaging ] ) .
so the next step is to determine the exact positioning of the slits for all observations . for this we analyzed and modeled the observed intensity profiles along the slit .
these essentially provide one - dimensional ( linear ) images of m15 , which uniquely constrain the telescope pointing . to model an observed intensity profile along the slit we proceed as follows .
for each star in the wfpc2 catalog we use the observed johnson b and v magnitudes to estimate the flux @xmath136 at 5216 ( the center of the wavelength range for the spectra ) .
we choose a trial position of the slit on the sky . for each pixel along the slit
we calculate the flux contributed by each star , using the formulae derived in section [ ss : psflsf ] . adding the contributions of all stars in the catalog yields the predicted intensity profile along the slit for the given assumed slit position .
this profile can be compared to the observed profile in a @xmath137 sense ( using an arbitrary scaling that corresponds to the efficiency of the telescope+instrument system ) .
looping over a fine grid of possible slit positions yields the position with minimum @xmath137 , which corresponds to the estimated position of the slit on the sky . to obtain adequate fits we found that we had to add a smooth , low - intensity model for the light contribution of faint stars that are missing from the wfpc2 catalog due to incompleteness .
as an example of the procedure , figure [ f : intprof ] shows the final data - model comparison for one of the slit positions . the accuracy with which hst can achieve a requested roll angle is @xmath138 .
the accuracy with which the slit pa on the sky is known is dominated by the calibration uncertainty in the relative angle of the slit within the instrument , which is @xmath139 .
this is small enough that it can be neglected in the present context .
we therefore did not include the slit pa as a free parameter in the fits to the intensity profiles , but kept it fixed at the value commanded to the telescope .
to confirm the validity of this approach we did a number of test runs in which we fitted observed m15 intensity profiles with the slit pa treated as a free parameter .
these tests indeed implied pa values consistent with the value commanded to the telescope , to within the quoted uncertainty .
the intensity profile modeling yields for each observation the actual pointing position of the telescope during the exposures .
the intended pointings are known as well , so it is possible to determine for each observation the offset @xmath140 of the actual pointing from the intended pointing .
the results are shown in figure [ f : pointings ] .
observations from the same visit are shown with the same plot symbol .
the offsets satisfy @xmath141 and @xmath142 for all observations .
these absolute pointing accuracies are not so important in the present context ; the aim of the observational strategy was to scan the central region of m15 , and not to point at any object in particular .
however , it is important to know the accuracy of our determinations of @xmath140 . to this end
we need to understand the sources of the inferred pointing offsets .
the first source of pointing error is the limited astrometric accuracy of the wfpc2 catalog .
figure [ f : catimage ] shows that the star that was used for target acquisitions was on a different ccd chip in the wfpc2 observations than the center of m15 .
the astrometric calibration of wfpc2 is considerably less accurate across ccd chips than within a given ccd chip ( @xmath0 as compared to @xmath143 , according to the hst data handbook ; voit et al .
since the slew that we applied from the acquisition star to the m15 center was based on the astrometry of our catalog , an error in this astrometry would yield a fixed non - zero component to the offsets @xmath140 for all observations .
the second source of pointing error is the limited accuracy with which the positions of the apertures are known within the stis instrument .
an acq acquisition centers the target in the 0.2x0.2 aperture . on - board software
then applies a telescope slew to place the target behind the science aperture ( the 52x0.1 in the present case ) .
if there is an error in the expected position of the science aperture , this yields a fixed non - zero pointing error relative to the slit .
we obtained the data for visit 2 in a configuration that is rotated by @xmath144 with respect to that for the other visits .
so any component in the offsets @xmath140 due to this effect would have the opposite sign for visit 2 as compared to the other visits .
figure [ f : catimage ] shows that the average offset for visit 2 is @xmath145 , whereas the average for all the other visits is @xmath146 .
so there is indeed evidence for pointing errors of this nature .
the third source of pointing error is the limited accuracy of the target acquisitions that start each visit . for most visits we only used an acq acquisition procedure , and not the more accurate acq / peak procedure .
the former has a nominal accuracy between @xmath59 ( leitherer et al .
2001 ) and @xmath60 ( voit et al .
1997 ) .
the fourth source of pointing error is the limited accuracy of the @xmath147 slews from the acquisition star to the m15 center .
the nominal accuracy of such slews is @xmath148 ( downes & rose 2001 ) .
the first four sources of error can explain both the size and the scatter in the pointing offsets for different visits . however , these error sources predict identical offsets for all observations in the same visit . in reality
, there is some scatter among the pointing offsets for observations in the same visit , but this scatter is very small : @xmath143 in the @xmath69 direction , and @xmath1 in the @xmath149 direction .
we attribute this to the fifth source of pointing error : small pointing drifts during a visit as a result of thermal effects on the telescope ( downes & rose 2001 ) .
the small scatter in the offsets for observations in the same visit implies that our models for determining the slit positions from the data must be very accurate
. the numerical accuracy can be no worse than @xmath1 in each coordinate .
an error of this size in the stis @xmath69 coordinate corresponds to a wavelength shift @xmath150 of at most @xmath134 pixels , cf .
the bottom panel of figure [ f : calibtot ] .
this corresponds to @xmath151 .
this accuracy is sufficient for the purposes of the project , especially since other effects already set a lower limit to the velocity accuracy of @xmath126 ( cf . section [ ss : calibdiscuss ] ) .
each row in each of the 18 calibrated and rectified two - dimensional spectra contains the spectrum for a small aperture on the sky . from the analysis in section [ ss : slitpos ] , the positions of these apertures are known with a ( two - dimensional ) accuracy better than @xmath59 .
figure [ f : apertures ] shows the inferred aperture positionings for the central region of m15 . in total
there are 19,200 spectra in the dataset .
most spectra contain very little signal from any star , and are therefore of little use in the analysis .
however , some spectra do contain considerable signal from one or more m15 stars .
many stars contribute significantly to more than one aperture . for proper analysis we therefore developed an algorithm to extract the spectra of individual stars .
the algorithm starts with the same modeling procedure used in sections [ ss : psflsf ] and [ ss : slitpos ] . from the known positions of all apertures @xmath114 and all stars @xmath152 we calculate predictions for three sets of quantities : ( i ) @xmath153 , the total flux observed in an aperture @xmath114 ; ( ii ) @xmath154 , the total flux contributed by star @xmath152 to aperture @xmath114 ; and ( iii ) @xmath155 , the velocity offset with which star @xmath152 contributes flux to aperture @xmath114 .
let @xmath156 be the total flux of star @xmath152 .
we define @xmath157 to be the fraction of the flux of star @xmath152 that falls in aperture @xmath114 .
we also define @xmath158 to be the fraction of the total flux in aperture @xmath114 that is contributed by star @xmath152 . for each aperture @xmath114
there is one star @xmath152 in the catalog for which @xmath159 has the largest value .
we denote the index number of this star as @xmath160 .
we denote an aperture as ` usable ' if the following three statements are true : ( i ) @xmath161 ( i.e. , there is a single star that contributes at least half of all the flux in the aperture ) ; ( ii ) @xmath162 ( i.e. , there is a non - negligible amount of flux from this star in the aperture ; this limit corresponds to a @xmath43-band magnitude of @xmath163 ) ; and ( iii ) @xmath164 ( the aperture contains a non - negligible fraction of the total flux of this star ) .
apertures that are not usable are ignored in all of the subsequent analysis . for each star
@xmath152 we then make a list of all the apertures @xmath114 for which @xmath165 .
let @xmath166 be the number of apertures for which this is the case .
we denote a star to be ` significant ' for our study if @xmath167 .
we call the spectra for the usable apertures that make up @xmath166 the ` usable ' spectra for star @xmath152 .
in the full dataset there are 1741 usable spectra for 636 significant stars .
the goal is to combine the @xmath166 usable spectra for each significant star @xmath152 into a final spectrum for that star .
this can be done by taking the exposure - time weighted average of the usable spectra .
however , there is the possibility to include or exclude each usable spectrum in or from the average , so there are @xmath168 possible combinations to do the averaging .
specifically , one does not want to include spectra whose main effect is to increase the amount of noise or blending in the average .
we developed an algorithm to optimize the quality of the final spectrum .
the stis data reduction pipeline creates an error frame for each spectrum .
we use the data frames and the error frames to calculate for each usable spectrum the median flux @xmath169 from the star ( i.e. , the median observed flux multiplied by @xmath159 ) and the median noise . from these quantities
we calculate the signal - to - noise ratio @xmath3 for all of the @xmath168 possible combinations of spectra .
we also calculate for all combinations the fraction @xmath170 of the total flux in the averaged spectrum that is due to the significant star .
we rejected all combinations in which @xmath170 is smaller than some value @xmath171 .
of the remaining combinations we adopt the one that produces the highest @xmath3 . the instrumental velocity shift @xmath172 of the final average spectrum of star @xmath152
is calculated as the weighted average of the @xmath155 for the apertures @xmath114 that are used in the combination ; the weighting is done using the product of exposure time and the flux @xmath169 .
we adopted @xmath173 in the extraction , to ensure that the fraction of the flux in the final spectrum that is blended in from other stars is less than 25% .
for the spectra to be useful for kinematical analysis they must have a sufficient @xmath3 . as demonstrated in paper ii ,
an average @xmath174 per pixel is a minimum requirement ( although not a sufficient requirement ) to obtain meaningful line - of - sight velocity measurements .
all extracted spectra that did not satisfy this requirement were discarded .
this ultimately yielded ` acceptable ' spectra for 131 stars , out of which 15 stars are within @xmath175 from the cluster center , 39 stars are within @xmath176 , and 87 stars are within @xmath177 .
figure [ f : spectra ] shows the spectra for four stars in the final sample .
all four spectra correspond to stars on the giant branch . from top to bottom , the @xmath43 magnitudes range from @xmath178 to @xmath179 while the average @xmath3 per pixel ranges from @xmath180 to @xmath181 .
although the bottom spectrum is quite noisy , the mg b triplet lines at @xmath182 are still clearly visible .
the faintest star for which we were able to extract a spectrum with @xmath183 from the stis data has magnitude @xmath184 , which is just above the main sequence turnoff .
the spectra in figure [ f : spectra ] are all reasonably ` flat ' , and do not show the undulations seen in figure [ f : undersamp ] .
this is because the spectra are sums of individual apertures for several slit positions .
each of the individual spectra has undulations , but these tend to cancel in the final co - addition .
approximately 10 per cent of the stars in our final list of 131 are based on only a single aperture .
the final spectra for these stars show undulations similar to those in figure [ f : undersamp ] .
as explained in section [ ss : undersampling ] , this does not compromise our ability to infer radial velocities for these stars .
the 131 individual stellar spectra that we extracted from the stis data form the end product of the present paper .
it is shown in paper ii that accurate velocities can be determined for 64 of these stars .
the spatial and color - magnitude distributions of these stars are discussed in paper ii .
that paper also discussed the implications of the inferred velocities for our understanding of the structure and kinematics of m15 .
the globular cluster m15 is one of the densest globular clusters in our galaxy .
a variety of interesting dynamical and structural evolutionary phenomena have been predicted to occur at high stellar densities , including mass segregation and the possible formation of a central black hole .
stellar kinematical data are required required to address these issues observationally .
while m15 has been one of the globular clusters that has been most intensively studied in the past decades , ground - based data have only been able to provide limited information in the central few arcsec .
we therefore executed a project to map the center of m15 spectroscopically at high spatial resolution with hst / stis .
our observational project is at the limit of what is feasible with hst / stis .
the pixel size of our spectra corresponds to @xmath65 , which exceeds the central velocity dispersion of m15 ( @xmath185 ; gebhardt et al .
extreme care therefore had to be taken in the reduction and velocity calibration of the data .
the main result of the present paper is that we have in fact succeeded in extracting accurately calibrated spectra from our data .
the implications of this for our understanding of m15 are discussed in paper ii . to be able to properly interpret the long - slit spectra , we started by creating a stellar catalog for m15 .
we used the hst / wfpc2 data discussed previously by g96 , but we performed an improved astrometric and photometric calibration .
the final stellar catalog contains 31,983 stars with their positions and u , b and v magnitudes .
it is spatially complete within @xmath49 from the cluster center , and extends as far as @xmath46 arcmin from the cluster center in some directions .
the catalog contains stars down to magnitude @xmath53 , and it is photometrically complete for stars brighter than @xmath52 . for proper calibration of the m15 spectra we performed a set of observations of a bright field star , with the star placed at various positions with respect to the slit .
after basic data reduction and correction for the velocity of hst around the earth we used these spectra for modeling the psf and lsf of stis .
this is particularly important for understanding the apparent velocity shift for a star that is offset from the center of the slit .
we derive a simple multi - gaussian model for the psf that reproduces the observed shifts with an rms residual of only @xmath126 ( for a known stellar position with respect to the slit ) .
spatial undersampling causes artificial undulations in the spectra , but this only affects low frequencies and does not affect the determination of radial velocities through cross - correlation .
the m15 observations consisted of 6 hst visits spread over 2 years .
the central region of the cluster was ` scanned ' by consecutively placing the @xmath0 slit at 18 adjacent positions .
the intensity profiles along the slits were used to accurately determine the actual slit positions on the sky . with use of our wfpc2 catalog and psf model
we find that these positions can be determined to an accuracy of @xmath1 in each coordinate .
positional errors in our knowledge of the slit positions translate directly into velocity errors , but with the achieved positional accuracy these errors are no larger than @xmath151 .
the reduced and calibrated data set contains 19,200 stis spectra , each for a different aperture position in m15 .
we developed an algorithm that co - adds the spectra for individual apertures , so as to extract spectra of individual stars with minimum blending and maximum @xmath3 .
our data allowed us to extract spectra with an average @xmath183 per pixel for a total of 131 stars , out of which 15 stars are within @xmath175 from the cluster center , 39 stars are within @xmath176 , and 87 stars are within @xmath177 . in paper
ii we show that reliable line - of - sight velocities can be derived for 64 of these stars , and that this provides enough new information to significantly improve our understanding of the dynamics and structure of m15 .
support for proposals # 8262 was provided by nasa through a grant from the space telescope science institute , which is operated by the association of universities for research in astronomy , inc .
, under nasa contract nas 5 - 26555 .
we thank pierre dubath for helpful advice in the early stages of this project , and brian yanny , don schneider and john bahcall for collaboration on the acquisition of the wfpc2 data .
we thank the anonymous referee for useful feedback that helped improve the presentation of the paper .
rrrccc 1 & 16.343 & 10.269 & 20.467 & 21.313 & 20.869 + 2 & 18.180 & 8.384 & 20.889 & 21.701 & 21.624 + 3 & 7.251 & 19.504 & 20.271 & 20.900 & 20.623 + 4 & 18.725 & 7.807 & 19.363 & 19.893 & 19.738 + 5 & 2.529 & 24.193 & 20.256 & 20.583 & 20.534 + 6 & 2.170 & 24.551 & 20.325 & 20.994 & 20.738 + 7 & 17.248 & 9.263 & 20.738 & 21.234 & 21.206 + 8 & 11.883 & 14.729 & 14.659 & 15.562 & 15.894 + 9 & 1.843 & 24.853 & 20.619 & 21.485 & 21.278 + 10 & 15.338 & 11.195 & 19.454 & 19.935 & 19.840 + 11 & 22.580 & 3.733 & 20.957 & 21.767 & 21.958 + 12 & 19.811 & 6.584 & 18.549 & 19.215 & 19.148 + 13 & 6.533 & 20.103 & 19.973 & 20.440 & 20.372 + 14 & 3.334 & 23.280 & 20.557 & 21.253 & 20.911 + 15 & 16.777 & 9.611 & 20.453 & 21.105 & 20.926 + 16 & 3.759 & 22.784 & 20.885 & 21.633 & 21.377 + 17 & 8.709 & 17.789 & 20.776 & 21.377 & 21.155 + 18 & 13.186 & 13.240 & 20.985 & 22.073 & 21.703 + 19 & 8.817 & 17.666 & 19.188 & 19.698 & 19.478 + 20 & 21.074 & 5.129 & 19.719 & 20.445 & 20.145 + 21 & 4.649 & 21.815 & 20.424 & 20.885 & 20.718 + 22 & 6.399 & 20.045 & 20.568 & 21.132 & 20.988 + 23 & 20.324 & 5.839 & 20.359 & 21.056 & 20.829 + 24 & 6.592 & 19.834 & 19.828 & 20.382 & 20.294 + 25 & 16.308 & 9.954 & 20.307 & 21.043 & 20.993 + ccccc @xmath196 & 2862 & 4 & 1 & 6 + @xmath197 & 2790 & 4 & 2 & 1 + @xmath198 & 2760 & 3 & 2 & 1 + @xmath199 & 2760 & 3 & 2 & 1 + @xmath200 & 2760 & 3 & 2 & 1 + @xmath201 & 3113 & 4 & 1 & 2 + @xmath202 & 3578 & 4 & 1 & 2 + @xmath203 & 3556 & 4 & 1 & 2 + @xmath204 & 3113 & 4 & 1 & 3 + @xmath205 & 3578 & 4 & 1 & 3 + @xmath206 & 3556 & 4 & 1 & 3 + @xmath207 & 3113 & 4 & 1 & 4 + @xmath208 & 3578 & 4 & 1 & 4 + @xmath209 & 3556 & 4 & 1 & 4 + @xmath210 & 3113 & 4 & 1 & 5 + @xmath211 & 3578 & 4 & 1 & 5 + @xmath212 & 3556 & 4 & 1 & 5 + @xmath213 & 3554 & 4 & 1 & 6 + | in this series of two papers we describe a project with the space telescope imaging spectrograph ( stis ) on hubble space telescope ( hst ) to measure the line - of - sight velocities of stars in the central few arcsec of the dense globular cluster m15 . the main goal of this project is to search for the possible presence of an intermediate mass central black hole .
this first paper focuses on the observations and reduction of the data .
we ` scanned ' the central region of m15 spectroscopically by consecutively placing the @xmath0 hst / stis slit at 18 adjacent positions .
the spectral pixel size exceeds the velocity dispersion of m15 .
this puts the project at the limit of what is feasible with stis , and exceedingly careful and complicated data reduction and analysis were required .
we applied corrections for the following effects : ( a ) drifts in the stis wavelength scale during an hst orbit ; ( b ) the orbital velocity component of hst along the line - of - sight to the cluster , and its variations during the hst orbit ; and ( c ) the apparent wavelength shift that is perceived for a star that is not centered in the slit .
the latter correction is particularly complicated and requires many pieces of information : ( 1 ) the positions and magnitudes of all the stars near the center of m15 ; ( 2 ) the accurate positionings of the stis slits during the observations ; ( 3 ) and the hst / stis point - spread function ( psf ) and line - spread function ( lsf ) . to address the first issue we created a stellar catalog of m15 from the existing hst / wfpc2 data discussed previously by guhathakurta et al .
( 1996 ) , but with an improved astrometric and photometric calibration .
the catalog is distributed electronically as part of this paper .
it contains 31,983 stars with their positions and u , b and v magnitudes . to address the second issue we model the observed intensity profiles along the stis slits to determine the slit positionings to @xmath1 accuracy in each coordinate . to address the third issue we obtained observations of a bright field star to which we fitted multi - gaussian psf and lsf models . upon reduction of the m15 spectroscopy
we ultimately obtain 19,200 one - dimensional stis spectra , each for a different aperture position in m15 , with a velocity scale accurate to better than @xmath2 .
we develop an algorithm that co - adds the spectra for individual apertures and use it to extract spectra of individual stars with minimum blending and maximum @xmath3 . in paper ii ( gerssen et al . )
we use these spectra to extract reliable line - of - sight velocities for 64 stars , half of which reside within @xmath4 from the cluster center .
these velocities constrain the central structure , dynamics and mass distribution of the cluster . #
1#2 # 1 ' '' '' # 1 ' '' '' # 1 ' '' '' _ # 1 _ ' '' '' |
our understanding of baryon resonances is undergoing continuous change . from the classical picture of the baryons made out of three constituent quarks , passing through attempts to represent some of them in terms of pentaquarks , to the more recent description of some of them in terms of meson baryon molecules . in this sense
, the low lying @xmath9 resonances , @xmath10 , @xmath11 @xmath12 , can be represented as dynamically generated states from the meson baryon interaction and are relatively well understood within the unitary chiral models @xcite .
the low lying @xmath4 domain remains far less understood , both experimentally and theoretically .
for instance , quark models seem to face difficulties in reproducing properties of the resonances in this sector @xcite .
the neat reproduction of the low lying @xmath13 states in the @xmath3-wave meson - baryon interaction , using chiral dynamics , suggests that the addition of a pseudoscalar meson in s - wave could lead to an important component of the structure of the @xmath14 resonances .
chiral dynamics has been used earlier in the context of the three nucleon problems , e.g. , in @xcite .
we present in this talk the study done in @xcite of two meson - one baryon systems where chiral dynamics is applied to solve the faddeev equations .
as shall be described below , our calculations for the total @xmath3 = -1 reveal peaks in the @xmath15 system and its coupled channels which we identify with the resonances @xmath16 , @xmath17 , @xmath18 , @xmath19 , @xmath20 and @xmath21 .
we start by taking all combinations of a pseudoscalar meson of the @xmath22 su(3 ) octet and a baryon of the @xmath14 octet which couple to @xmath23 with any charge . for some quantum numbers ,
the interaction of this two body system is strongly attractive and responsible for the generation of the two @xmath10 states @xcite and other @xmath3 = -1 resonances .
we shall assume that this two body system formed by @xmath24 and coupled channels remains highly correlated when a third particle is added , in the present case a pion .
altogether , we get twenty - two coupled channels for the net charge zero configuration : @xmath25 , @xmath26 , @xmath27 , @xmath28 , @xmath29 , @xmath30 , @xmath31 , @xmath32 , @xmath33 , @xmath34 , @xmath35 , @xmath36 , @xmath37 , @xmath38 , @xmath39 , @xmath40 , @xmath41 , @xmath42 , @xmath43 , @xmath44 , @xmath45 , @xmath46 . we assume the correlated pair to have a certain invariant mass , @xmath47 , and the three body @xmath48-matrix is evaluated as a function of this mass and the total energy of the three body system . at the end
we look for the value of @xmath49 as a function of these two variables and find peaks at certain values of these two variables , which indicate the mass of the resonances and how a pair of particles is correlated .
the input required to solve the faddeev equations , i.e. , the two body @xmath50-matrices for the meson - meson and meson - baryon interactions has been calculated by taking the lowest order chiral lagrangian following @xcite and using the dimensional regularization of the loops as done in @xcite , where a good reproduction of scattering amplitudes and resonance properties was found . instead
, a cut off could also be used to regularize the loops as shown in @xcite .
improvements introducing higher order lagrangians have been done recently , including a theoretical error analysis @xcite which allows one to see that the results with the lowest order lagrangian fit perfectly within the theoretical allowed bands .
a shared feature of the recent unitary chiral dynamical calculations is the on - shell factorization of the potential and the @xmath50-matrix in the bethe - salpeter equation @xcite , which is justified by the use of the n / d method and dispersion relations @xcite .
alternatively , one can see that the off - shell contributions can be reabsorbed into renormalization of the lower order terms @xcite .
we develop here a similar approach for the faddeev equations .
the full three - body @xmath48-matrix can be written as a sum of the auxiliary @xmath48-matrices @xmath51 , @xmath52 and @xmath53 @xcite @xmath54 where @xmath55 , @xmath56 , @xmath57 , @xmath58 , are the normal faddeev partitions , which include all the possible interactions contributing to the three - body @xmath48-matrix with the particle @xmath59 being a spectator in the last interaction .
the faddeev partitions satisfy the equations @xmath60 where @xmath61 ( @xmath62 ) is the initial ( final ) momentum of the ith particle in the global center of mass system , @xmath63 is the two - body @xmath50-matrix for the interaction of the pair @xmath64 and @xmath65 is the three - body propagator or green s function , with @xmath66 = 1 , 2 , 3 the first two terms of the diagrammatic expansion of the faddeev equations , for the case @xmath59=1 , are represented diagrammatically in fig.[fig1 ] , diagrammatic representation of the terms ( a ) @xmath67 ( b ) @xmath68.,scaledwidth=50.0% ] where the @xmath50-matrices are required to be off - shell
. however , the chiral amplitudes , which we use , can be split into an `` on - shell '' part ( obtained when the only propagating particle of the diagrams , labeled with @xmath69 in fig.[fig1 ] , is placed on - shell , meaning that @xmath70 is replaced by @xmath71 in the amplitudes ) , which depends only on the c.m energy of the interacting pair , and an off - shell part proportional to the inverse of the propagator of the off - shell particle .
this term would cancel the particle propagator , ( @xmath72 ) , for example that of the 3rd particle in the fig.[fig1]a ) resulting into a three body force ( fig.[fig2]a ) .
in addition to this , three body forces also stem directly from the chiral lagrangians @xcite ( fig.[fig2]b ) . the origin of the three body forces ( a ) due to cancellation of the propagator in fig.[fig1](a ) with the off - shell part of the chiral amplitude , ( b ) at the tree level .
] we find that the sum of the off - shell parts of all the six @xmath73 terms , together with the contribution from fig.[fig2](b ) cancels exactly when the su(3 ) limit is considered .
details of the analytical proof can be seen in the appendices of @xcite .
hence , only the on - shell part of the two body ( chiral ) @xmath50-matrices is needed in the evaluations .
this is one of the important findings of these works because one of the standing problems of the faddeev equations is that the use of different potentials which give rise to the same on shell scattering amplitudes give rise to different results when used to study three body system with the faddeev equations .
the different , unphysical , off shell amplitudes of the different potentials are responsible for it .
the use of chiral dynamics in the context of the faddeev equations has then served to show that the results do not depend on these unphysical amplitudes and only the on shell amplitudes are needed as input . in this sense , since these amplitudes can be obtained from experiment , it is suggested in @xcite to use these experimental amplitudes , and sensible results are obtained in the study of the @xmath74 system and coupled channels .
the first term with a non trivial structure , from the point of view of the on - shell factorization of the @xmath50-matrices in the faddeev equations , is the one involving three successive pair interactions , where a loop function of three particle propagators appears for the first time .
we show the diagrams with such a structure for the @xmath51 partition in fig.[fig3](a - d ) .
different diagrams involving three pair interactions corresponding to the @xmath51 partition . ]
the strategy followed in the former works is that the terms with two , three , and four interactions are evaluated exactly .
then it is observed that the ratio of the four to three body interaction terms is about the same as that of the three body to two body .
once this is realized , the coupled integral equations are converted into algebraic equations , which renders the technical work feasible in spite of the many coupled channels used .
there are two independent variables in the formalism @xmath75 , @xmath47 , as a function of which we plot the squared @xmath76-matrix ( @xmath77 ) ) , where @xmath78 sum all terms with the first two interactions having @xmath79 as spectators .
two @xmath0 resonances in the @xmath80 amplitude in @xmath81 = 1 , @xmath82 = 2 configuration.,scaledwidth=60.0% ] we now report on the four isospin i=1 states found in our study . in fig.[fig4 ] , we show a plot of the squared @xmath76-matrix and its projection , for @xmath83 in the total isospin @xmath81 = 1 configuration obtained by keeping the two pions in isospin @xmath82 = 2 .
we see two peaks ; one at @xmath75 = 1656 mev with the full width at half maximum @xmath84 30 mev and another at @xmath75 = 1630 mev with @xmath85 = 39 mev .
we identify the peak at @xmath75 = 1656 mev with the well established @xmath86 @xcite as a resonance in the @xmath80 system , which is a new finding .
a detailed description of all the states that appear in this sector can be seen in @xcite . here
we summarize the results in table [ table1 ] . .@xmath0 and @xmath1 states obtained from the interaction of two mesons and one baryon . [
cols="^,^,^,^ " , ] this state is very interesting and it has been suggested in @xcite that it could be responsible for the peak around 1920 mev of the @xmath7 reaction @xcite . it was suggested that the spin of the resonance could be found performing polarization measurements which are the state of the art presently . in case
the @xmath87 assignment was correct , an easy test can be carried out to rule out the @xmath88 state .
the experiment consists in performing the @xmath7 reaction with a circularly polarized photon with helicity 1 , thus @xmath89 with the @xmath90-axis defined along the photon direction , together with a polarized proton of the target with @xmath91 along the same direction . with this set up
, the total spin has @xmath92 .
since @xmath93 is zero with that choice of the @xmath90 direction , then @xmath94 and @xmath95 must be equal or bigger than @xmath96 .
should the resonant state be @xmath87 , the peak signal would disappear for this polarization selection , while it would remain if the resonance was a @xmath97 state .
thus , the disappearance of the signal with this polarization set up would rule out the @xmath97 assignment . such type of polarization
set ups have been done and are common in facilities like elsa at bonn , mami b at mainz or cebaf at jefferson lab , where spin-3/2 and 1/2 cross sections , which play a crucial role in the gdh sum rule , see e.g. ref .
@xcite , were measured in the two - pion photoproduction @xcite reaction .
the theoretical analysis of @xcite shows indeed that the separation of the amplitudes in the spin channels provides information on the resonances excited in the reaction . in @xcite
it was also shown that the presence of this resonance would lead to an enhanced @xmath98 invariant mass distribution close to the @xmath98 threshold , and also it would produce an enhanced cross section in the @xmath99 close to threshold of the reaction .
these features are now being tested at spring8/osaka @xcite .
to finalize this selection of topics , we would like to address a recent study that finds some unexpected few body systems .
these are states with several bound @xmath8 mesons , up to six . in the pdg
@xcite there are intriguing mesons with large spin , of the @xmath8 and @xmath100 type , whose quantum numbers match systems made with 3 , 4 , 5 and 6 @xmath8 mesons with their spins aligned .
these are the @xmath101 ( @xmath102 ) , @xmath103 ( @xmath104 ) , @xmath105 ( @xmath106 ) and @xmath107 ( @xmath108 ) resonances .
the idea of having these states as multirho states stems from the findings of @xcite where the @xmath109 interaction was studied using a unitary coupled channel approach with the input from the hidden gauge lagrangians @xcite . in the work
@xcite it was found that the interaction of two @xmath110 mesons in isospin @xmath111 and spin @xmath112 was strong enough to bind the @xmath109 system into the @xmath113 ( @xmath114 ) resonance .
the nature of this resonance as a @xmath115 molecule has passed the tests of radiative decay into @xmath116 @xcite , the decay of @xmath117 into @xmath118 and @xmath113 ( together with other resonances generated in @xcite ) @xcite , and @xmath117 into @xmath119 and @xmath113 ( and the other resonances of @xcite ) @xcite .
the realization of this strong attraction in j=2 , suggest that other @xmath8 mesons could cluster , always with their spin aligned such that all pairs would have j=2 .
this is what has been done in @xcite .
one studies the interaction of a @xmath8 with the @xmath113 , using the fixed center approximation to the faddeev equations , and the scattering matrix leads to a clear bump that we associate to the @xmath101 ( @xmath102 ) .
after this , one studies the interaction of two @xmath113 , where the input is the scattering amplitude of @xmath8 with the @xmath113 , studied before .
a peak is also found in the scattering matrix , which we associate to the @xmath103 ( @xmath104 ) .
further iterations including every time one extra @xmath8 allow us to find also peaks in amplitudes that we associate to the @xmath105 ( @xmath106 ) and @xmath107 ( @xmath108 ) states .
the agreement found with experiment , at the cost of no free parameters , is excellent , as one can see in fig .
[ fig : mvsn ]
. it would be interesting to apply these ideas to other systems where one of the @xmath8 mesons is replaced by a @xmath120 in order to see if one obtains the several high spin @xmath120 stated of @xcite .
work along this direction has just appeared @xcite and the predictions are equally successful .
mesons , @xmath121 . only single scattering contribution ( dotted line ) ; full model ( solid line ) ; experimental values from the pdg@xcite , ( circles ) .
, scaledwidth=70.0% ]
we conclude the discussion by emphasizing that all the low lying @xmath122 @xmath0 and @xmath1 resonances in the @xmath123 @xcite , up to the 1800 mev energy region , get dynamically generated as two meson - one baryon states in these calculations .
in addition , we predict the quantum numbers of the @xmath19 and also find evidence for a @xmath122 @xmath0 resonance at @xmath84 1620 mev . on the other hand we also reported on a hidden strangeness three body system , which leads to a new @xmath5 state with mass around 1920 mev and which we think is responsible for the peak seen in the @xmath7 reaction .
we suggest to perform experiments on polarization and also the study of the @xmath99 reaction close to threshold in order to test the predictions tied to the existence of this resonance below the @xmath124 threshold .
finally we reported on very novel few body systems , states made of many @xmath8 mesons with their spins aligned .
we could find that these states are bound , although they have a width from decay to pions .
yet , up to j=6 , the widths are still on the measurable range such that we could associate the states found to several existing mesons with spin up to j=6 . as a general remark
, the combination of the chiral unitary techniques with few body systems is proving to be a very fruitful field to which we can only encourage our colleagues to jump to .
this work is partly supported by dgicyt contract number fis2006 - 03438 and the generalitat valenciana in the program prometeo .
we acknowledge the support of the european community - research infrastructure integrating activity study of strongly interacting matter ( acronym hadronphysics2 , grant agreement n. 227431 ) under the seventh framework programme of eu .
j. nieves and e. ruiz arriola , nucl .
a * 679 * , 57 ( 2000 ) d. jido _ et .
phys . a * 725 * , 181 ( 2003 ) c. garcia - recio , m. f. m. lutz and j. nieves , phys .
b * 582 * , 49 ( 2004 ) t. hyodo , s. i. nam , d. jido and a. hosaka , phys .
c * 68 * , 018201 ( 2003 ) b. borasoy , r. nissler and w. weise , eur .
j. a * 25 * , 79 ( 2005 ) l. y. glozman and d. o. riska , phys .
rept . * 268 * , 263 ( 1996 ) .
e. epelbaum _ et .
* 86 * , 4787 ( 2001 ) a. martinez torres , k. p. khemchandani and e. oset , eur .
j. a * 36 * , 211 ( 2008 ) [ arxiv:0712.1938 [ nucl - th ] ] .
a. martinez torres , k. p. khemchandani , l. s. geng , m. napsuciale and e. oset , phys .
d * 78 * , 074031 ( 2008 ) [ arxiv:0801.3635 [ nucl - th ] ] .
d. jido and y. kanada - enyo , phys .
c * 78 * , 035203 ( 2008 ) a. martinez torres , k. p. khemchandani , u. g. meissner and e. oset , eur .
j. a * 41 * , 361 ( 2009 ) k. h. glander _ et al . _ , eur .
phys . j. a * 19 * , 251 ( 2004 ) r. bradford _ et al . _
[ clas collaboration ] , phys .
c * 73 * , 035202 ( 2006 ) m. sumihama _ et al .
_ [ leps collaboration ] , phys . rev .
c * 73 * , 035214 ( 2006 ) m.
bando , t. kugo and k. yamawaki , phys .
rept . * 164 * ( 1988 ) 217 .
m. harada and k. yamawaki , phys .
rept . * 381 * , 1 ( 2003 ) .
u. g. meissner , phys .
rept . *
161 * , 213 ( 1988 ) .
h. nagahiro , j. yamagata - sekihara , e. oset and s. hirenzaki , phys .
d * 79 * , 114023 ( 2009 ) .
l. s. geng and e. oset , phys .
d * 79 * , 074009 ( 2009 ) .
a. martinez torres , l. s. geng , l. r. dai , b. x. sun , e. oset and b. s. zou , phys .
b * 680 * , 310 ( 2009 ) . | we report on four @xmath0 s and three @xmath1 s , in the 1500 - 1800 mev region , as two meson - one baryon s - wave @xmath2 resonances found by solving the faddeev equations in the coupled channel approach , which can be associated to the existing @xmath3 = -1 , @xmath4 low lying baryon resonances . on the other hand we also report on a new , hidden strangeness @xmath5 state , mostly made of @xmath6 , with mass around 1920 mev , which we think could be responsible for the peak seen in the @xmath7 around this energy .
finally we address a very novel topic in which we show how few body systems of several @xmath8 mesons can be produced , with their spins aligned up to j=6 , and how these states found theoretically can be associated to several known mesons with spins j=2,3,4,5,6 .
example.eps gsave newpath 20 20 moveto 20 220 lineto 220 220 lineto 220 20 lineto closepath 2 setlinewidth gsave .4 setgray fill grestore stroke grestore |
let @xmath0 be a real connected semisimple lie group with no compact factors , @xmath1 a maximal compact subgroup of @xmath0 , and @xmath2 a discrete subgroup of @xmath0 .
a general definition of automorphic forms for @xmath2 ( cf @xcite and @xcite ) depends on a finite dimensional representation of @xmath1 .
it includes classical holomorphic automorphic forms on fuchsian groups , maass wave forms , and automorphic forms on bounded symmetric domains ( these terms are defined in [ auto def ] ) .
a convenient construction of automorphic forms by means of relative poincar series introduced in @xcite and @xcite , allows to `` extend '' an automorphic form for a subgroup @xmath6 to the whole discrete group @xmath2 .
for @xmath7 we obtain poincar series , which can be constructed for any smooth absolutely integrable function on @xmath0 ( @xcite 5 ) . for cocompact @xmath2 and in some other cases which are discussed in
[ relps ] , ( relative ) poincar series provide cusp forms
. it may happen that a poincar series is identically zero .
however , it has been shown ( see @xcite ch . 3 and @xcite ch .
5 ) that for holomorphic automorphic forms on a symmetric bounded domain there exist non - zero poincar series for suitable weights .
it is not easy , however , to construct relative poincar series for non - trivial @xmath8 , and even less so to prove that a certain collection spans the corresponding space of cusp forms . in the case when @xmath0 is of real rank one and with finite center , and @xmath2 is a lattice in @xmath0 , i.e. @xmath3 has finite haar measure , we propose a program of constructing spanning sets for cusp forms based on dynamical properties of the homogeneous flow on @xmath3 and a livshitz type theorem that we prove for such a flow ( theorem [ livshitz ] ) .
we carry out this program in the case when , in addition , @xmath1 has non - trivial 1-dimensional representations . in this case the symmetric space @xmath9 is hermitian of real rank one , and by the classification of symmetric spaces , it is a complex hyperbolic space @xmath10 for some @xmath11 .
let @xmath2 be a lattice in @xmath12 , and @xmath13 be a half
integer such that @xmath14 is an integer . for each loxodromic element @xmath15
we construct a relative poincar series @xmath16 which is a @xmath17valued cusp form in @xmath18 ( for exact definitions see [ auto def ] , [ relps ] and [ pseries ] ) .
the main result of this paper is the following theorem .
[ main]the relative poincar series @xmath19 span @xmath18 .
this is an extension of a result by s. katok @xcite for the classical case of fuchsian groups ( @xmath20 ) .
the method is based on duality between the introduced relative poincar series and closed geodesics on @xmath21 ( theorems [ integral ] and [ lift int ] ) , dynamical properties of the frame flow on @xmath22 which imply a livshitz theorem type result ( theorem [ livshitz ] ) , and the existence of a three dimensional lie subalgebra associated to the frame flow which enables the harmonic analysis on @xmath22 that completes the proof .
the proof utilizes two essential features of the complex hyperbolic space : real rank one and the presence of the complex structure . a natural question of finding
finite spanning sets for cusp forms has been addressed for the classical @xmath20 case in @xcite .
the main ingredient here was an approximate version of livshitz theorem @xcite for geodesic flows , based on more subtle dynamical considerations .
the same scheme will probably work for complex hyperbolic spaces of all dimensions and will be pursued elsewhere .
the paper is organized as follows . in [ autoforms ]
we review the general definition of automorphic forms and the relative poincar series , discuss their convergence and further implications under additional assumptions .
[ hom flow ] describes the homogeneous flow on @xmath3 and its relation to the geodesic flow on the unit tangent bundle @xmath23 . in [ dyn ]
we prove a special livshitz theorem for the homogeneous flow on @xmath3 and describe a program of constructing spanning sets for cusp forms in the real rank one case . in
48 we restrict ourselves to the case of the complex hyperbolic space , construct relative poincar series associated to closed geodesics in @xmath5 , and prove theorem [ main ] .
we want to express our gratitude to d. akhiezer for his suggestion to consider a three dimensional subalgebra associated to the homogeneous flow , which was crucial in the completion of this work .
we also would like to thank the erwin schroedinger institute for mathematical physics in vienna , where the second author spent three months working on this paper in 1997 and 1998 , for its hospitality and financial support . and finally
, we would like to thank the referee for the comments which helped us to clarify the exposition .
the research has begun when the second author was partially supported by nsf grant dms-9404136 .
let @xmath0 be a real connected semisimple lie group with no compact factors ( @xcite 1.13.12 ) , @xmath24 its lie algebra , and @xmath1 a maximal compact subgroup of @xmath0 .
we shall denote the homogeneous space by @xmath25 .
the group @xmath0 acts on itself by left multiplications , and this action projects to the action on @xmath26 .
let @xmath27 be the fixed point of @xmath1 in @xmath26 , then the natural projection @xmath28 is given by @xmath29 the cartan decomposition corresponding to @xmath1 is @xmath30 where @xmath31 is the lie algebra of @xmath1 , @xmath32 , and @xmath33 is the orthogonal complement to @xmath31 with respect to the killing form on @xmath34 .
the differential @xmath35 at the identity of @xmath0 identifies @xmath33 with @xmath36 .
the kernel of @xmath35 is @xmath31 ( @xcite p. 208 ) . to each @xmath37
is associated a left - invariant differential operator also denoted by @xmath38 , @xmath39 this linear map from @xmath24 into the algebra @xmath40 of left - invariant differential operators on @xmath0 extends to an isomorphism of @xmath41 , the universal enveloping algebra of the lie algebra @xmath24 ( with complex coefficients ) onto @xmath40 . in particular
, one may consider differential operators @xmath38 for @xmath42 , the complexification of @xmath24 .
on @xmath43 the center @xmath44 of @xmath41 corresponds to the left and right invariant operators and is isomorphic to a polynomial ring in @xmath45 letters where @xmath45 is the real rank of @xmath0 .
the general definition of _ automorphic forms _ in the sense of harish - chandra @xcite and borel @xcite assumes only that @xmath2 is a discrete subgroup of @xmath0 . in this paper
we study _ cusp forms _ ,
which are automorphic forms with some additional hypotheses imposed , in case when @xmath2 is not cocompact , concerning the behavior close to certain boundary points .
naturally , if @xmath2 is cocompact , every automorphic form is a cusp form .
let @xmath2 be a lattice in @xmath0 , and @xmath46 be a finite dimensional complex vector space .
in what follows @xmath47 will act on @xmath46 on the right , and @xmath48 be a fixed ( anti)representation of @xmath1 in @xmath47 .
let @xmath49 be a hermitian @xmath50invariant inner product on @xmath46 and @xmath51 the corresponding norm .
we define the norm @xmath52 ( @xmath53 ) on measurable @xmath46valued @xmath2invariant on the left functions on @xmath0 as usual by @xmath54 where @xmath55 is the haar measure on the group @xmath0 , and the norm @xmath56 as the essential supremum of @xmath57 on @xmath0 , and let @xmath58 ( @xmath59 ) be the space of such @xmath60 for which @xmath61 . a vector valued function @xmath62 is called _
@xmath44finite _ if @xmath63 is annihilated by an ideal @xmath64 of @xmath44 of finite codimension .
if @xmath64 has codimension 1 , this means that @xmath60 is an eigenfunction of every operator in @xmath44 .
[ aut - form]a smooth vector valued function @xmath62 is called a _ cusp form for @xmath2 _ if ( [ aut - form].1 ) @xmath60 is @xmath2invariant on the left and @xmath1equivariant on the right , i.e. @xmath65 for any @xmath66 , @xmath67 and @xmath68 ; ( [ aut - form].2 ) @xmath60 is @xmath44finite ; ( [ aut - form].3 ) @xmath69 .
\1 . condition ( [ aut - form].1 ) and ( [ aut - form].2 ) imply that @xmath60 is a real analytic function on @xmath3 @xcite , hence a function @xmath60 which satisfies ( [ aut - form].1 ) and ( [ aut - form].2 ) satisfies ( [ aut - form].3 ) if and only if @xmath57 is bounded on @xmath3 .
if @xmath2 is cocompact , the condition ( [ aut - form].3 ) is automatic .
\3 . if @xmath70 is trivial , @xmath60 is @xmath1invariant on the right and is often referred to as a cusp form of _ weight @xmath27_. in case @xmath71 such cusp forms are called _ maass wave forms_. they are eigenfunctions of the non euclidean laplacian .
automorphic forms may be defined using an _ automorphy factor _ ,
i.e. a smooth map @xmath72 that satisfies the @xmath73cocycle property , i.e. @xmath74 for all @xmath75 and @xmath76 . if a smooth function @xmath77 satisfies the equation @xmath78 for any @xmath66 and @xmath76 , then its `` lift '' to @xmath0 , @xmath79 defined by @xmath80 satisfies ( [ aut - form].1 ) with @xmath81 .
let @xmath82 be an automorphy factor .
a smooth function @xmath77 is called a _ cusp form of type @xmath82 _ if its lift to the group @xmath0 defined by ( [ lift ] ) is a cusp form of the definition [ aut - form ] with @xmath81 .
we denote the space of cusp forms of type @xmath82 ( as well as the space of their lifts to @xmath0 ) by @xmath83 .
the space @xmath84 is a hilbert space with the inner product @xmath85 corresponding to the norm @xmath86 introduced above ( inside the integral is the inner product on @xmath46 ) . using the cartan decomposition of the group @xmath0 corresponding to ( [ cd ] ) , @xmath87 where @xmath68 and @xmath88 depends only on @xmath89 , and the cocycle property of @xmath82 ( [ cocycle ] ) , we obtain for the lifts of automorphic forms
@xmath90 and @xmath91 with @xmath92 , an analogue of the petersson inner product @xmath93 where @xmath94 is a @xmath0invariant volume form on the symmetric space @xmath26
. since @xmath2 is a lattice , we have @xmath95 therefore the integral ( [ ip ] ) converges for cusp forms ; in fact , it converges if either @xmath96 or @xmath97 is a cusp form .
thus @xmath98 .
notice that in general a function satisfying ( [ aut - form].1 ) and ( [ aut - form].2 ) and belonging to @xmath99 is not necessarily a cusp form ( e.g. for @xmath100 and maass wave forms , a residual eisenstein series has behavior like @xmath101 , for @xmath102 in the cusp ; it satisfies ( [ aut - form].1 ) and ( [ aut - form].2 ) , is in @xmath103 but is not bounded ) , although for holomorphic forms on a symmetric bounded domain it is ( see remark 1 below ) .
let @xmath26 be a symmetric bounded domain .
if @xmath104 is a holomorphic function on @xmath26 , the condition ( [ aut - form].2 ) is satisfied @xcite .
let @xmath105 be the subspace of @xmath58 consisting of holomorphic functions on @xmath26 which satisfy ( [ type - mu ] ) .
then @xmath106 . \1 . for holomorphic automorphic forms on
a symmetric bounded domain @xmath26 all spaces @xmath107 ( @xmath59 ) coincide for suitable @xmath82 by satake s theorem ( @xcite exp .
9 ( supp . ) , @xcite ch .
11 , 5 ) .
although in general , @xmath108 , for holomorphic automorphic forms on a symmetric bounded domain @xmath26 , @xmath109 ( @xcite ch . 4 , @xcite , @xcite ch .
11 , 5 ) . \3 .
the classical case of holomorphic automorphic forms on fuchsian groups corresponds to @xmath110 .
let @xmath8 be a subgroup of @xmath2 .
the following construction allows to `` extend '' an automorphic form for @xmath8 to an automorphic form for @xmath2 .
[ rps]let @xmath8 be a subgroup of @xmath2 , @xmath111 be an automorphic form of type @xmath82 for @xmath8 , and @xmath112 be its lift to the group @xmath0 by the formula ( [ lift ] ) such that 1 .
@xmath112 is @xmath44finite , 2 .
@xmath113 , then the series @xmath114 , called the _ relative poincar series for _ @xmath8 , converges absolutely and uniformly on compact sets , and represents a function satisfying ( [ aut - form].1 ) , ( [ aut - form].2 ) , and belonging to @xmath99 .
the proof follows the lines of the argument of harish
chandra for the poincar series ( see @xcite or @xcite ) .
it is proved that the series converges absolutely and uniformly on compact sets and satisfies ( [ aut - form].1 ) and ( [ aut - form].2 ) . for cocompact @xmath2
this proves that @xmath115 is a cusp form .
it @xmath2 is a lattice , it follows that @xmath116 which , without additional assumptions does not imply that @xmath115 is a cusp form .
however , for holomorphic automorphic forms on a symmetric bounded domain , according to satake s theorem , the relative poincar series @xmath115 are cusp forms .
if @xmath25 is a symmetric bounded domain , then any absolutely integrable function on @xmath26 , e.g. any polynomial on @xmath26 , produces a holomorphic cusp form for the trivial @xmath7 ( @xcite ch .
11 , 1 ) .
if @xmath8 is a parabolic subgroup of @xmath2 , the above construction gives so called poincar eisenstein series . for general @xmath8 ,
it is not easy to find a function @xmath104 which satisfies conditions 1 and 2 of theorem [ rps ] .
we shall give a construction for cyclic loxodromic subgroups of isometries of complex hyperbolic spaces in [ pseries ] .
from now on we suppose , in addition to the standing assumptions of [ auto def ] , that @xmath0 is of real rank one and with finite center ( and therefore non compact simple ) . then any non zero @xmath117 defines a maximal abelian subspace of @xmath33 , @xmath118 , consisting of real semisimple elements ( cf .
@xcite pp .
401 , 431 ) .
we fix a non zero @xmath117 and let @xmath119 be the corresponding maximal split abelian subgroup of @xmath0
. acting by right multiplications on @xmath0 , it defines a standard flow @xmath120 on @xmath0 , @xmath121 an orbit of @xmath67 , @xmath122 projects by ( [ proj ] ) to a geodesic @xmath123 on @xmath26 passing through @xmath124 ( @xcite p. 208 ) .
the subgroup @xmath125 itself projects on @xmath26 to the `` standard geodesic '' @xmath126 passing through @xmath27 .
there is a subgroup @xmath127 which , acting on @xmath26 on the left by isometries , fixes @xmath128 pointwise .
then @xmath129 , the centralizer of @xmath125 .
an element @xmath67 is called _ hyperbolic _ ( or _ regular _ )
if it is conjugate to an element in @xmath125 .
an element @xmath67 is called _ loxodromic _ if it is conjugate to an element in @xmath130 .
any loxodromic ( and , therefore , hyperbolic ) element in @xmath0 fixes a geodesic in @xmath26 , called its _ axis _ and has two fixed points in the boundary @xmath131 .
it is easy to see that geodesics in @xmath26 are exactly the axes of loxodromic elements in @xmath0 .
geodesics naturally lift to @xmath132 which can be identified with the unit tangent bundle @xmath133 . recall that as subgroups of @xmath0 acting on the space @xmath25 on the left by isometries , @xmath1 fixes the point @xmath134 , and @xmath127 fixes @xmath128 pointwise , and hence the unit tangent vector to @xmath128 at @xmath27 denoted by @xmath135 .
thus we have two natural mappings : @xmath136 given by ( [ proj ] ) and @xmath137 given by @xmath138 where @xmath139 is the differential of @xmath140 . defining the mapping @xmath141 such that @xmath142 we see that @xmath132 is a sphere bundle over @xmath9 with the fiber at each point identified with @xmath143 , the unit tangent space at this point . for any geodesic @xmath144 in @xmath26 passing through a point @xmath145
there exists a @xmath67 mapping @xmath128 into @xmath144 in such a way that @xmath146 .
this follows from the fact that the group @xmath0 acts transitively on the unit tangent bundle @xmath147 .
the transformation @xmath140 is not unique but is determined up to the right multiplication by the subgroup @xmath148 which fixes every point of @xmath128 .
thus we obtain a family of lifts of @xmath144 , @xmath149 to the group @xmath0 parametrized by the group @xmath148 , which are orbits of the flow @xmath150 . since the subgroups @xmath125 and @xmath148 commute , the flow @xmath150 projects to the factor @xmath132 to the _ geodesic flow _ denoted by @xmath151 : @xmath152 we see that all lifts @xmath149 project to the orbit @xmath153 of the geodesic flow on @xmath132 .
let @xmath2 be a lattice in @xmath0 and @xmath154 .
the action of @xmath125 descends to the factor @xmath3 and defines a _ homogeneous flow _
@xmath155 on @xmath22 denoted by @xmath155 , @xmath156 the mappings @xmath157 , @xmath158 and @xmath159 also descend to the corresponding left factors by @xmath2 , and the homogeneous flow @xmath155 projects to the _ geodesic flow _ on @xmath160 , denoted by @xmath161 , by the formula : @xmath162 for the base point @xmath163 we shall use the same notation : @xmath164 respectively , we have two differential operators : @xmath165 defined on the set of functions on @xmath3 differentiable along the orbits of @xmath155 , and @xmath166 defined on the set of functions on @xmath167 differentiable along the orbits of @xmath161 .
the following characterization of closed geodesics in @xmath168 is immediate . closed geodesics in @xmath169 are the axes of loxodromic elements in @xmath2 .
they are in one to finite correspondence with conjugacy classes of primitive loxodromic elements in @xmath2 .
it is easy to see that closed geodesics in @xmath169 lift to closed orbits of the geodesic flow @xmath161 on @xmath167 while their lifts to @xmath3 are not necessarily closed .
we shall denote a closed geodesic in @xmath169 corresponding to the axis of a loxodromic element @xmath15 in @xmath5 as well as its lift to @xmath167 by @xmath170 $ ] , and a family of lifts to @xmath3 by @xmath171_w\mid w\in w\}$ ] .
it is a standard fact that there exists a left @xmath0invariant riemannian metric on @xmath25 relative to which @xmath26 is a symmetric space ( @xcite ch .
11 , th . 8.6 ) . since the real rank of @xmath0 is equal to one , this
metric is of negative sectional curvature , which is the main reason of the following properties of the homogeneous flow @xmath155 and the geodesic flow @xmath161 which we review below . a standard procedure ( @xcite ch . 11
, 6 ) provides @xmath0invariant riemannian metrics on @xmath0 and @xmath132 ( this can be done for any compact subgroup of @xmath0 in place of @xmath148 ) in such a way that the distance ( which we will denote in both spaces by @xmath172 ) does not increase after projection : for any @xmath173 , @xmath174 the riemannian volume on @xmath0 coincides with the haar measure . we have seen that the homogeneous flow @xmath155 on @xmath3 , considered as a fibered bundle over @xmath167 , projects to the geodesic flow @xmath161 on the base @xmath167 .
in addition , if @xmath175 and @xmath176 belong to the same fiber , i.e. @xmath177 for some @xmath178 , we have @xmath179 by left
invariance of the metric and since @xmath180 and @xmath181 commute .
the geodesic flow @xmath161 on @xmath167 is _ anosov ( hyperbolic ) _
( @xcite th .
17.6.2 ) with the corresponding @xmath182invariant splitting of the tangent bundle of @xmath167 : @xmath183 and foliations @xmath184 ( the orbit foliation ) , @xmath185 , and @xmath186 , respectively .
it follows that the homogeneous flow @xmath155 is an isometric extension of @xmath161 and , as such , is _ partially hyperbolic _
, i.e. there is a @xmath187 @xmath188invariant splitting of the tangent bundle of @xmath3 : @xmath189 with the following properties .
the integral manifolds of the distribution @xmath190 form the _ neutral _ foliation denoted by @xmath191 ; its leaf through a point @xmath140 being @xmath192 , where @xmath193 is the orbit of the flow .
the flow restricted to the leaves of @xmath191 is isometric .
the integral manifolds of the distributions @xmath194 form the _ stable ( unstable ) _ foliation denoted by @xmath195 .
let us denote the distance along the leaves of the foliations @xmath196 and @xmath197 by @xmath198 and @xmath199 , respectively
. then there exist @xmath200 such that for any @xmath201 lying on the same leaf of @xmath202 ( @xmath203 ) , @xmath204 for @xmath205 and for @xmath206 .
the foliations @xmath207 are transversal and project to the corresponding foliations @xmath184 , @xmath185 , and @xmath186 of @xmath161 with the same estimates ( [ hyper ] ) .
the homogeneous flow @xmath155 preserves the haar measure on @xmath3 while @xmath161 preserves the riemannian volume on @xmath167 .
the hyperbolic properties of the geodesic flow on @xmath167 imply the anosov closing lemma and its strengthening ( see @xcite th .
6.4.15 and prop . 6.4.16 ) . as a consequence ,
the livshitz theorem holds .
since the space @xmath208 may not be compact , it is natural to formulate it for the class @xmath209 of bounded lipschitz functions .
let @xmath210 be such that it has zero integrals over all closed orbits of the geodesic flow on @xmath211 .
then there exists a lipschitz function @xmath60 on @xmath208 differentiable in the direction of the flow @xmath161 , and such that @xmath212 this theorem has been proved in @xcite for anosov flows on compact manifolds .
it has been proved later @xcite that if the function @xmath104 is @xmath213 , the solution is also @xmath213 .
the original proof works with minor alterations for manifolds with cusps ( the proof for fuchsian groups with cusps is given in @xcite appendix ) . by the moore s ergodicity theorem ( @xcite th .
2.2.6 ) the homogeneous flow on @xmath3 is ergodic and hence topologically transitive , but it is not anosov , the closing lemma does not hold , and hence a straightforward generalization of livshitz theorem does not hold either . however , the same conclusion holds under a stronger hypothesis , and the proof is similar to the proof of the livshitz theorem for the geodesic flow .
[ livshitz]let @xmath214 be such that for every closed geodesic @xmath170 $ ] in @xmath208 and every @xmath178 the integral in @xmath3 @xmath215_w}f({{\tilde\varphi}}_s(g))ds=0.\ ] ] then there exists a lipschitz function @xmath60 on @xmath3 constant on @xmath148cosets and differentiable in the direction of the flow @xmath155 and such that @xmath216 we shall give a proof for real valued functions .
then the theorem will obviously hold for functions valued in any @xmath217 .
the homogeneous flow on @xmath3 is topologically transitive , i.e. there exists a @xmath67 whose orbit @xmath218 is dense in @xmath3 .
we define a function @xmath60 on this orbit by the formula @xmath219 we need to prove that @xmath60 satisfies a lipschitz condition on @xmath220 and hence can be extended to @xmath3 as a lipschitz function .
given @xmath221 , let @xmath222 be such that @xmath223 we will show that @xmath224 as it is customary , `` @xmath225 '' means that the expression is `` @xmath226 '' for some constant @xmath144 .
we will use this notation to avoid an accumulation of constants . since the homogeneous flow is not anosov
, we can not hope to be able to approximate any @xmath227close piece of the dense orbit by a closed one as for the geodesic flow .
nevertheless , the close relation with the geodesic flow allows us to prove the following using the notations above , for a piece of a dense orbit of @xmath155 ( [ eps ] ) , there is a lift of a closed orbit of @xmath161 to @xmath3 , @xmath228 with @xmath229 , such that for @xmath230 @xmath231 by ( [ ineq ] ) the projection of @xmath232 onto @xmath167 is a dense orbit @xmath233 of the geodesic flow @xmath161 with @xmath234 , and it follows that @xmath235 let @xmath236 and @xmath237 be the corresponding pieces of orbits .
let us denote @xmath238 @xmath239 , @xmath240 and @xmath241 .
since the geodesic flow @xmath161 is anosov , by anosov closing lemma there exists a @xmath242 such that @xmath243 and its orbit @xmath244 is closed , i.e. @xmath245 with the period @xmath246 satisfying @xmath229 .
the point @xmath247 can be chosen such that the leaf of the unstable foliation containing @xmath248 , @xmath249 , and a leaf of the stable foliation containing @xmath247 , @xmath250 intersect , and since they are transversal , they intersect in a point @xmath251 , and @xmath252 .
lift @xmath249 to @xmath253 .
since @xmath254 its lift @xmath253 intersects @xmath255 , so let @xmath256 , and similarly , @xmath257 intersects @xmath258 , @xmath259 , and @xmath260 ( see fig . 1 which represents the picture in the direction transversal to the orbits ) . since the distance between @xmath261 and @xmath262 decreases exponentially for @xmath263 , and @xmath264 , we have @xmath265 and since all leaves are transversal we have the same estimate on @xmath257 ; @xmath266 similarly , since the distance between @xmath267 and @xmath268 decreases exponentially for @xmath263 , and @xmath269 , we have @xmath270 and the same estimate on @xmath257 : @xmath271 therefore , we have simultaneous estimates @xmath272 and @xmath273 @xmath230 .
it follows from ( [ exp ] ) that @xmath274 and since by the hypothesis @xmath275 we obtain the required estimate ( [ lip ] ) .
this proves the claim .
thus @xmath60 can be extended from the dense orbit to a lipschitz function in @xmath3 .
since @xmath276 on the dense orbit , it follows that @xmath60 is differentiable in the direction of the homogeneous flow and @xmath276 in @xmath3 .
a similar argument shows that the function @xmath60 is constant on @xmath148cosets . for
, let @xmath277 .
there are @xmath278 and @xmath279 on the dense orbit of @xmath155 @xmath227close to @xmath175 and @xmath176 respectively .
then the projection to @xmath167 is an @xmath227close orbit of @xmath161 .
find a closed orbit of @xmath161 @xmath227close to the projection , and lift it back to @xmath3 in the same manner we described before .
it follows from the exponential estimates that @xmath280 and as @xmath281 , we obtain @xmath282 .
our program of constructing spanning sets for @xmath283 consists of three steps .
\(1 ) _ construction of relative poincar series associated to closed geodesics in @xmath284_. fix a finite dimensional complex vector space @xmath46 and an automorphy factor @xmath72 . for each loxodromic element
@xmath285 find a function @xmath286 which satisfied conditions of theorem [ rps ] for the subgroup @xmath287 .
having such a function , we could apply theorem [ rps ] to obtain a relative poincar series @xmath288 , and prove that they are cusp forms .
\(2 ) _ the period formula_.
let @xmath289 , @xmath285 loxodromic , and @xmath290 be the lift of @xmath104 to @xmath0 .
then for any lift @xmath170_w,\,\,w\in w$ ] of closed geodesic @xmath170 $ ] to @xmath0 , @xmath291_w}{{\tilde f}}dt\ ] ] with an explicit constant @xmath144 depending only on @xmath170_w$ ] .
\(3 ) _ cohomological equation for cusp forms_. in order to prove that the relative poincar series @xmath292 span @xmath83 , we assume that there is a cusp form @xmath293 such that @xmath294 for all relative poincar series @xmath288 . in order to apply the special livshitz theorem ( theorem [ livshitz ] ) we first need to show that @xmath295 and then use ( 2 ) .
thus we obtain a lipschitz and differentiable in the direction of the flow @xmath155 solution @xmath60 of the cohomological equation @xmath296 , which means that @xmath290 is a _
coboundary_. proving that cusp forms can not be coboundaries will imply the result .
we are able to carry out this program in the case when @xmath1 has non trivial @xmath73dimensional representations . in this case
the symmetric space is hermitian of real rank one , and by the classification of symmetric spaces ( h p. 518 ) , it is a complex hyperbolic space @xmath10 for some @xmath11 . the rest of the paper is devoted to this case .
the group @xmath4 is the group of @xmath297 complex unimodular matrices preserving the hermitian form @xmath298 on @xmath299 . in other words ,
@xmath300 @xmath301 where @xmath302 denotes the @xmath303 identity matrix , @xmath304 and @xmath305 is used for matrix multiplication .
the maximal compact subgroup of @xmath0 @xmath306 the symmetric space @xmath9 is called the _ complex hyperbolic space _ and is denoted by @xmath307 .
it can be identified with the projectivised space of _ negative _ vectors @xmath308 , i.e. such that @xmath309 : @xmath310 or , equivalently , with the set of _ negative _ lines in @xmath299 , or in homogeneous coordinates , with the unit ball in @xmath311 : @xmath312 the last identification is obtained by the biholomorphic embedding @xmath313 @xmath314 of @xmath311 onto the affine chart of @xmath315 defined by @xmath316 .
@xmath317 is a bounded domain in @xmath311 and hence a complex manifold .
an invariant riemannian metric on @xmath317 of the form @xmath318 obtained by a standard construction ( @xcite ch . 9 , 6 ) is of negative sectional curvature and is called the _
bergman metric_. we normalize it according to @xcite ( iii.1.3 ) so that the sectional curvature is pinched between @xmath319 and @xmath320 . using @xcite ch .
2.5 , one easily obtains the following relation between the volume form @xmath94 on @xmath317 corresponding to the riemannian metric and the euclidean volume form @xmath321 on @xmath311 : @xmath322 using the above terminology , we call a vector @xmath323 _ positive _ if @xmath324 and a _ null vector _ if @xmath325
. we shall always assume that negative and null vectors are represented in the form ( [ emb ] ) , and keep the notation @xmath326 for them .
for each point @xmath327 we write @xmath328 . _ the real tangent space at @xmath329 _ , @xmath330 refers to the tangent space of the underlying @xmath331-dimensional real @xmath187 manifold and has a basis @xmath332 the vector fields @xmath333 and @xmath334 form a basis of the complexification @xmath335 of @xmath336 .
the real tangent space @xmath336 can be identified with the @xmath337dimensional complex subspace of @xmath335 of the vectors in the form @xmath338 , where @xmath339 .
we shall always refer to the tangent vectors being in the form @xmath340 ( @xmath341 .
the group @xmath4 acts on @xmath317 by biholomorphic transformations ( automorphisms ) : for @xmath342 @xmath343 ( sometimes we will also use @xmath344 for this action ) , which are isometries of @xmath317 with respect to the bergman metric .
this action corresponds to the left multiplication on @xmath0 .
the differential of @xmath345 at @xmath329 , @xmath346 is given by the jacobian matrix which can be written as @xmath347 where @xmath348 , @xmath349 , @xmath350 , @xmath351 .
a direct calculation shows that @xmath352 the group of biholomorphic automorphisms of @xmath317 is actually @xmath353 where @xmath354 is the center of @xmath0 ( @xmath227 is the primitive @xmath355 root of @xmath73 ) .
the lie algebra of the group @xmath0 is @xmath356 .
the cartan decomposition ( [ cd ] ) is @xmath357 @xmath31 is the lie algebra of @xmath1 consisting of @xmath358 complex matrices of the form @xmath359 where @xmath360 is an @xmath303 skew hermitian matrix , @xmath361 , and @xmath362 , and @xmath33 consists of matrices @xmath363 where @xmath364 .
the differential at the identity @xmath35 of the natural mapping @xmath365 maps @xmath366 to the tangent vector @xmath367 thus identifying @xmath368 with @xmath369 . using the cartan decomposition of the group @xmath0 ( [ cdg ] ) , we can write any element @xmath370 as a product @xmath371 .
the matrix @xmath68 is in the form @xmath372 , where @xmath373 and @xmath374 ; the matrix @xmath375 is hyperbolic whose entries depend only on @xmath376 , and represents the projection on the symmetric space @xmath377 , and by @xcite iii.2.7 , @xmath378 recall ( @xcite ch .
9 ) that the canonical complex structure on @xmath317 is a tensor field @xmath379 , where for each @xmath380 , @xmath381 is an automorphism of the tangent space @xmath336 such that @xmath382 ( @xmath64 denotes the identity map of @xmath336 ) . in local
coordinates it is given by @xmath383 restricting @xmath384 to @xmath385 we obtain an automorphism @xmath386 such that @xmath387 for all @xmath388 . since @xmath0 is semisimple , there exists an element @xmath389 such that @xmath390 and @xmath391=0\}$ ] ( @xcite ch .
we have @xmath392 and @xmath393 for all @xmath394 . for each @xmath395 the automorphism @xmath381
can be extended uniquely to a complex linear mapping of the complexification @xmath396 of @xmath336 onto itself also denoted by @xmath381 , and satisfying @xmath397 .
the eigenvalues of @xmath381 are therefore @xmath398 and @xmath399 .
for @xmath400 we have @xmath401 and the corresponding eigenspaces are @xmath402 then @xmath403 and @xmath404 can be written as a direct sum of complex vector spaces @xmath405 thus by ( [ 1,0 - 0,1 ] ) and ( [ mathfrak p ] ) @xmath406 and @xmath407 any @xmath67 can be also decomposed according to harish - chandra ( @xcite p. 688
, @xcite ch .
viii , 7 ) as @xmath408 where @xmath409 , @xmath410 , @xmath411 , @xmath412 , and @xmath413 , the complexification of @xmath1 .
for @xmath414 we obtain @xmath415 where @xmath416 and @xmath417 .
since @xmath376 we see that the harish - chandra map @xmath418 given by @xmath419 is exactly the map @xmath157 of ( [ proj - b ] ) and the bounded domain @xmath420 is the unit ball @xmath421 .
notice that in ( [ g+g- ] ) @xmath422 .
the bergman metric on @xmath317 introduced in [ model ] is hermitian , i.e. invariant by the complex structure @xmath384 , and thus defines a hermitian inner product on each tangent space @xmath423 .
let @xmath2 be a lattice in @xmath0 and @xmath424 .
the corresponding maximal split abelian subgroup of @xmath0 @xmath425 projects to the `` standard geodesic '' @xmath426 the subgroup of @xmath1 which commutes with @xmath125 is @xmath427 as has been explained in
[ flow - def ] , acting by right multiplications on @xmath0 , @xmath125 defines a homogeneous flow @xmath155 on @xmath3 , which is called the _ frame flow _ in this case .
the terminology comes from the fact that the group @xmath428 may be identified with a principal @xmath429bundle over @xmath317 . a fiber at the point @xmath395 geometrically represents the space of @xmath337frames : unitary frames of tangent vectors at @xmath329 with respect to the hermitian inner product .
geometrically , the flow @xmath155 acts as follows : along the geodesic it leaves invariant , it acts by hyperbolic isometries whose differentials move the last vector along the geodesic at constant speed ( this represents the projection of @xmath155 to the geodesic flow @xmath161 on @xmath167 ) , and the remaining vectors of the frame by parallel translation ( @xcite ch .
notice that the distance with respect to the bergman metric between two points differs from the parameter @xmath430 of the geodesic flow on the corresponding geodesic by the factor of @xmath431 : @xmath432 . as has been explained in the general setup in
[ flow - def ] , a loxodromic element @xmath433 has the axis in @xmath317 , and two fixed points in the boundary @xmath434 , one _ attracting _ , and one _ repelling_. the fixed points correspond to the null eigenvectors of @xmath433 with eigenvalues @xmath435 and @xmath436 , and the remaining @xmath437 eigenvectors are positive with eigenvalues of absolute value @xmath73 .
as we have pointed out earlier , the automorphisms of @xmath317 are in fact elements of @xmath428
. therefore each loxodromic automorphism may be also represented by an element of @xmath438 with a real eigenvalue not equal to @xmath73 .
sometimes it is possible to find such an element in @xmath12 .
eigenvectors ( in @xmath439 ) and hence fixed points ( on the boundary @xmath440 ) do not depend on the choice of a representative in @xmath438 .
the following proposition gives an explicit formula for the transformation @xmath246 conjugating a loxodromic element in @xmath438 to an `` almost hyperbolic '' element with a nice fundamental domain .
this formula will be used in [ periods ] .
[ t - explicit ] let @xmath433 be a loxodromic element in @xmath438 with eigenvectors @xmath441 and corresponding eigenvalues @xmath442 , @xmath443 , @xmath444 , @xmath445 , @xmath446 .
let @xmath447 , @xmath444 , be normalized so that @xmath448 and @xmath449 be chosen so that the matrix @xmath450 has real positive determinant .
let @xmath451 be the dirichlet fundamental domain for @xmath452 centered at @xmath453 .
then \1 ) @xmath454 , \2 ) @xmath455 has null eigenvectors + @xmath456 and @xmath457 , \3 ) the fundamental domain @xmath458 of @xmath459 is bounded by two hypersurfaces : @xmath460 and @xmath461
for any @xmath67 , along with the jacobian matrix @xmath462 and its determinant , the function @xmath463 is a 1-cocycle .
taking @xmath464 for an integer @xmath13 as an _ automorphy factor _ , we see that for @xmath68 , @xmath465 is a 1dimensional representation of @xmath1 .
the operator @xmath466 defined in ( [ type - mu ] ) is well defined on @xmath428 .
thus we obtain a space of cusp forms of weight @xmath14 denoted classically by @xmath467 ( see e.g. @xcite ) . using the formula ( [ d ] )
we see that @xmath468 and therefore the inner product is given by the formula @xmath469 it is convenient to choose the following local ( partial ) coordinates @xmath470 on @xmath12 : @xmath471 where @xmath135 is the unit tangent vector at 0 to the `` standard geodesic '' @xmath128 , introduced in [ flow - def ] , and @xmath472 it is easy to check that the left multiplication by @xmath473 corresponds to the action on @xmath395 by a biholomorphic transformation @xmath474 ( [ bihol ] ) , on @xmath475 by the jacobian matrix @xmath476 ( [ jac ] ) , and on @xmath477 by @xmath478 ( [ j ] ) .
the lift of the automorphic form @xmath479 to @xmath0 ( [ lift ] ) has a nice expression in these coordinates : @xmath480 the following construction associates a cusp form to each loxodromic element @xmath481 . as a loxodromic automorphism , @xmath433 has two fixed points on the boundary @xmath434 , @xmath26 and @xmath38 : @xmath482 and @xmath483 , @xmath26 repelling and @xmath38 attracting .
then @xmath484 on the closure of @xmath485 except for the points @xmath26 and @xmath38 , and transforms under @xmath433 as follows : @xmath486 let @xmath487 . for the rest of the paper we shall assume that @xmath488 .
the function @xmath489 is an automorphic form of type @xmath490 for the subgroup @xmath287 , and it satisfies the condition 1 of theorem [ rps ] since it is holomorphic .
the condition 2 is also satisfied . to see that we write using ( [ volume form ] ) @xmath491 where @xmath451 is a dirichlet fundamental domain for @xmath8 .
since the denominator of the expression in the second integral is equal to zero only at @xmath492 and @xmath493 , the expression represents a continuous function on @xmath451 .
since @xmath451 is a bounded domain , the integral is finite .
thus for any @xmath494 we produce a relative poincar series @xmath495 of weight @xmath496 which belongs to @xmath103 , and , by satake s theorem is a cusp form .
if @xmath497 is conjugate to @xmath433 in @xmath2 , then it is easy to see that @xmath498 . if two primitive loxodromic elements @xmath499 have the same axis , then @xmath500 , where @xmath180 belongs to a compact subgroup of @xmath1 conjugate to @xmath148 .
it follows from discreteness of @xmath2 that if @xmath501 finitely many elements in @xmath2 may have the same axis ) .
it is a consequence of the period formula ( corollary [ same axis ] ) that their relative poincar series coincide .
[ integral]for any @xmath502 and any loxodromic element @xmath15 , @xmath503 here @xmath504 is any point on the axis of @xmath433 , and the integration is over the axis of @xmath433 in @xmath317 , @xmath430 is the parameter of the geodesic flow , @xmath505 , and @xmath506 is a constant .
let @xmath507 since the series ( [ theta ] ) converges absolutely , we can interchange summation and integration and , using standard `` rankin selberg method '' ( see @xcite p. 246 ) , obtain @xmath508 where @xmath451 is a fundamental domain for @xmath509 .
we now make the change of variables @xmath510 , where @xmath246 is as in proposition [ t - explicit ] , mapping the `` standard geodesic '' @xmath511 into the axis of @xmath433 .
the following lemma is checked easily .
[ lemma]@xmath512 , where @xmath505 .
according to ( [ volume form ] ) we have @xmath513 let us denote @xmath514 .
then @xmath515 further calculations can be divided onto three parts : * step 1 . *
show that @xmath516 * step 2 . * show that @xmath517 * step 3 . *
show that @xmath518 * step 1 . * @xmath519 leaves invariant @xmath520 .
the fundamental domain @xmath521 for @xmath522 over which we take the integral can be described as @xmath523 with the disc @xmath524 of radius @xmath525 `` attached '' at every point @xmath526 ( this disc has only one common point with @xmath527 ) .
change coordinates on the disc @xmath524 : @xmath528 @xmath529 then on the disc @xmath530 , @xmath531 we have @xmath532 so @xmath533 where @xmath534 is a holomorphic function of @xmath535 . by the cauchy integral formula
@xmath536 hence this integral does not depend on the value of @xmath537 and , taking into account that @xmath538 , we get @xmath539 * step 2 .
* denote @xmath540 .
@xmath522 leaves invariant @xmath541 for any @xmath542 .
repeat step 1 @xmath543 times more ( i.e. totally we perform step 1 @xmath437 times ) .
it is proved by induction that @xmath544 and the step @xmath545 essentially repeats the argument in step 1 .
for @xmath546 we obtain @xmath547 * step 3 . *
the last integral is over @xmath548 , the fundamental domain for a `` standard hyperbolic element '' @xmath549 of @xmath550 acting on the unit disc in @xmath17 .
the change of coordinates @xmath551 where @xmath552 maps the unit disc to the upper half
plane with polar coordinates @xmath553 , so that @xmath554 is mapped onto the upper half
annulus @xmath555 using formulas @xmath556 we obtain @xmath557 but @xmath558 hence @xmath559 and we have @xmath560 the last integral does not depend on @xmath561 as the integral of a @xmath522-invariant holomorphic function of @xmath562 . for the same reason it does not depend the choice of the point @xmath563 and the path from @xmath563 to @xmath564 in @xmath554 . hence we may take @xmath565 , and since on @xmath128 @xmath566 , and the last integral can be rewritten as @xmath567 where the integration is over a segment of @xmath128 . using lemma [ lemma ] again
, we go back to the integral over the axis of @xmath433 , and obtain @xmath568 finally we obtain @xmath569 with @xmath570 the following corollary is immediate from theorem [ integral ] . [
same axis]let @xmath571 be two primitive loxodromic elements having the same axis
. then @xmath572 .
the integral ( [ per ] ) is well defined and is called the _ period of @xmath104 over the closed geodesic @xmath573$]_. the reason for this definition is the following result .
[ lift int]let @xmath290 be the lift of the cusp form @xmath574 to @xmath0
. then for any lift @xmath170_w$ ] of the closed geodesic @xmath170 $ ] to @xmath0 we have @xmath575_w}{{\tilde f}}dt.\ ] ] here @xmath576 with @xmath577 , and @xmath144 is the constant from theorem [ integral ] .
we make a change of variables @xmath140 used in [ flow - def ] which maps the `` horizontal '' geodesic to the axis of @xmath433 in such a way that @xmath578 .
according to [ flow - def ] all lifts of the segment of the geodesic @xmath579 $ ] are given by @xmath580 , and all lifts of the corresponding segment of the `` horizontal '' geodesic , by @xmath581}. using local coordinates ( [ local - coor ] ) on @xmath0 we obtain @xmath582_w}f(gx)(e^{-i\psi}\zeta_x)^{2(n+1){{\kappa}}}j(g , x)^{2(n+1){{\kappa } } } dt\\ & = & \bar{{\alpha}}^{(n+1){{\kappa}}}e^{-2(n+1){{\kappa}}i\psi}\int_{[{{\gamma}}_0]_w}f(z)\zeta_z^{2(n+1){{\kappa } } } dt.\end{aligned}\ ] ] the last two integrals are in @xmath0 over the lifts @xmath583_w$ ] and @xmath170_w$ ] , and @xmath584 , @xmath585 and @xmath329 , @xmath586 are evaluated at the corresponding value of the parameter @xmath430 : @xmath587 , @xmath588 , @xmath589 .
we used the cocycle identity to obtain the last equality : @xmath590 the required formula now follows from theorem [ integral ] .
notice that since @xmath591 is @xmath2invariant , the integral is in @xmath3 .
now we consider @xmath592 with the inner product ( [ l2 ] ) .
the infinitesimal generators of the frame flow ( [ ff ] ) and the one
parameter subgroup @xmath593 belong to the lie algebra @xmath594 .
the corresponding left
invariant differential operators @xmath595 are defined on a dense set of functions in @xmath596 differentiable along the orbits of the corresponding flows , and are given by the matrices @xmath597 complemented by the third differential operator @xmath598 they generate a three dimensional lie subalgebra of @xmath599 with the commutation relations : @xmath600=2{{\mathfrak d}}',\,[\tfrac{\partial } { \partial\psi},{{\mathfrak d } } ' ] = -2{{\mathfrak d}},\,[{{\mathfrak d}},{{\mathfrak d}}']=-2\tfrac{\partial}{\partial\psi}.\ ] ] then @xmath601 belong to the complexification of @xmath599 , @xmath602 and have the following commutation relations : @xmath603=2{{\mathfrak d}}^+,\,\,[\psi,{{\mathfrak d}}^-]=-2{{\mathfrak d}}^-,\,\,[{{\mathfrak d}}^+,{{\mathfrak d}}^-]=\psi.\ ] ] they generate a three - dimensional real lie algebra isomorphic to @xmath604 .
the properties of these operators are exactly the same as in the case @xmath20 ( cf @xcite 3 , @xcite 2 ) .
a standard fourier analysis argument shows that the space @xmath592 can be decomposed into a direct sum of orthogonal subspaces @xmath609 such that @xmath610 notice that the lift of a cusp form @xmath502 to the group @xmath0 belongs to the space @xmath611 .
suppose there is a cusp form @xmath616 , such that @xmath617 for all loxodromic elements @xmath15 .
first we show that @xmath618 .
if @xmath3 is compact , it follows from remarks 1 and 2 following definition [ aut - form ] .
alternatively , @xmath5 has a finite number of cusps @xcite , and it is sufficient to show that @xmath619 and its first derivatives vanish at each cusp @xmath158 . let @xmath537 be a partial cayley transform ( @xcite ch .
4 ) mapping biholomorphically a siegel domain @xmath620 to @xmath621 in such a way that @xmath622 .
by @xcite @xmath623 .
then @xmath624 contains `` parallel translations '' @xmath625 , ( @xmath626 ) , with @xmath627 , and @xmath628 is invariant under @xmath629 : @xmath630 .
then @xmath631 has a fourier
jacobi expansion ( @xcite ch . 3 , 5 ; @xcite ch .
11 ) @xmath632 and since @xmath104 is a cusp form , by satake s theorem ( @xcite ch .
5 ) @xmath633 for @xmath634 , and the claim follows .
the uniform boundedness of @xmath57 on @xmath3 of finite volume would imply the required result .
if @xmath3 is compact , it follows from the lipschitz condition .
if @xmath3 is not compact , it is sufficient to show that @xmath57 is bounded at each cusp @xmath158 .
the proof is similar to the proof for fuchsian groups in @xcite .
it is based on an application of a partial cayley transform @xmath537 described above .
let @xmath64 be the geodesic in @xmath639 given by @xmath640 , @xmath641 .
then @xmath642 is the geodesic in @xmath643 going to the cusp @xmath158 .
on @xmath64 we have @xmath644 , where @xmath645 . then on @xmath642 , considered as an orbit of @xmath155 , @xmath646
may be estimated using ( [ cohom ] ) for any @xmath647 by the integral @xmath648 over @xmath64 , which is finite .
the finiteness of the volume implies that for any @xmath221 there exists a neighborhood of the cusp @xmath649 such that @xmath650 , and the uniform boundedness of @xmath57 now follows from the lipschitz condition .
the argument of guillemin and kazhdan ( @xcite th .
3.6 ) for negatively curved surfaces is applicable to this situation since it depends only on the commutation relations ( [ com ] ) and the fact which immediately follows from it : for @xmath652 @xmath653 it follows immediately from the fact that has been , apparently , first pointed out in @xcite ( see also @xcite 5 and @xcite p. 203 ) that for any @xmath661 @xmath662 here @xmath663 is the linear combination of the partial derivatives @xmath664 where @xmath665 are coordinates on the bounded domain @xmath317 and @xmath666 . for @xmath667
this can be easily checked by a direct differentiation along the orbit of the frame flow @xmath668 using the decomposition @xmath669 corresponding to the local coordinates ( [ local - coor ] ) with coefficients of @xmath670 independent on @xmath671 .
then @xmath672 since @xmath619 is holomorphic we have @xmath673 | let @xmath0 be a semisimple lie group with no compact factors , @xmath1 a maximal compact subgroup of @xmath0 , and @xmath2 a lattice in @xmath0 .
we study automorphic forms for @xmath2 if @xmath0 is of real rank one with some additional assumptions , using dynamical approach based on properties of the homogeneous flow on @xmath3 and a livshitz type theorem we prove for such a flow . in the hermitian case @xmath4 we construct relative poincar series associated to closed geodesics on @xmath5 for one dimensional representations of @xmath1 , and prove that they span the corresponding spaces of cusp forms . |
+ different classes of inflationary models produce wildly varying predictions for observables with some models being ruled out by recent advances in the precision of cosmic microwave background observations . the planck 2015 results @xcite found the following value for the spectral index of the curvature perturbation , @xmath0 , and the following upper bound on the ratio of the power spectra of the tensor to scalar perturbations , @xmath1 : @xmath2 at 1-@xmath3 with negligible tensors and @xmath4 .
these ( and the fact that non - gaussian and isocurvature perturbations have not been observed ) strongly suggest that primordial inflation is single - field with a concave scalar potential featuring an inflationary plateau . indeed , while the minimal versions of chaotic and hybrid inflation are all but ruled out ( but see refs .
@xcite ) , models with such an inflationary plateau ( e.g. starobinsky @xmath5 inflation @xcite , or higgs inflation @xcite ) have received enormous attention . in these models
( and similar other such models , e.g. @xmath6-attractors @xcite ) , the approach to the inflationary plateau is exponential . in this case , however , distinguishing between models is difficult @xcite .
in contrast , a power - law inflationary plateau was considered in ref .
@xcite , where shaft inflation was introduced , based in global supersymmetry with a deceptively simple but highly non - perturbative superpotential , where @xmath7 is a mass scale . in this paper
, we consider a new family of single - field inflationary models , which feature a power - law approach to the inflationary plateau but are not based on an exotic superpotential like in shaft inflation .
we call this family of models power - law plateau inflation .
power - law plateau inflation is characterised by a simple two - mass - scale potential , which , for large values of the inflaton field , features the inflationary plateau but for small values , after the end of inflation , the potential is approximately monomial .
the family is parameterised by two real parameters , whose optimal values are determined by contrasting the model with observations .
we find that the predicted spectral index of the curvature perturbation is near the sweet spot of the planck observations . for a sub - planckian inflaton
, the lyth bound does not allow for a large value of the tensor to scalar ratio @xmath1 .
however , for mildly super - planckian inflaton values we obtain sizeable @xmath1 , which is easily testable in the near future . even though our treatment is phenomenological , and the form of the potential of power - law plateau inflation is data driven , we develop a simple toy model in global and local supersymmetry to demonstrate how power - law plateau inflation may well be realised in the context of fundamental theory .
we use natural units , where @xmath8 and newton s gravitational constant is @xmath9 , with @xmath10 being the reduced planck mass .
+ we start from the following proposed potential : @xmath11 where @xmath7 is a mass scale , @xmath12 and @xmath13 are real parameters and @xmath14 is a canonically normalised , real scalar field .
@xmath15 is a constant density scale and we assume @xmath16 because otherwise our model is indistinguishable from monomial inflation with @xmath17 , which is disfavoured by observations .
initially we impose sub - planckian values of @xmath14 to compare with perturbative models and avoid supergravity ( sugra ) corrections . for the slow roll parameters , to first order in @xmath18 ,
we find : @xmath19\ , , \label{eq : firstorderepsiloncontainingphi}\ ] ] @xmath20\ , , \label{eq : firstorderetacontainingphi}\ ] ] where the prime denotes a derivative with respect to the scalar field and the second term in the square brackets is the first order correction .
considering this correction allows us to approach the bound @xmath21 . in the above and throughout
, @xmath6 is defined as : @xmath22 these result in the following expressions for the tensor to scalar ratio and the spectral index of the curvature perturbation : @xmath23\ , , \label{eq : firstorderrwithphi}\ ] ] @xmath24\ , , \label{eq : firstordernswithphi}\ ] ] where the @xmath25 term has been omitted in eq . because it is negligible .
we can express : @xmath26^{\frac{1}{n+2 } } \big\{1-\frac{1}{2}\big[n(n+2)q\alpha^{-2 } \big(n+\frac{n+1}{n+2}\big)\big]^{-\frac{n}{n+2}}\big\ } \ , , \label{eq : firstorderphioverm}\ ] ] where we found the critical value @xmath27 , which ends inflation , by setting @xmath28 to 1 in eq .
.. ] therefore , we can write eqs . and in terms of the remaining e - folds of inflation , @xmath29 , as : @xmath30^{-2\frac{n+1}{n+2 } } \big\{1+(n-1)\big[n(n+2)q\alpha^{-2 } \big(n+\frac{n+1}{n+2}\big)\big]^{-\frac{n}{n+2}}\big\}\ , , \label{eq : firstorderr}\ ] ] and @xmath31^{-\frac{n}{n+2}}\big\}. \label{eq : firstorderns}\ ] ] to test if our potential is a successful model we need to see if it produces satisfactory values for the tensor to scalar ratio and the spectral index when compared to the planck observations @xcite
. we have four variables and parameters , namely @xmath12 , @xmath13 , @xmath6 , and @xmath29 . to start with , we consider @xmath32 and @xmath33 .
we restrict @xmath12 and @xmath13 to integer values and keep @xmath6 in the range 0.01 to 0.1 for our initial investigation . at this point
we maintain @xmath14 at sub - planckian values ; super - planckian @xmath14 will be considered later .
we also ensure at all times .
[ fig : allplots ] gives an overview of how @xmath1 and @xmath34 vary for differing @xmath12 , @xmath13 and @xmath6 , within this scope .
[ fig : allplotszoomed ] shows the relevant sections of the graph in more detail .
the slanted black line marks the highest allowed @xmath12 value ( non - integer ) for the relevant @xmath13 , @xmath6 , @xmath29 combination . in terms of the spectral index @xmath0 .
solid ( dashed ) lines correspond to @xmath32 ( @xmath33 ) respectively . on the left we have @xmath35 and on the right @xmath36 .
@xmath12 increases along the length of the lines , counter - intuitively , _
right to left_. the vertical black dashed straight lines represent the 1-@xmath3 bounds on @xmath0 from the planck data .
the slanted solid black line depicts the limit below which values of @xmath14 are sub - planckian . ] in terms of the spectral index @xmath0 .
solid ( dashed ) lines correspond to @xmath32 @xmath37 respectively . on the left we have @xmath35 and on the right @xmath36 .
@xmath12 increases along the length of the lines , counter - intuitively , _
right to left_. the vertical black dashed straight lines represent the 1-@xmath3 bounds on @xmath0 from the planck data .
the slanted solid black line depicts the limit below which values of @xmath14 are sub - planckian . ]
we can see from fig .
[ fig : allplotszoomed ] that sensible @xmath34 values exist in most cases but it would be better if @xmath1 were maximised to make contact with future observations . from the graph , it is clear that increasing @xmath12 decreases @xmath1 .
lowering @xmath12 increases @xmath14 so we need to balance this whilst staying sub - planckian .
increasing @xmath13 increases @xmath1 only marginally for all values of @xmath12 except @xmath38 , but as shown in fig .
[ fig : windowedalphazoom ] , @xmath38 is ruled out .
however , again , this also increases @xmath14 .
we can also see that a larger value of @xmath6 is better for maximising our @xmath1 values , as expected since @xmath39 to lowest order in eq .
, this again also increases @xmath14 though .
table [ table : obsfromgraphs ] summarises these results .
.varying effects of the variables / parameters . [ cols="^,^,^,^",options="header " , ] and @xmath1 superimposed on the planck graph . ] and @xmath1 superimposed on the planck graph , with @xmath29 varying between its min and max values for which @xmath34 is within the planck 1-@xmath3 bounds . ]
in this section we will present a toy - model in supergravity ( sugra ) which can produce the scalar potential of plateau inflation for @xmath40 and @xmath35 .
we will follow the approach of ref .
@xcite but only in form , not assuming the same theoretical framework ( hence , we only consider a toy model ) and with an important difference : we consider a minimal khler potential so that we can avoid producing too large @xmath0 , but retain the successes of the @xmath40 , @xmath35 plateau inflation model . at first
, we consider only global supersymmetry ( susy ) and sub - planckian fields .
we introduce the non - renormalisable superpotential : @xmath41 where @xmath42 are chiral superfields and @xmath7 is a large , but sub - planckian scale .
the f - term scalar potential is then @xmath43 .
\label{vf0susy}\ ] ] the above potential is minimised when .
rotating the fields in configuration space ( assuming a suitable r - symmetry ) we can introduce a canonically normalised , real scalar field @xmath14 such that .
then the scalar potential becomes @xmath44 we consider that there is also a d - term contribution to the scalar potential . mirroring ref .
@xcite , we take @xmath45 where @xmath46 is the scale of a grand unified theory ( gut ) . thus , in total , the scalar potential reads @xmath47 minimising the potential in the @xmath48 direction requires @xmath49 inserting the above in eq .
we obtain @xmath50 which is the @xmath40 , @xmath35 power - law plateau inflation model . to confine ourselves in global susy we have to require that .
given that , the available parameter space is not a lot .
however , generalising the above into sugra may allow super - planckian values for the inflaton , while @xmath51 . in sugra
, we continue to consider the superpotential in eq . and we will also consider a minimal khler potential @xmath52 then the f - term scalar potential is @xmath53.\end{aligned}\ ] ] considering that @xmath54 is sub - planckian , since , we have @xmath55.\ ] ] again , the potential is minimised when . writing , we obtain @xmath56 we consider the same d - term contribution to the scalar potential , given in eq . .
the total scalar potential is now @xmath57 minimising the above along the @xmath48 direction , we find @xmath58 inserting this into eq .
we obtain @xmath59 our finding is technically different from eq . because of the exponential factor in the first term in the denominator . however , this term is important only when , when the exponential is unity . in the opposite case , when , the first term in the denominator is negligible .
so the exponential factor makes no difference and the potential is practically the same with the one in eq . .
this is certainly so if the inflaton remains sub - planckian ( i.e. when @xmath60 ) . using the potential in eq .
it can be checked that the @xmath28-problem of sugra inflation is overcome due to the d - term .
in fact , one can show that when .
however , for a super - planckian inflaton , things are different from power - law plateau inflation , since the sugra correction dominates .
so , our sugra toy model is similar but not identical to the the @xmath40 , @xmath35 power - law plateau inflation model .
to have an idea of the value of the observables in this toy model , we can investigate the slow - roll parameters , which are found to be @xmath61,\end{aligned}\ ] ] where and @xmath6 is given in eq . .
then the spectral index is @xmath62.\ ] ] taking @xmath63 and @xmath64 simplifies the above considerably as @xmath65 and it is evident that as already mentioned . in this limit ,
it is easy to find @xmath66 where we have taken .
then , the observables become @xmath67 these should be contrasted with the predictions of power - law plateau inflation given by eqs . and .
at lowest order and taking we find @xmath68 we see that the predictions of our sugra toy - model are more pronounced with respect to @xmath29 , with both the spectral index and the tensor to scalar ratio smaller . also , the dependence of @xmath1 on @xmath6 is more prominent in the case of power - law plateau inflation . for more realistic values of the inflaton ,
however , where ( i.e. ) we would expect @xmath0 and @xmath1 to lie inbetween the above extremes
. note that the predicted values are well in agreement with the planck data , in all cases .
we have studied in detail a new family of inflationary models called power - law plateau inflation .
the models feature an inflationary plateau , which is approached in a power - law manner , in contrast to the popular starobinsky / higgs inflation models ( and their variants ) but similarly to shaft inflation .
we have shown that power - law plateau inflation is in excellent agreement with planck observations . to avoid supergravity corrections we mostly considered a sub - planckian excursion for the inflaton in field space . as expected
, this resulted in very small values for the ratio of the spectra of tensor to scalar curvature perturbation @xmath1 . in an attempt to improve our results and produce observable @xmath1
we have considered minimising the remaining number of e - folds of primordial inflation when the cosmological scales exit the horizon . to this end , we assumed late reheating as well as a subsequent period of thermal inflation , driven by a suitable flaton field .
we have managed to achieve which might be observable in the future ( see table [ tab : running2 ] ) . for economy
we have also investigated the possibility that our model is characterised by a single mass scale .
we have found that the spectral index of the scalar curvature perturbation @xmath0 satisfies well the planck observations but the model produces unobservable @xmath1 . abandoning sub - planckian requirements allows the model to achieve much larger values of @xmath1 .
indeed , for natural values of the mass scales ( planck and gut scale ) , i.e. without fine - tuning , we easily obtain @xmath1 as large as a few percent ( up to 9% , see table [ tab : maximising r ] ) , which is testable in the near future .
our predicted values for @xmath1 and @xmath0 fall comfortably within the 1-@xmath3 bounds of the planck observations , while different models of the power - law plateau inflation family are clearly distinguishable by future observations ( see figs .
[ fig : ourresultsonplanck ] and [ fig : ourresultsonplanck2 ] ) . from our analysis
, we have found that the best choice of model in the power - law plateau inflation family has the scalar potential , which is also a member of the shaft inflation family of models @xcite .
such a potential was originally introduced by s - dual inflation in ref .
@xcite , where , however , the inflaton was non - canonically normalised so the predicted value for @xmath0 was too large and incompatible with the planck data . following ref .
@xcite but crucially considering canonically normalised fields ( i.e. minimal khler potential ) we have constructed a toy - model realisation in global and local supersymmetry for our preferred power - law plateau inflation model .
all in all , the level of success of power - law plateau inflation , and the fact that it offers distinct and testable predictions make this a worthy candidate for primordial inflation , which may well be accommodated in a suitable theoretical framework , as our toy models suggest .
p. a. r. ade _ et al .
_ [ planck collaboration ] , arxiv:1502.02114 [ astro-ph.co ] k. dimopoulos and c. owen , arxiv:1606.06677 [ hep - ph ] .
b. garbrecht , c. pallis and a. pilaftsis , jhep * 0612 * ( 2006 ) 038 .
a. a. starobinsky , phys .
b * 91 * ( 1980 ) 99 ; sov .
. lett .
* 9 * ( 1983 ) 302 .
f. l. bezrukov and m. shaposhnikov , phys . lett .
b * 659 * ( 2008 ) 703 .
r. kallosh , a. linde and d. roest , jhep * 1311 * ( 2013 ) 198 ; jhep * 1408 * ( 2014 ) 052 .
f. l bezrukov , d.d gorbunov , phys .
b * 713 * ( 2012 ) 365 ; a. kehagias , a. m. dizgah and a. riotto , phys .
d * 89 * ( 2014 ) no.4 , 043527 .
k. dimopoulos , phys . lett .
b * 735 * ( 2014 ) 75 ; pos planck * 2015 * ( 2015 ) 037 .
g. k. chakravarty , g. gupta , g. lambiase and s. mohanty , phys .
b * 760 * ( 2016 ) 263 ; g. k. chakravarty , u. k. dey , g. lambiase and s. mohanty , arxiv:1607.06904 [ hep - ph ] | a new family of inflationary models is introduced and analysed .
the behaviour of the parameters characterising the models suggest preferred values , which generate the most interesting testable predictions .
results are further improved if late reheating and/or a subsequent period of thermal inflation is taken into account .
specific model realisations consider a sub - planckian inflaton variation or a potential without fine - tuning of mass scales , based on the planck and grand unified theory scales .
a toy model realisation in the context of global and local supersymmetry is examined and results fitting the planck observations are determined . |
the key elements of the membrane microwave circuits studied in this work are shown schematically in fig . [
fig : model](a ) .
the circuits are created through a series of patterning steps of an aluminum - coated @xmath15 nm thick @xmath7 nanomembrane , and consist of a mechanical beam resonator , a planar vacuum gap capacitor , a spiral inductor ( @xmath16 ) , and a @xmath17 ohm coplanar waveguide feedline .
the vacuum gap capacitor , formed across the nanoscale cuts in the membrane defining the beam resonator , is connected in parallel with the coil inductor to create an @xmath14 resonator in the microwave c band .
each @xmath14 resonator sits within a @xmath18 @xmath19 m @xmath20 @xmath19 m square membrane and is surrounded on all sides by a ground plane .
the coplanar waveguide feedline is terminated by extending the center conductor from one side of the membrane to the other , where it is shorted to the ground plane . electrical excitation and read - out of the @xmath14 resonator
is provided by inductive coupling between the center conductor and the spiral inductor . note that although thinner membranes could have been used , our choice of a @xmath15 nm thick membrane allows for compatibility with single - mode near - ir photonic devices , and is guided by an ultimate goal of integrating planar optical components with electromechanical ones as per ref .
@xcite .
the electromechanical coupling between the beam resonator and the @xmath14 circuit in general depends upon the particular resonant mode of the beam , and is given in terms of the linear dispersion ( @xmath21 ) of the microwave circuit resonance frequency ( @xmath22 ) with respect to modal amplitude coordinate @xmath23 , @xmath24 here @xmath25 is the vacuum gap capacitance across the beam , @xmath26 is the total capacitance of the circuit , and @xmath27 is the motional participation ratio . in the case of uniform in - plane beam motion , and assuming @xmath25 behaves approximately as a parallel plate capacitor , the cavity dispersion simplifies to @xmath28 , where @xmath29 is the nominal capacitor gap size .
the vacuum coupling rate , describing the interaction between light and mechanics at the quantum level , is given by @xmath30 , where @xmath31 is the zero - point amplitude , @xmath32 is the motional mass , and @xmath33 is the mechanical resonance frequency of a given mechanical mode of the beam . in this work
we consider a patterned beam resonator of width @xmath34 @xmath19 m and length @xmath35 @xmath19 m which supports two in - plane resonant modes which can be coupled efficiently to microwave or optical cavities @xcite .
the beam unit cell , shown in fig .
[ fig : model](b ) , has a lattice constant @xmath36 and contains a central hole of width @xmath37 and height @xmath38 .
a pair of upper and lower aluminum wires of thickness @xmath39 nm and width @xmath40 nm at the edges of the beam form one half of the vacuum gap capacitor electrodes .
simulations of the mechanical modes of the beam are performed using a finite - element method ( fem ) solver @xcite , and include the internal stress of the nitride film ( @xmath41 gpa ) .
the simulated fundamental in - plane flexural mode of the patterned and wired beam , a displacement plot of which is inserted into the microwave circuit of fig .
[ fig : model](e ) , occurs at a frequency of @xmath42 mhz . as shown in fig .
[ fig : model](c , d ) , a higher frequency mode also results from bragg diffraction of acoustic waves due to the patterning of holes along the beam s length . in the structure studied here the nominal hole parameters
are chosen to be @xmath43 @xmath19 m and @xmath44 @xmath19 m , which results in a @xmath45 mhz phononic bandgap around a center frequency of @xmath46 mhz .
a defect is formed in the phononic lattice by increasing the hole width ( @xmath37 ) over the central @xmath47 holes of the beam , resulting in a localized `` breathing '' mode of frequency @xmath48 mhz that is trapped on either end by the phononic bandgap . from the simulated motional mass of both mechanical resonances ,
the zero - point amplitude is estimated to be @xmath49 fm and @xmath50 fm for the flexural and breathing modes , respectively . as motional capacitance scales roughly with mechanical resonator size , realizing large electromechanical coupling to nanomechanical resonators depends crucially on minimizing parasitic capacitance of the microwave circuit as per eq .
( [ eq : coupling ] ) . utilizing a planar spiral inductor coil of multiple
turns greatly increases the coil inductance per unit length through mutual inductance between coil turns , and consequently , reduces coil capacitance .
one can determine the capacitance ( @xmath51 ) and inductance ( @xmath16 ) of a given coil geometry by numerically simulating its self resonance frequency with and without a known small shunting capacitance .
figure [ fig : model](f ) displays a method of moments numerical simulation @xcite of the self resonance frequency ( @xmath52 ) of a series of square planar coil designs with constant area ( @xmath53 @xmath19 m @xmath54 @xmath19 m ) but varying wire - to - wire pitch @xmath55 , or equivalently , coil turns @xmath56 .
here we assume a coil wire width and thickness of @xmath57 nm and @xmath58 nm , respectively , deposited on top of the @xmath15 nm nitride membrane . while the coil capacitance is roughly constant at @xmath59 ff , the coil inductance varies over @xmath60 orders of magnitude , in good agreement with an analytical model for planar inductors @xcite .
an additional stray capacitance of @xmath61 ff is estimated for the full integrated microwave circuit ( see app .
[ app : a ] for details ) . figure [ fig : model](g ) displays the simulated motional capacitance and vacuum coupling rate versus capacitor slot size @xmath62 for both the flexural and breathing modes of the beam resonator assuming a coil of pitch @xmath63 @xmath19 m ( @xmath64 , @xmath65 nh , @xmath66 ff , @xmath67 ghz ) . here
, @xmath68 is calculated for each specific mechanical mode utilizing a perturbation theory depending on the integral of the electric field strength at the dielectric and metallic boundaries of the vacuum gap capacitor @xcite . for a gap size of @xmath69 nm ,
the vacuum coupling rate is estimated to be @xmath70 hz ( @xmath71 hz for @xmath72 ) for the flexural mode and @xmath73 hz ( @xmath39 hz for @xmath72 ) for the breathing mode .
note that here we assume the outer electrode of the vacuum gap capacitor extends along the entire length of beam in the case of the flexural mode , whereas for the breathing mode we limit the outer capacitor electrode to the central @xmath74 lattice constants of the beam where the breathing mode has significant amplitude .
also , for the breathing mode simulations the two vacuum gap capacitors are assumed to be connected in parallel , which doubles the vacuum coupling rate due to the mode symmetry .
fabrication of the membrane microwave circuits begins with the lpcvd growth of @xmath15 nm thick stoichiometric @xmath7 layers on the top and bottom surfaces of @xmath75 @xmath19 m thick silicon wafer , and involves a series of electron beam lithography , dry etching , aluminum evaporation , and chemical wet etching steps .
an optical image of the fully fabricated and wirebonded chip is shown in fig .
[ fig : fab](a ) .
zoom - in scanning electron microscope ( sem ) images of the inductor coil and nanobeam regions of the device are shown in fig .
[ fig : fab](b - e ) .
the main fabrication steps are depicted in fig .
[ fig : fab](f ) and discussed in more detail in app .
[ app : c ]
. one important feature of our fabrication method is the use of the tensile membrane stress ( @xmath41 gpa ) to fabricate capacitive slot gaps that shrink upon release of the membrane , providing a controllable way to create ultra - small gaps . as can be seen in the device figures of fig .
[ fig : fab](b , c ) , stress release cuts are used above and below the nanobeam region so as to allow the membrane to relax on either side of the beam . a simulated plot of the membrane relaxation is shown in fig .
[ fig : fab](g ) .
comparison of the simulated slot gap change ( @xmath76 ) and measured slot gap change for a series of fabricated devices with different cut angles @xmath77 is shown in fig .
[ fig : fab](h ) , indicating that slot gap adjustments up to @xmath45 nm can be reliably predicted and produced . in the measured device of this work we use this feature to controllably close the capacitor slot @xmath62 from an initial slot size of @xmath78 nm right after dry etching , down to a final slot size of @xmath0 nm after membrane release . as shown in fig .
[ fig : fab](b ) , in the device studied here each nanobeam is coupled on one side to one coil , and on the other side to another coil .
the capacitor electrodes also extend across the whole length of the device in order to maximize coupling to the low frequency flexural mode of the beam .
the two coils have different lengths , resulting in different @xmath14 resonant frequencies .
as will be presented elsewhere @xcite , such a double - coil geometry can be used to perform coherent microwave frequency translation using the intermediate nanomechanical resonator as a parametric converter @xcite . in the following , however , we will focus on the lower frequency circuit ( larger coil ) only .
the device is cooled to a fridge temperature of @xmath79 mk using a cryogen free dilution refrigerator , and connected to a microwave test set - up consisting of low noise control and readout electronics for electromechanical characterization ( see app . [ app : b ] for details ) .
sweeping a narrowband microwave source across the @xmath74 ghz-@xmath47 ghz frequency range , and measuring in reflection , we find a high-@xmath9 , strongly coupled microwave resonance at @xmath80 ghz corresponding to the larger coil of @xmath81 turns .
this is very close to the expected @xmath14 resonance frequency based upon the above simulations , indicating that the stray and motional capacitance of the circuit are close to the expected values . using a two - tone pump and probe scheme we are able to study the coherent interaction between the microwave electrical circuit and the coupled nanobeam mechanical resonator . in
the driven linearized limit @xcite , the circuit electromechanical system is approximately described by an interaction hamiltonian @xmath82 , where @xmath83 ( @xmath84 ) is the microwave photon annihilation ( creation ) operator for the @xmath14 resonator mode of the circuit and @xmath85 ( @xmath86 ) are the phonon annihilation ( creation ) operators of the mechanical resonance .
@xmath87 is the parametrically enhanced electromechanical coupling strength , with @xmath88 corresponding to the number of intra - cavity microwave drive photons inside the resonator . as schematically
indicated in fig .
[ fig : eit ] , pumping with a strong drive at a detuning @xmath89 from the @xmath14 resonance of the circuit produces a two - photon resonance condition with a second ( weaker ) probe tone as it is swept across the microwave resonance .
interference in the reflected probe signal occurs between that part of the probe field which enters the microwave resonator and is directly re - emitted , and that part of the probe field which enters the cavity , interacts with the mechanical resonator , and is then re - emitted from the cavity . probing this interference as a function of the probe detuning @xmath90 yields the optomechanical analog of electromagnetically induced transparency ( eit ) @xcite . for red - sideband pumping ( @xmath91 ) the expected probe reflection spectrum
is given by , @xmath92 where @xmath93 is the detuning of the probe frequency ( @xmath94 ) from the cavity resonance ( @xmath22 ) , @xmath95 is the external microwave cavity damping rate due to coupling to the coplanar waveguide port , @xmath96 the intrinsic cavity damping rate , and @xmath97 is the total loaded cavity damping rate . here
we have made approximations assuming the system is sideband resolved ( @xmath98 ) and that the probe signal is weak enough so as to not saturate the drive tone .
the cooperativity associated with the coupling of the microwave cavity field to the mechanical resonator is given by @xmath99 , where @xmath100 is the back - action - induced damping of the mechanical resonator by the microwave drive field .
in this work we focus on the fundamental in - plane flexural mode of the beam .
the phononic crystal breathing mode at a frequency of @xmath46 mhz is not accessible in our current single microwave resonator circuit given the high drive power required to excite the circuit at the large cavity detuning required for two - photon resonance . in future work a double resonant system @xcite
may be employed to overcome this limitation and allow for efficient excitation and detection of high frequency mechanical resonators such as the breathing mode . by stepping the pump detuning frequency ( @xmath101 ) and sweeping the probe signal across the cavity resonance , an eit - like transparency window in the microwave cavity response
is found at a drive detuning of @xmath102 mhz , close to the theoretically simulated resonance frequency ( @xmath103 mhz ) of the fundamental in - plane flexural mode .
figure [ fig : eit](b ) shows a series of measured probe spectra ( blue points ) at different applied drive powers for a drive detuning fixed close to the two - photon resonant condition of @xmath104 mhz .
fits to the measured spectra are performed using eq .
( [ eq : eit ] ) and plotted as solid red curves in fig .
[ fig : eit](b ) . from each fit
we extract the loaded microwave resonator properties ( @xmath105 , @xmath95 , @xmath22 ) , the parametric coupling rate ( @xmath106 ) , the mechanical frequency ( @xmath33 ) , and the intrinsic mechanical damping rate ( @xmath107 ) .
these fit values are plotted versus drive power in fig .
[ fig : eit](c ) . the microwave cavity parameters ( @xmath108 mhz , @xmath109 mhz )
are found to be approximately constant over @xmath110 orders of magnitude in drive power , up to an intra - cavity photon number of @xmath111 . for @xmath112 the intrinsic damping of the cavity begins to rise , and above @xmath113 ( @xmath114 dbm ) the @xmath14 circuit goes normal .
conversion from drive power to intra - cavity photon number @xmath88 is performed using the thermometry calibrations described in the next section . at low drive power ( @xmath115 )
the fits yield high confidence estimates of both @xmath116 and @xmath107 , with the intrinsic mechanical damping of the resonator estimated to be @xmath117 hz at the lowest drive powers ( see inset to fig . [
fig : eit](b ) ) . at high drive powers ( @xmath118 )
the transparency window saturates and becomes too broad to accurately determine either @xmath116 or @xmath107 .
as such we only provide fit estimates for @xmath116 and @xmath107 below a cooperativity of @xmath45 .
figure [ fig : eit](d ) shows a plot of the measured mechanical frequency shift ( @xmath119 ) and damping ( @xmath120 ) versus drive detuning @xmath101 .
here we adjust the drive power as a function of drive detuning so as to maintain a constant intra - cavity drive photon number , and fit the transparency window using a fano lineshape ( see si ) .
data was taken at @xmath121 mk as well as at an elevated fridge temperature of @xmath122 mk .
the intra - cavity drive photon number in both cases was chosen to yield a peak cooperativity of order unity .
we observe broadening of the mechanical linewidth that peaks at a detuning @xmath101 equal to the mechanical resonance frequency , and stiffening ( softening ) of the mechanical mode for drive detuning above ( below ) the mechanical resonance frequency .
plots of the theoretical damping and frequency shift due to radiation pressure backaction @xcite are shown as solid back curves in fig .
[ fig : eit](d ) .
we find a parametric coupling rate @xmath123 khz ( @xmath124 khz ) and intrinsic mechanical damping rate @xmath125 hz ( @xmath126 hz ) that fit both the damping and spring shift curves at @xmath121 mk ( @xmath127 mk ) , in close agreement with the estimated values from the fixed detuning data in fig .
[ fig : eit](b ) .
measurement of the mechanical resonator noise is used to calibrate the delivered microwave power to the circuit and to study the backaction cooling of the mechanical resonator . in the resolved sideband limit ( @xmath128 ) ,
efficient scattering of drive photons by mechanical motion occurs for @xmath129 , in which either anti - stokes ( @xmath130 ) or stokes ( @xmath131 ) scattering is resonant with the cavity .
blue detuned pumping at @xmath132 results in stokes scattering of the drive field , down - converting a photon to the cavity resonance and emitting a phonon into the mechanical resonator in the process .
red detuned pumping at @xmath130 , as illustrated in fig .
[ fig : thermal](a ) , leads to predominantly anti - stokes scattering in which a drive photon is up - converted to the cavity resonance and a phonon is absorbed from the mechanical resonator .
the per - phonon anti - stokes scattering rate for this pumping geometry is @xmath133 , to a good approximation equal to the backaction damping rate @xmath134 which leads to cooling of the mechanical resonator @xcite .
figure [ fig : thermal](b ) shows a plot of the measured area underneath the lorentzian noise peak of the fundamental in - plane mechanical resonance versus fridge temperature . here
, data for blue detuned ( @xmath131 ) driving has been averaged over several different temperature sweeps , with the area at each temperature normalized to units of phonon occupancy ( @xmath135 ) using the high temperature measurement ( @xmath136 mk ) as a reference point . in these measurements
the drive power was kept at a low enough value to ensure @xmath137 and negligible backaction damping or amplification .
the mechanical flexural mode is seen to thermalize with the fridge temperature all the way down to @xmath138 mk , at which point the mechanical mode temperature saturates .
the source of this temperature saturation in the mechanics is not fully understood , but is thought to be due to coupling to two - level systems ( tls ) in the amorphous @xmath7 membrane @xcite .
these tls can be driven by the microwave input signal into an elevated temperature state , and , as presented in the si , can also strongly couple with the high impedance microwave cavity resonance .
this latter property may interfere with the mechanical transduction process , leading to unreliable thermometry of the mechanical mode . for a known temperature of the mechanical resonator , one may also employ the above low - cooperativity thermometry measurement to calibrate the vacuum coupling rate @xmath139 between the mechanics and the microwave circuit @xcite ( see app .
[ app : h ] for details ) . as the reflected drive signal and the scattered photons by the mechanical mode experience the same amount of gain , normalizing the measured reflected noise spectrum ( @xmath140 ) by the measured reflected drive tone amplitude ( @xmath141 ) yields a lorentzian of the following form for a drive detuning of @xmath142 , @xmath143 the background offset @xmath144 yields the added noise of the measurement amplifier chain ; @xmath145 for our current set - up . integrating the normalized spectral density for the reference fridge temperature of @xmath146 mk ( @xmath147 ) , and assuming @xmath148 , yields a vacuum coupling rate of @xmath149 hz , comparable to that estimated from numerical simulation ( @xmath150 hz for @xmath69 nm ) . with @xmath139
calibrated , the conversion factor between drive power and intra - cavity drive photon number can now be determined from the coherent two - tone spectroscopy measurements of @xmath87 , as displayed in fig .
[ fig : eit](c ) . increasing the drive power to large cooperativity levels results in backaction cooling of the mechanical resonator for detuning @xmath142 .
figure [ fig : thermal](c ) plots the measured occupancy of the mechanical resonator versus drive power for three different fridge temperatures , @xmath146 , @xmath127 , and @xmath151 mk .
for the lowest of these temperatures ( @xmath152 mk ) , the measured noise power spectral density from low to high drive power are shown in fig .
[ fig : thermal](d ) . at low drive powers
we find excellent agreement between the inferred @xmath153 and the bath occupancy corresponding to the fridge temperature , @xmath154 , for all three temperatures . at intermediate drive powers
the mechanical mode is both damped and cooled according to @xmath155 . at the highest drive powers we measure both
an increase in broadband added noise and lorentzian microwave cavity noise .
we attribute this excess noise to absorption of the input microwave drive .
these two additional noise inputs can lead to noise squashing in the measured output spectrum and heating of the mechanical resonator @xcite . using a model that includes microwave ( thermal ) noise in the input line ( @xmath156 ) and in the microwave cavity ( @xmath157 ) ( see app .
[ app : f ] for details ) , we fit the measured spectra at higher drive power for the mechanical mode occupancy @xmath153 ( blue symbols ) and the microwave cavity noise occupancy @xmath158 ( orange symbols ) . the lowest mechanical occupancy is found to be @xmath6 for a drive photon number of @xmath159 and a fridge temperature of @xmath152 mk , and is similar to the lowest occupancies realized to date for other backaction cooled electromechanical resonators @xcite .
measurements at the lowest fridge temperature of @xmath160 mk resulted in inconsistent and fluctuating cooling curves , attributable we believe to drive - power - dependent coupling of individual tls to the microwave cavity ( see app . [
app : i ] for details ) .
utilizing @xmath7 nanomembranes as a substrate for superconducting microwave circuits enables the formation of high - impedance circuit elements with large per photon electric field strengths . in the current work
, the reduced thickness and low dielectric constant of the nanomembrane helps realize a microwave resonator with an estimated vacuum field strength as large as @xmath161 v / m .
this feature gives rise to the large electromechanical coupling that we observe to the fundamental flexural mode of an integrated phononic crystal nanobeam .
dynamical backaction cooling via a strong microwave drive tone results in an occupancy of @xmath6 for the @xmath5 mhz flexural mode of the beam , limited here by heating of the circuit due to absorption of the microwave drive at the highest powers .
substantial further reduction in the coil and stray capacitance should be possible through tighter coil wiring and optimized layout of the capacitor wiring , respectively , greatly reducing the required drive power ( and corresponding heating ) for backaction cooling to the quantum ground state .
our results also indicate that capacitive coupling to smaller , much higher frequency nanomechanical resonant modes is possible utilizing the @xmath7 platform .
in particular , the planar nature of the membrane circuit allows for integration with slab phononic crystals which can be used to guide and localize mechanical excitations over a broad ( @xmath45 mhz-@xmath162 ghz ) frequency range @xcite .
numerical simulations show that the electromechanical coupling strength to the localized `` breathing '' mode at @xmath163 mhz of the phononic crystal nanobeam of our device is large enough for efficient photon - phonon parametric coupling ; however , a doubly - resonant microwave cavity is needed to drive the system in such a deeply sideband resolved limit @xcite .
coupling phononic crystal structures to superconducting microwave circuits would allow not only for exquisite studies of phonon dynamics using the toolbox of circuit - qed @xcite , but could also be used to realize a quantum network involving superconducting qubits , phonons , and optical photons @xcite .
the authors would like to thank joe redford , lev krayzman , matt shaw , and matt matheny for help in the early parts of this work .
lh thanks andreas wallraff for his support during his master s thesis stay at caltech .
this work was supported by the darpa meso program , the institute for quantum information and matter , an nsf physics frontiers center with support of the gordon and betty moore foundation , and the kavli nanoscience institute at caltech .
ap was supported by a marie curie international outgoing fellowship within the @xmath164 european community framework programme , nemo ( ga 298861 ) .
61ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo
secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ] + 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty @noop @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop ( ) @noop * * , ( ) @noop ( ) @noop * * , ( ) link:\doibase 10.1103/revmodphys.86.1391 [ * * , ( ) ] @noop * * , ( ) @noop ( ) @noop ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) http://dx.doi.org/10.1038/nature08524 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1364/oe.20.024394 [ * * , ( ) ] http://stacks.iop.org/1742-6596/264/i=1/a=012025 [ * * , ( ) ] http://stacks.iop.org/1367-2630/13/i=1/a=013017 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.109.130503 [ * * , ( ) ] @noop * * , ( ) http://dx.doi.org/10.1038/nature13029 [ * * , ( ) ] @noop link:\doibase 10.1109/4.792620 [ * * , ( ) ] @noop link:\doibase 10.1364/oe.23.003196 [ * * , ( ) ] @noop `` , '' ( ) @noop * * , ( ) , @noop ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) http://dx.doi.org/10.1038/nature09898 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.104.033901 [ * * , ( ) ] @noop * * , ( ) @noop `` , '' ( ) @noop @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1126/science.aac5138 [ * * , ( ) ]
@noop `` , '' ( ) @noop * * , ( ) link:\doibase 10.1126/science.1257219 [ * * , ( ) ] @noop ( ) in link:\doibase 10.1109/memsys.2000.838579 [ _ _ ] ( ) pp . @noop * * ( ) @noop _ _ ( , ) link:\doibase 10.1103/physrev.124.1866 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.4710520 [ * * , ( ) , http://dx.doi.org/10.1063/1.4710520 ] http://dx.doi.org/10.1038/nature08681 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.105.163601 [ * * , ( ) ] link:\doibase 10.1126/science.1226487 [ * * , ( ) ] , link:\doibase 10.1103/physreva.75.032329 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.94.123602 [ * * , ( ) ]
our device is fabricated and simulated on a 300 thick and @xmath165 large @xmath166 membrane .
the coil wire is 500 wide and 120 thick , with a @xmath167 m pitch , 42 turns forming a square with lateral length of only @xmath168 , well in the lumped element limit . according to finite element simulations , which includes wire cross - overs , the coil is inductive up to its self resonance frequency of @xmath169 ghz , where the half wavelength roughly matches the total wire length of @xmath170 mm .
we repeat this simulation with a small additional shunt capacitor of known value ( @xmath171 ff ) and extract the new self resonance frequency @xmath172 . solving the two simple relations @xmath173 and @xmath174
, we extract @xmath175 and @xmath176 . these results are valid close to - but below - the self resonance frequency of the coil . in this limit
we realize a maximum impedance of @xmath177 , far exceeding the vacuum impedance @xmath178 , and approaching the resistance quantum @xmath179 . knowing the inductance @xmath16 of the fabricated inductor , as well as the actually measured resonance frequency of @xmath180 , yields a total capacitance of @xmath181 and a total circuit impedance of @xmath182 ( see fig .
[ fig : setup ] ) .
the modulated capacitance @xmath183 is a function of the capacitor slot size , which we estimate from sem images to be on the order of @xmath184 .
numerical finite element simulations yield a nanobeam capacitance of @xmath185 for this gap size ( see figs . in main text ) , which gives a participation factor of @xmath186 . using a self resonance frequency simulation of the full electrical circuit including @xmath183
, we can attribute the remaining stray capacitance of @xmath187 to the coil to capacitor wiring ( @xmath188 % ) , the presence of a second resonant circuit ( @xmath126 % ) , the coupling wire ( @xmath189 % ) , non - ideal crossovers ( @xmath189 % ) . the remaining 0.35 ( @xmath190 % ) we attribute to frequency dependence , packaging and our uncertainty of the relative permittivity of silicon nitride at low temperature @xmath191 .
as expected , for these full circuit simulations we extract the same inductance @xmath16 as for the coil only simulations .
the value of @xmath175 is consistent with both , the modified wheeler and the current sheet method @xcite , to within @xmath192 nh . in order to estimate the electromechanical coupling of the high frequency acoustic mode
, we consider that the identical microwave circuit is coupled to both sides of the nanobeam . here
the outer capacitor length is taken to match the acoustic defect region of @xmath193 lattice constants , see fig .
[ fig : simulation ] .
we find a reduced participation ratio @xmath194 due to the reduced @xmath183 in this case .
further improvements in reducing the circuit s stray capacitance will have a big impact for efficient coupling to high frequency modes .
for the measurements of coherent and incoherent electromechanical response , we combine the output of a vector network analyzer with up to two microwave sources , feed the microwave tones to the base plate of a cryogen free dilution refrigerator using ut-085 stainless steel coaxial cables with feedthroughs for thermalization at each temperature stage and an additional attenuation of 50 db to suppress room temperature johnson noise ( see fig . [ fig : setup ] ) .
we couple to the sample in a reflective geometry using a circulator and a low loss , high dielectric constant , copper printed circuit board ( pcb ) . on the pcb and chip we use 50 @xmath3 coplanar waveguides to route the microwave tones all the way to the membrane with very little reflections . on the membrane the center conductor is shorted to ground through a small wire , which couples the waveguide inductively to the lc resonant circuit . on the output side we use another circulator for isolating the sample from 4 k noise , otherwise entering in reverse direction .
crimped niobium titanium superconducting cables are used to connect directly to a low noise , high electron mobility transistor amplifier ( hemt ) at 4 k. from there we use loss loss ut-085 stainless steel - beryllium copper cables and amplify once more at room temperature . in order to suppress spurious response peaks in high drive power cooling measurements
, we add a phase and amplitude adjusted part of the pump tone to the output signal , as shown in fig .
[ fig : setup ] . properly adjusted , this cancels the directly reflected pump tone before entering the low noise amplifier ( lna ) and spectrum analyzer ( sa ) and avoids the occurrence of spurious resonances in the measured spectrum .
after the final amplification we use an electronically controlled microwave switch to distribute the signal to either the spectrum analyzer or the second vector network analyzer port .
after a thorough rca clean we grow a 300 thick film of stoichiometric si@xmath195n@xmath196 at a temperature of 835 , using low pressure chemical vapor deposition on both sides of a doubly polished 200 @xmath197 thick , high resistivity ( @xmath198 k@xmath3-cm ) , si @xmath199 wafer .
after cooldown , the dielectric film has a stress of @xmath8 gpa due to the differential expansion coefficient .
we spin a protective layer of photoresist and dice the wafer in @xmath162 mm @xmath200 mm chips .
the chips are cleaned using weak sonication in ace and ipa and prebaked at 180 @xmath201c for 2 min on a hotplate .
we then spin the front side with zep 520a at 4000 rpm ( for protection ) , bake at 180 @xmath201c for 2 min , spin the back side with zep 520a at 2000 rpm and bake at 180 @xmath201c for 2 min .
patterning of the 16 membrane areas of size @xmath202 mm @xmath203 mm each , is done with 100 kev electron beam exposure with a 200 na electron beam , 50 nm fracturing size and a dose of 250 @xmath204 on the chip back side .
this layer is carefully aligned to the chip corners .
we develop with znd - n50 for 2.5 min and rinse in mibk for 0.5 min .
this is followed by an icp - rie etch of the silicon nitride in the developed areas , using a c@xmath196f@xmath205/sf@xmath206 ( 34/12 sccm ) plasma , generated with an icp power of 1000 w , rf power of 30 w and a dc bias of 84 v , at a pressure of 15 and a temperature of 25 @xmath201c for 7 min 15 s. we finish this layer by a thorough cleaning of the chips using weak sonication in tce , ipa , zdmac , ace and ipa .
this layer initially follows the same procedure to pattern the top side of the chip ( no resist on the back side ) with the nanobeams , pull - in cuts and the global and pattern alignment markers of size @xmath207 , with these process parameters : 300 pa beam , 2.5 nm fracturing , 275 @xmath204 dose , 7 min 50 s etch time .
we then use an o - ring sealed holder to expose only the back side of the chip to @xmath208 % koh in water at 85 @xmath201c ( stir bar at 400 rpm ) .
this anisotropic si wet etch is stopped when the wafer becomes semi - transparent ( dark orange ) in the membrane area , when illuminated with an led on the sealed side of the chip .
the color indicates a silicon thickness of @xmath209 which is usually achieved after @xmath210 h of etching . after cleaning
the chip in ultra - pure deionized water and ipa , we wet etch both the front and back side of the chip in @xmath208 % koh in water at 65 @xmath201c ( stir bar at 100 rpm ) for @xmath211s .
this partially undercuts ( @xmath212 ) the nanobeams for a clean subsequent inverse shadow evaporation process @xcite , used to pattern the small gapped capacitors .
the chips are then rinsed in hot water , fresh piranha solution ( mix 45 ml h@xmath213so@xmath196 with 15 ml h@xmath213o@xmath213 at 85 @xmath201c with stir bar at 300 rpm ) for 8 min followed by a water and ipa rinse .
this layer patterns all of the electrical circuit , except for the coil wires .
we start with a prebake at 180 @xmath201c for 2 min , and spin the front side with zep 520a at 2000 rpm , followed by another bake at 180 @xmath201c for 2 min .
we use 100 kev electron beam lithography to pattern the ground plane and transmission lines ( 200 na beam , 50 nm fracturing , 290 @xmath204 dose with pec ) , as well as the capacitor wires , and the wires connecting the capacitors with the coil end and center ( 10 na beam , 10 nm fracturing , 275 @xmath204 dose ) .
this layer is carefully aligned to the etched negative markers from the previous step .
we develop the chips in the same way and use a @xmath214 plasma ash process ( 50 sccm o@xmath213 , 0.74 bar , 13.56 mhz , 35 w , 2 min ) to descum the surface before deposition of aluminum . for the deposition we use an electron beam evaporator ( 0.3 nm / s , 65 nm thickness at @xmath215 mbar to @xmath216 mbar ) .
we then do a lift - off process in 80 @xmath201c nmp for @xmath217 1 h and carefully rinse in ace and ipa .
now we pattern a scaffolding layer to fabricate the cross - overs .
after prebaking , we spin lor 5b at 3000 rpm and bake at 180 @xmath201c for 5 min , followed by spinning pmma 950k a2 at 4000 rpm and baking at 180 @xmath201c for 5 min .
we then beam write the negative pattern of the cross - over support structure using aligned electron beam lithography ( 200 na beam , 25 nm fracturing , 1000 @xmath204 dose ) .
the resist is developed using mibk : ipa ( 1:3 ) for 1 min , and rinsed in ipa for 30 s. we then wet etch the scaffolding layer using mf-319 for 8 s , followed by a water rinse and ipa which stops the etch .
finally we remove the remaining pmma layer with ace ( 30 s ) and reflow the lor cross - over support layer on a hot plate at 200 @xmath201c for 10 min .
this creates a structurally stable arc shaped cross over scaffolding . in order to pattern the narrow pitch coils
, we spin pmma 495 a8 at 2000 rpm , bake , spin pmma 950k a2 at 2000 rpm and bake again .
then we lithographically define the coil wires , which overlap the capacitor wires ( 10 na beam , 10 nm fracturing , 1800 @xmath204 dose ) and develop the resist as described previously .
development is followed by the same plasma ashing , deposition of aluminum ( 1 nm / s , @xmath58 nm , p@xmath218 mbar ) and lift off , during which the nmp ( at 80 @xmath201c , @xmath60 h ) dissolves the lor scaffolding layer . after a careful rinse with ace and ipa
we reproduce the previous layer recipe to pattern a small ( 500 nm@xmath2194 @xmath19 m ) dc contact wire that symmetrically covers all overlap regions between capacitor wire and coil wire ( two per coil and capacitor ) . here
we use an in - situ ion gun etch process ( normal incidence with 4 cm gridded kaufman source , 400 v , 21 ma for 5 min ) right before the aluminum deposition of thickness 140 nm , in order to establish reliable contact .
contact is tested after lift - off on dc test structures of the same contact size located in the center of the chip .
high resistance contacts with low capacitance at microwave frequencies would lead to additional parasitic in - series capacitances of the fabricated circuit . for the final release step
we prepare a silicon enriched solution of tmah to selectively etch the silicon without aluminum corrosion @xcite .
we use a custom built reactor vessel with thermometer port and a hotplate with magnetic stir bar to mix 60 g of tmah ( @xmath220 % , 6n ) and 250 g water , and then add 5.12 g of silicon powder ( -325 mesh , 5n ) and stir at 300 rpm . after the chemical reaction calms down
we start heating the solution up to 80 @xmath201c .
when the solution is clear , we wait for 1 h and prepare a clear mixture of 5.21 g of tmah ( @xmath220 % , 6n ) and 2.11 g of the oxidizing agent ammonium persulfate in a small beaker .
we add the mixture to the solution ( stir bar at 1000 rpm ) , wait 10 min to 15 min , reduce the stir speed and add the sample in a vertical position .
the sample is securely clamped , but with the membranes open to a steady flow of solution on the back and front side of the chip .
we keep the solution at 80 @xmath201c and wait for the membranes to become fully transparent ( 1 h to 2 h ) . as a last step we carefully remove the sample , rinse it throughly in hot water , cold water , ipa , ultra purified ipa , and dry it using a co@xmath213 critical point dryer .
we use the following convention for the fourier transform . given an operator @xmath221 we define @xmath222 we consider a resonator mode @xmath83 at frequency @xmath22 , which is coupled to a single waveguide with coupling strength @xmath95 , and to the environment with the coupling strength @xmath96 ( see fig . [ fig : schematic ] ) .
we follow general input - output theory @xcite to write the time derivative of the annihilation operator @xmath223 where @xmath224 is the total resonator linewidth , @xmath225 represents the annihilation operator of the coherent input mode , @xmath226 the waveguide mode operator , and @xmath227 is the respective field operator of the resonator environment .
we take the fourier transform to remove the time derivative , and simplify to get the frequency dependence @xmath228 the resonator output field is defined as @xmath229 which we can use to calculate the complex scattering parameter as measured by a network analyzer @xmath230 where the incoherent bath mode terms drop out .
we use this function to simultaneously fit the real and imaginary part of the measured cavity response and extract the intrinsic and extrinsic cavity coupling rates .
is coupled to the coherent waveguide modes @xmath231 and @xmath232 with the external coupling strength @xmath95 .
it is also coupled to a bath of noise photons , ideally at the refrigerator temperature @xmath157 , with the intrinsic coupling strength @xmath96 .
in addition , the waveguide bath mode @xmath233 can be populated with thermal noise photons @xmath156 , which also couples with @xmath95 .
the mechanical resonator mode @xmath85 is coupled to the microwave resonator with the parametrically enhanced electromechanical coupling strength @xmath106 . in addition , it is coupled to a bath of noise phonons , ideally at the refrigerator temperature @xmath135 , with the intrinsic coupling rate @xmath107 . ] it is useful to define the intra - cavity photon number @xmath88 due to a classical coherent drive tone at frequency @xmath234 . we replace the field operators in eq .
[ eq : ares ] with the classical amplitudes @xmath235 and discard the resonator and waveguide bath modes to get @xmath236 here we have introduced the resonator drive detuning @xmath237 and the input photon flux @xmath238 .
the power at the cavity input can be expressed as @xmath239 with @xmath240 the drive power in dbm and @xmath241 the total attenuation of the input line in db .
fano line shapes generally originate from interference between a resonant mode and a background mode @xcite .
experimental imperfections , such as leakage or reflections in the feedline circuit , can lead to such asymmetric cavity line shapes .
we can model this effect by introducing a complex valued external resonator to waveguide coupling parameter @xmath242 , where @xmath243 is a version of the fano parameter . while small @xmath243 values do not change the magnitude of the inferred external coupling ( or the drive photon number ) , they correctly model small asymmetries in the lorentzian cavity response @xcite . for simplicity
we define the generalized coupling @xmath244 and substitute into eq .
[ eq : barecavity ] , to get the generalized resonator line shape @xmath245
in this section we follow previous work @xcite to calculate the coherent response and the full noise spectrum of the system .
in contrast to earlier treatments we also include thermal noise in the feedline circuit , which gives rise to an increased mechanical occupation and an asymmetric cavity noise line shape . the hamiltonian of the coupled microwave cavity - mechanical system ( see fig . [
fig : schematic ] ) can be written as @xmath246 where @xmath247 ( @xmath248 ) is the annihilation ( creation ) operator of the mechanical mode at frequency @xmath33 , and @xmath139 is the electromechanical coupling strength , i.e. the resonator frequency shift due to a mechanical displacement corresponding to half a phonon on average .
we excite the microwave resonator mode using a strong drive tone at frequency @xmath249 , detuned from the resonator frequency by @xmath237 .
the linearized hamiltonian in the rotating frame is then given as @xmath250 where @xmath251 is the parametrically enhanced optomechanical coupling strength .
the linearized langevin equations are given as @xmath252 taking the fourier transform and simplifying we obtain @xmath253 where we have introduced the uncoupled susceptibilities of the cavity and the mechanical mode @xmath254 in the sideband resolved limit @xmath255 , and for positive detuning of the drive tone @xmath91 ( red side pumping ) , the linearized langevin equations can be written approximately as , @xmath256 now we can calculate the cavity output mode @xmath257 we first calculate the coherent part of the system response using eq .
[ eq : aout ] and drop incoherent noise terms to get @xmath258 substituting the bare response of the cavity and the mechanical oscillator we get the coherent eit response function valid for small probe drive strengths @xmath259 in order to take into account potential interference with a continuum of parasitic modes , we can follow the procedure outlined above . substituting @xmath260 we get @xmath261 which can be used to fit asymmetric eit spectra .
using the fourier transforms defined above , we can write the spectral density of an operator @xmath221 as @xmath262 the auto - correlation function of the detected normalized field amplitude ( or the photo current ) of the output mode @xmath263 is then given as @xmath264 substituting @xmath265 and @xmath266 from eq .
[ eq : aout ] we find a general expression for the single sided noise spectrum @xmath267 here , @xmath156 and @xmath157 represent the bath of noise photons from the waveguide and the microwave resonator environment respectively ; @xmath135 corresponds to the phonon bath at the mechanical frequency ( see fig . [
fig : schematic ] ) .
we assume thermal input noise correlations for all input noise terms , i.e. @xmath268 , @xmath269 , @xmath270 , @xmath271 , @xmath272 and @xmath273 . in the sideband resolved regime and positive detuning ( red sideband pump ) we can drop the terms proportional to @xmath274 and @xmath275 . in order to represent a realistic experimental setup ,
we introduce the fixed gain @xmath276 in units of db and the system noise temperature @xmath277 in units of resonator quanta and referenced to the cavity output .
we can now write the full expression for the single sided power spectral density as measured by a spectrum analyzer , valid in the presence of all relevant noise baths @xmath278 . \label{eq : fulls2}\end{gathered}\ ] ] we minimize the number of free parameters by eliminating the resonator bath @xmath157 using the relation @xmath279 with the eqs .
[ eq : linlang1 ] and [ eq : linlang2 ] we can calculate @xcite the mechanical occupation @xmath153 @xmath280 which we use to also replace the mechanical bath occupation @xmath135 in eqs .
[ eq : fulls2 ] . at large drive photon numbers we observe a power dependent increase of the measured noise background .
using a cavity filter to remove broad band phase noise of the microwave source did not remove this background .
similarly , the power levels of the observed microwave signals are believed to be far from saturating the hemt amplifier .
we therefore take the most conservative approach and attribute this background rise entirely to a power dependent thermal waveguide photon bath @xmath156 .
such a noise term could originate from a rise of the electronic temperature of the on - chip feedline circuit . figure [ fig : noise ] a shows a comparison of noise spectra with and without waveguide noise . in the presence of a broad band thermal input field ,
the background rises and the resonator noise peak shrinks , even though the resonator noise bath is kept constant .
the reason is that the cavity filters the broad band input noise , effectively changing the background of the cavity noise peak . if the waveguide noise bath matches the cavity noise bath , no cavity noise peak
is observed even though the cavity temperature is increased .
compared to a model which attributes this background change to a modification of the amplifier noise temperature or an increased attenuation at the output of the detection circuit , i.e. a change of @xmath281 only , we extract almost twice the cavity occupation @xmath158 using eq .
[ eq : fulls2 ] .
this also affects the lowest observed mechanical occupation and raises it from @xmath282 to 0.58 .
figure [ fig : noise ] b shows the extracted noise baths for the dataset measured at 26 mk . in our measurements the cavity noise bath exceeds the waveguide noise bath for all relevant pump powers .
in this case the microwave resonator bath @xmath158 manifests itself as a broad band resonator noise peak on top of the background .
this power dependent noise peak shows a small asymmetry for the highest pump powers .
such an asymmetry is qualitatively expected from interference between narrow band cavity noise and broad band waveguide noise .
we follow a similar procedure as outlined above and introduce a complex waveguide coupling constant to find better agreement with the measured data in this limit .
we make the substitution @xmath260 in the first term proportional to @xmath156 in eq .
[ eq : fulls2 ] and expand it . for @xmath283 we can only keep @xmath243 to linear order and simplify the expressions .
we can then write the asymmetric noise power spectrum with two additional terms as @xmath284.\ ] ] the additional two terms are odd functions with a vanishing integral .
this ensures the same fit results as obtained compared to using the symmetric model eq .
[ eq : fulls2 ] . defined in this way
, the asymmetry scales with the waveguide noise bath and the fano parameter @xmath243 , which is independent of any other parameters .
we find excellent agreement between this model and the measured broad band noise spectra at high pump powers ( see main text ) .
it is important to point out that only the relevant bath occupancies @xmath153 , @xmath158 , @xmath156 as well as @xmath243 ( in the case of the highest drive powers ) are actual fit parameters , while all other relevant parameters are extracted from - or verified in - independent ( low drive power ) measurements .
in the weak coupling regime we can relate the single sided displacement spectrum @xmath285 using @xmath286 with the photon scattering rate @xmath287 for the optimal detuning @xmath142 , and the factor @xmath288 taking into account the limited collection efficiency of photons leaving the cavity .
at low drive powers and sufficient shielding from room temperature johnson noise , it is a very good approximation to set @xmath289 .
eliminating the waveguide noise input allows for a significant simplification of the power spectrum @xmath290 with the mechanical noise bath @xmath135 related to the mechanical occupation @xmath153 in eq .
[ eq : bath ] . at low drive photon numbers
we see no indication of tls , pump phase noise , or waveguide heating of the cavity .
the chosen attenuation and shielding of input and output microwave lines connecting the sample to room temperature johnson noise limits the expected cavity occupation to @xmath291 ( see for example ref .
@xcite for an independent temperature measurement in a similar setup with less attenuation ) . under these assumptions , which are verified by our low power measurements ( constant background noise , no cavity noise peak )
, we can simplify the power spectrum to the more standard form used in cavity electro- and optomechnics @xmath292 without resonator occupation , eq . [ eq : bath ] simplifies to @xmath293 where we introduced the cooperativity @xmath294 and assumed moderate coupling strength @xmath295 in the last step .
for very small drive powers where @xmath296 we can simplify the expected thermal noise spectrum further . dropping terms associated with backaction allows to measure the displacement noise in a self - calibrated way .
this compact model is particularly useful to back out @xmath139 and the system noise temperature with a minimal number of assumptions . starting with eq .
[ eq : fullssimplified2 ] we can make the replacement @xmath297 and drop the backaction term in the denominator . both is valid for @xmath298 .
we then insert @xmath299 with the drive photon number defined in eq .
[ eq : drivephotons ] .
we furthermore assume that the gain of the system is flat over the relevant detuning such that we can introduce the directly reflected pump power measured at the spectrum analyzer @xmath300 .
the cavity output power is related to the cavity input power via eq .
[ eq : barecavity ] @xmath301 finally , we can write the measured noise spectrum , normalized by the measured reflected pump tone @xmath302 only directly measurable system parameters and the temperature of the mechanical mode need to be known to extract @xmath139 without any further assumptions about the particular gain , attenuation or noise temperature of the chosen measurement setup . knowing @xmath139 , it is easy to accurately back out the input attenuation @xmath303 db and drive photon number @xmath88 ( for example from an eit measurement ) . furthermore , from the measured offset @xmath304 we conveniently infer the system noise temperature in units of photons @xmath305 .
the system gain @xmath306 db is then easily determined from the not - normalized wide band background of the measured noise spectrum @xmath140 .
the demonstrated motional sideband cooling is facilitated by the small capacitor gap size , the high impedance and small stray capacitance of the electrical circuit .
these are very desirable properties in the context of quantum electrodynamics ( qed ) , e.g. with atomic systems .
the vacuum fluctuations of the tested resonator give rise to a large root mean square voltage of @xmath307 @xmath19v . with a gap size of only @xmath308 nm
, the electric field across the capacitor @xmath183 is as large as @xmath309 v / m , about @xmath310 times larger than in typical coplanar waveguide resonators and @xmath311 times that of typical small 3d microwave cavities of similar frequency .
for the largest pump photon numbers @xmath312 we apply an rms voltage of @xmath313 kv , corresponding to a maximum field strength of @xmath314 gv / m . to our knowledge
this is a record high value for a low loss superconducting microwave resonator .
large vacuum fields enable efficient dipole coupling @xmath315 to microscopic systems with small dipole moments @xmath316 , such as molecular two level systems ( tls ) @xcite , single atoms , or charge quantum dots .
we demonstrate this by observing coherent coupling between the resonator vacuum field and a nanoscopic tls on the substrate ( see fig .
[ fig : tls ] ) .
the demonstrated ac - stark control of this nanoscopic system complements previous strain based control techniques @xcite , and will be useful to suppress the negative impact of tls in engineered solid state quantum systems .
the quantum nonlinearity of the tls furthermore represents a unique calibration tool , which is generally missing in electro- and opto - mechanical systems .
more importantly , it could point the way how to better understand the origin and location of tls , which are believed to be the limiting factor not only for ground state cooling of electromechanical systems but more generally for improving the integration density of low loss superconducting circuits . in order to observe vacuum rabi mode splitting
, we reduce the probe power significantly such that the probe photon number @xmath317 . in order to control the interaction ,
we use a second off - resonant drive tone at frequency @xmath234 to stark - shift the tls into and out of resonance with the resonator . for sufficient detuning
@xmath318 we can adiabatically eliminate direct transitions of the tls @xcite and define the new stark - shifted tls frequency as @xmath319 with @xmath320 the rabi frequency due to the off - resonant drive tone . using this linearized model
we find very good agreement for different drive detunings @xmath321 and drive strength in the range @xmath322 for which the linearization is valid , see fig .
[ fig : tls ] .
the attenuation @xmath323 db , entering via the drive photon number @xmath88 , is the only fit parameter and agrees with our previous calibration to within 0.5 db .
the presented stark - shift measurements therefore independently confirm the previously calibrated drive photon numbers and the electromechanical coupling @xmath139 .
similar stark - shift calibrations are quite common in superconducting qubit experiments @xcite .
opto- and electromechanics experiments on the other hand typically lack the necessary strong vacuum nonlinearities for an absolute calibration with the vacuum field .
it is now clear that the presence of a coupled tls can strongly modify the resonator lineshape .
this could explain why at low drive powers , where the tls happens to be close to the resonator frequency , the electromechanical transduction efficiency is reduced and the mechanical occupancy appears to be lower than expected ( see fig .
4 c in the main text ) .
although the studied tls should be far detuned in the relevant power range , there may be other weakly coupled tls which do not stark tune as easily , but still absorb and scatter photons at a high rate . at high drive powers on the other hand , all tls are either saturated due to the high intra - cavity photon numbers , or far detuned from the drive and resonator frequencies . on resonance , the vacuum rabi split linewidths
are approximately given by an equal mixture of resonator and tls linewidth , which we use to estimate @xmath324 mhz .
we extract a tls coupling of @xmath325 mhz from the minimum separation of the measured spectroscopic lines ( see fig .
[ fig : tls ] c ) .
this puts our system very close to the strong coupling limit @xmath326 .
we can also put a lower bound on the electric dipole moment of this tls @xmath327 debye depending on the tls orientation and the exact position on the capacitor .
this value is comparable to many atomic , nano- and microscopic systems , where strong coupling is not readily observed .
the presented techniques therefore open up new possibilities to realize hybrid systems and can help study the loss mechanism that plague superconducting quantum processors more directly , selectively and with higher sensitivity . | we present a platform based upon silicon nitride nanomembranes for integrating superconducting microwave circuits with planar acoustic and optical devices such as phononic and photonic crystals . utilizing tensile stress and lithographic patterning of a silicon nitride nanomembrane
we are able to reliably realize planar capacitors with vacuum gap sizes down to @xmath0 nm . in combination with spiral inductor coils of micron pitch , this yields microwave ( @xmath1 ghz ) resonant circuits of high impedance ( @xmath2 k@xmath3 ) suitable for efficient electromechanical coupling to nanoscale acoustic structures .
we measure an electromechanical vacuum coupling rate of @xmath4 hz to the low frequency ( @xmath5 mhz ) global beam motion of a patterned phononic crystal nanobeam , and through parametric microwave driving reach a backaction cooled mechanical mode occupancy as low as @xmath6 .
thin films of silicon nitride ( @xmath7 ) , when grown stoichiometrically via low - pressure chemical vapor deposition ( lpcvd ) on silicon substrates , can be used to form membranes with large tensile stress ( @xmath8 gpa ) , thickness down to tens of nanometers and planar dimensions as large as centimeters @xcite .
the large tensile stress of these films allows one to pattern membranes into extreme aspect ratio nanostructures which maintain precise planarity and alignment @xcite .
the additional energy stored in tension also results in a significant reduction in mechanical damping in high tension nitride films @xcite , with @xmath9-frequency products as large as @xmath10 hz and @xmath11 hz having been observed at room temperature @xcite and millikelvin temperatures @xcite , respectively . as an optical material , @xmath7 thin films have been used to support low loss guided modes for microphotonic applications , with a measured loss tangent in the near - ir of @xmath12 @xcite . owing to their unique elastic and dielectric properties ,
@xmath7 nanomembranes have recently been utilized in a variety of cavity - optomechanical and cavity - electromechanical experiments @xcite involving the interaction of membrane motion and radiation pressure of either optical or microwave light .
these experiments include optical back - action cooling of a millimeter - scale membrane close to its quantum ground state of motion @xcite , measurement of radiation pressure shot noise @xcite and optical squeezing @xcite , and parametric conversion between optical and microwave photons @xcite .
thin film @xmath7 has also been patterned into various other optomechanical geometries , such as deformable photonic crystals @xcite , nanobeams coupled to microdisk resonators @xcite , and optomechanical crystal cavities which can be used to co - localize ( near - ir ) photons and ( ghz ) phonons into wavelength - scale modal volumes @xcite .
here we explore @xmath7 nanomembranes as a low - loss substrate for integrating superconducting microwave circuits and planar nanomechanical structures .
in particular , we exploit the thinness of the nanomembrane to reduce parasitic capacitance and greatly increase the attainable impedance of the microwave circuit .
we also use the in - plane stress to engineer the post - release geometry of a patterned membrane @xcite , resulting in planar capacitors with vacuum gaps down to tens of nanometers . combining the large capacitance of planar vacuum gap capacitors and the low stray capacitance of compact spiral inductor coils formed on a @xmath7 nanomembrane
, we show theoretically that it is possible to realize large electromechanical coupling to both in - plane flexural modes and localized phononic bandgap modes of a patterned beam structure .
two - tone microwave measurements of an @xmath13 ghz @xmath14 circuit at millikelvin temperatures in a dilution refrigerator confirm the predictions of strong electromechanical coupling to the low - frequency flexural mode of such a beam , and microwave backaction damping is used to cool the mechanical resonance to an average phonon occupancy of @xmath6 .
these results , along with recent theoretical and experimental efforts to realize @xmath7 optomechanical crystals @xcite , indicate the viability of @xmath7 nanomembranes as an all - in - one substrate for quantum electro - opto - mechanical experiments .
such membrane systems could be used , for instance , to realize a chip - scale quantum optical interface to superconducting quantum circuits @xcite . |
galaxy interactions are considered a key process in galaxy formation since these violent events could be able to redistribute mass and angular momentum very efficiently ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
as a consequence , chemical elements can be stirred up affecting their distribution and metallicity patterns such as gradients .
there are numerous observational works dedicated to the study of galaxy pairs ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
but only in the last few years , it has been possible to focus on the effects of galaxy interactions on chemical abundances .
interacting galaxies , galaxy pairs , ultra - luminous infrared galaxies , and merger remnants have been found to exhibit low gas - phase metallicity at a given stellar mass compared with the mean mass - metallicity relation ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
recently , @xcite analysed the metallicity gradients of galaxy pairs in detail confirming the central dilution and gradient flattening previously reported by several authors .
these trends have been interpreted as a consequence of low - metallicity inflows triggered by galaxy interactions . indeed
, these interacting galaxies are in majority spirals which , in the nearby universe , hold a gas - phase metallicity gradient ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the presence of metallicity gradients could explain the decrease of the central abundances when low - metallicity inflows are triggered during interactions .
there are also some galaxies in pairs with abundances departing significantly from the mass - metallicity relation such as those reported by @xcite .
hence , there are still effects of galaxy interactions which demand further understanding . by using cosmological chemo - hydrodynamical simulations , @xcite showed for the first time that the interacting galaxies tend to have lower central metallicity when compared to isolated galaxies of similar stellar mass .
these authors showed that the decrease of central abundances could be explained in terms of gas inflows triggered during the interactions .
more recently other numerical works have confirmed these findings by using different simulations and hypothesis @xcite .
particularly , @xcite studied the evolution of metallicity profiles by using hydrodynamical simulations to recreate the major interaction of galaxies with a gas mass fraction similar to that used in our less gas - rich simulations .
however , these authors did not consider new star formation activity and the subsequent chemical enrichment of the interstellar medium as we do in our work . in this paper
, we use a chemical model grafted onto ` gadget-2 ` , which allows the description of the chemical enrichment of baryons as the merger evolves .
our code also includes a self - consistent model of sn feedback @xcite , so mass - loaded galactic outflows can be triggered without including mass - depend parameters .
this code includes the chemical model of @xcite which describes the enrichment of baryons by type ia and type ii sne .
hence , the simulations analysed in this work follow the dynamical evolution together with the chemical evolution of baryons as the interaction takes place .
this paper is organized as follows . in section 2
, we describe the code and the initial conditions designed for this work . in section 3 , we explore wet interactions at low redshift while section 4 analyses a more gas - rich encounter , in order to reproduce typical wet interactions at high redshift .
finally , we summarize our findings in section 5 .
we performed a set of hydrodynamical simulations of interacting pre - prepared disc galaxies of comparable masses .
the simulations were run by using a version of ` gadget-2 ` @xcite modified by @xcite to include the chemical model of @xcite and a new sn feedback model developed to provide an improved description of the interstellar medium ( ism ) and the injection of energy and metals generated by sne .
the multiphase and sn feedback model developed by @xcite allow the coexistence of gas clouds with different thermodynamical properties and are able to describe the injection of energy from young stars into the ism producing the self - regulation of the star formation activity as well as triggering galactic outflows without the need to introduce mass - dependent parameters or to change discontinuously particle momentum to start a wind or temporary suppression of radiative cooling .
this scheme has been successful at reproducing the star formation activity during quiescent and starburst phases and is able to trigger mass - loaded winds with strength dependent on the potential well of the systems ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
chemical elements are synthesized in the stellar interiors , and ejected into the ism by sne .
the chemical model describes the enrichment by snii and snia , providing information on certain individual chemical elements .
we also assume that each sn injects an amount of energy into the ism , @xmath1 ( cf table [ tab1 ] ) .
snii and snia are originated from different progenitors with different rates , yields and time - scales .
we assume that snii are generated by stars more massive than 8 m@xmath2 after @xmath3 yrs .
we adopt the yields given by @xcite . for snia
the chemical model assumes lifetimes randomly distributed in the range @xmath4 $ ] yrs and yields from w7 model of @xcite .
the sn feedback model is able to generate galactic outflows powerful enough to transport enriched material outside the galaxies @xcite .
thus , these simulations provide us with the evolution of the dynamical and the chemical properties of the baryons as galaxy interactions proceed .
we run three major merger simulations ( 1:1 ) varying the orbital parameters ( simi , simii , and simv ) with the same star formation and sn feedback model .
we also performed a run ( simiv ) with the same initial conditions as simii but decreasing the energy release by each sn event ( e@xmath5 ) and a run ( simiii ) with the same orbital parameters but without star formation and feedback activity .
this last experiment is comparable to those analysed by @xcite . in order to represent wet mergers at low redshift , these experiments ( from simi to simv )
were run with a 20 percent of the disc in form of gas . finally , simvi has the same orbital , star formation and feedback parameters as simii but the gas fraction in the disc has been increased to up 50 per cent in order to simulate a wet encounter at high redshift .
the parameters of the simulations are summarized in table [ tab1 ] .
initial conditions correspond to disc galaxies composed of a dm halo , a bulge component ( modeled by a nfw and a hernquist profiles , respectively ) and an exponential disc , with a total baryonic mass of @xmath6 . for the less gas - rich simulations , we use @xmath7 dark matter particles , @xmath8 stars distributed in the stellar disc and the bulge , and @xmath8 particles to represent the disc gas component with an initial gas resolution of @xmath9 m@xmath10 . for the more gas - rich experiment , we re
- calculate the number of particles so as to keep similar mass resolution .
total stellar masses change according to the gas content in the galaxy discs , being @xmath11 for the less gas - rich experiment and @xmath12 for the more gas - rich simulation .
a gravitational softening of @xmath13 kpc was adopted for the gas particles , @xmath14 kpc for the stars , and @xmath15 kpc for the dark matter .
galaxies live in dm halos with circular velocities of @xmath16 .
all these experiments are set to have the same elliptical orbits , according to @xcite , with pericentre distance of @xmath17 kpc and circularity parameter @xmath18 .
these orbital values are consistent with those obtained from the analysis of cosmological simulations ( e.g. * ? ? ?
we have kept them fixed so that changes in the chemical and dynamical properties can only be adscribed to the different angular momentum orientation .
a larger orbital distribution might be of interest to analyse possible dependences in the future .
the gas component is initially assumed to have been pre - enriched by the existing stars , following a metallicity gradient statistically consistent with observations .
we assigned an initial amount of chemical elements to the gas particles in order to reproduce the mean observational gradient reported by @xcite , which for the less gas - rich simulations is @xmath19 dex kpc@xmath20 with a central oxygen abundance of 9.2 .
thus , the initially simulated galaxies are located onto the observed mass - metallicity relation @xcite . for the more gas - rich simulation ( simvi ) ,
the initial metallicity gradient is somewhat steeper ( @xmath21 dex kpc@xmath20 ) and the central metallicity lower ( by 0.7 dex ) than the less gas - rich galaxies . by doing that
, these more gas - rich simulated galaxies are then initially onto the mass - metallicity relation observed at high redshift by @xcite .
note that the exact adopted values are not critical since we are interested in the relative evolution of the metallicity .
.characteristics of the numerical experiments .
@xmath22 : angle between the two total angular momentum of the galaxies .
e@xmath5 : amount of sn energy release by each event .
@xmath23 : gas fraction of the discs . [ cols="<,^,^,^",options="header " , ]
in this section , we analyse the effects of galaxy interactions on the central metallicity in wet local encounters . taking simi as an example
, we show in fig .
[ leo1 ] the evolution of the gas - phase o / h abundance of one member of the galaxy pair as the interaction proceeds .
the evolution of the interaction can be followed by the relative distances between the center of masses of the interacting systems , plotted in the same figure .
as it can be seen , there is a clear evidence of a decrease in the central gas - phase metallicity which can be correlated with the first and the second pericentre .
several previous works have shown that during a close encounter , strong tidal torques can develop , re - distributing angular momentum and producing gas infall ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) .in the case that a metallicity gradient exists , the infalling gas will tend to have lower metallicity than that of baryons in the central region producing a dilution as it is clearly shown by our simulation ( see also * ? ?
* ; * ? ? ?
in order to quantify the chemical changes as a function of time , we compute the central abundances inside concentric spheres of different radius in the range from 1 kpc to 10 kpc . as an example , fig . [ leo1 ] shows the gas - phase oxygen abundance measured inside a sphere of 2 kpc , 4 kpc , and 6 kpc of radius .
as it can be appreciated , as one moves inwards , the signals get noisier since the number of particle decreases .
if one moves to larger radius , the signal decreases as chemical abundances start to be dominated by the external regions .
as a compromised between these two effects , we select @xmath24 kpc ( blue line ) to define the central region where we will perform the analysis of central metallicities . in order to understand the decrease in the central abundances shown by fig .
[ leo1 ] , we analyse the evolution of the central gas inflows and their abundances . we compute the central gas inflow at a certain time as the ratio between the new accreted gas mass within @xmath25 and the total gas content within the same radius .
[ general]a shows the evolution of this central gas inflow as the interacting galaxies approach each other .
we find that the gas inflow remains constant in around @xmath26 until the first close passage when it rises up to @xmath27 .
from the second pericentre , the gas inflow steadily increases up to around @xmath28 .
this gas inflow is dominated by low - metallicity material as it can be appreciated from fig .
[ general]b ( solid line ) where we plot its o / h abundance .
the accretion of this low - metallicity gas dilutes the central oxygen abundance as shown by the continuous decrease detected after the first pericentre ( fig .
[ general]b , dotted line ) .
the main drops in central abundances correlate with the occurrence of the pericentres driven by important gas inflows ( fig .
[ general]a ) . additionally , the gas inflow is also responsible for triggering new star formation activity ( fig .
[ general]c ) . in simi ,
the sfr increases by a factor 3 from the first pericentre as the interaction continues fueling the gas inflow ( fig .
[ general]a , c ) . as a consequence of this increase in the sfr
, the enrichment by snii also increases , diminishing the effects of chemical dilution previously imprinted by the low - metallicity gas inflows .
the increase of the sfr is also reflected in the @xmath0-enhancement , [ o / fe ] , of the gas in the central region ( dotted line of fig . [ general]d ) .
this central @xmath0-enhancement is present from the first pericentre but is much stronger around the second one .
note that the infalling gas into the central region is low @xmath0-enhanced as expected in a disc galaxy such as that of the milky way ( solid line of fig .
[ general]d ) , indicating that the increase of oxygen in the central region is principally due to snii associated to the central starburst .
these general results are in agreement with those reported by @xcite although these authors did not consider chemical enrichment self - consistently .
as we will discuss in the next section , our simulations allow us to make further advances on this issue .
previous works ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) have already reported how star formation histories of interacting systems depend on the initial conditions and orbital parameters . hence , the trends shown in fig . [ general ] for simi might vary among the analysed simulations as we will discuss in next sections . of one galaxy in simi .
a ) fraction of the gas falling within @xmath29 as a function of time . dashed horizontal line represents the initial fraction .
b ) oxygen abundance of the infalling gas ( solid ) and the mean central abundance within @xmath29 ( dotted line ) .
c ) evolution of star formation rate .
d ) @xmath0-enhancement of the infalling gas ( solid ) and its mean central value ( dotted line ) .
arrows indicate the first and second pericentres.,width=317 ] as already shown for simi , the @xmath0-enhancement starts to be detected after the first close passage .
however , the most significant increase is recorded after the second one . in order to analyse
if the @xmath0-enhancement could be used to time the interaction as suggested by @xcite , we study the ism enrichment history along the interactions and its dependence on different orbital configurations .
[ alpha1 ] shows the evolution of [ o / fe ] in the central regions for different simulations with different orbital parameters : simi , simii and simv .
we also include simiv which assumes a lower energy per sn event but otherwise has the same parameters as simii .
we find that before the first close passage ( red points ) , the different simulations exhibit no significant variations in the @xmath30(o / h ) nor [ o / fe ] values .
however , different feedback parameters and orbital configurations impact differently on the ism chemical enrichment during the interacting phase between the first two pericentres ( green points ) . as a general result
, we find an increase of the [ o / fe ] , except for the coplanar retrograde orbit ( simv ) . in order to understand these differences in the @xmath0-enhancement , fig .
[ alpha3 ] shows the central @xmath0-enhancement and the sfr for one of the interacting galaxies in each simulation .
as expected , the central @xmath0-enhancement ( red dotted lines ) correlates with the star formation activity ( green lines ) .
we find that between the first and the second pericentre the [ o / fe ] reaches the highest values for coplanar direct orbits ( simii and simiv ) , decreases as inclination increases ( simi ) , and is almost negligible for the coplanar retrograde interaction ( simv ) .
note that in agreement with previous results by @xcite , we find that coplanar direct interactions trigger stronger starburts than coplanar retrograde configurations during the pre - merger phase ( see fig .
8 of their paper ) .
we also find that the simulation with standard feedback ( simii ) has slightly lower levels of the [ o / fe ] and sfr , than the case of low energy feedback ( simiv ) because in the later , sn feedback drives weaker galactic outflows , being less efficient at transporting material out of the central regions . as discussed previously by @xcite , the @xmath0-enhancement observed in interacting galaxies before the final coalescence of systems might date the interaction .
however , we find that this is valid when the @xmath0-enhancement is only regulated by the star formation process . in this case , the increase of [ o / fe ] during the interacting phase would indicate the occurrence of a recent starburst ( approximately @xmath31 1 gyr ) , because the ism have not had enough time to be heavily enriched by snia yet . on the contrary
, the detection of a decreasing [ o / fe ] value would indicate that snia had time to release fe into the ism , which means that the time elapsed since the last starburst is longer than approximately @xmath32 gyr .
our simulations which follow the dynamical and chemical evolution along the interaction suggest that other factors are also involved in the regulation of the @xmath0-enhancement . as can be seen for simi , simii and simiv ( fig .
[ alpha3 ] ) , the increase in the @xmath0-enhancement after the first pericentre is followed by a drop before the second pericentre .
we find that this dilution of the central @xmath0-enhancement is caused by the low [ o / fe ] enhancement of the very important gas inflow ( open blue symbols in fig .
[ alpha3 ] ) , triggered during the second close approach of the galaxies .
this gas inflow vanish the clear traces in the central abundances left by the star formation activity , restarting the clock to date the interaction . hence , our simulations suggest that the @xmath0-enhancement can only indicate the proximity to the first pericentre , providing extra information on the parameters of the encounter since those with orbits close to coplanar retrograde configurations would tend to have very mild enhancement ( simv in fig .
[ alpha3 ] ) .
however , our findings also suggest that the chance to observe close galaxy pairs with central @xmath0-enhancement might be low , because of the short time interval over which it could be detected ( less than approximately 0.2 gyr according to our experiments ) .
-enhancement of the central region ( red dotted lines ) and of the infalling gas ( blue open symbols ) and the star formation rate ( green lines ) for different simulations : i , ii , iv and v. arrows in each panel indicate the first and second pericentres.,width=317 ] as shown in the previous section , galaxy - galaxy interactions have a clear effect on the central abundances and the star formation of wet interacting systems at low redshift caused by the dynamical perturbation induced on the gas component . in order to illustrate how this affect the gas - phase metallicity profiles ,
we show a sequence of oxygen abundance maps which plot the spatial distribution for the interacting galaxies of simi at different stage of the interaction , with their corresponding oxygen abundance profiles ( fig .
[ leo - map ] ) .
it can be appreciated how the metallicity profile changes as the interacting system gets closer .
this evolution is mainly related to the flattening of the profile caused by a decrease of central metallicities and a simultaneous increase of the abundances in the external region of the disc .
we also show in fig .
[ leo - map ] ( right panel ) that the abundance gradients computed by radial binning of all the gaseous particles are in good agreements with the gradients determined with only the abundances of star - forming regions , as it is done with observations . in order to quantify the evolution of internal metallicity profiles ,
we carry out a linear regression fit to estimate the central oxygen abundances , ( o / h)@xmath33 , and slopes .
the internal profiles are defined in the range of 3@xmath34 kpc .
the inner radial limit corresponds to three times the gravitational softening used for baryons , and for the external radius , @xmath35 kpc corresponds to approximately three scale - lengths of the initial stellar disc .
[ fit1 ] shows the linear fitting coefficients of the radial profiles for different simulations .
as it can be appreciated , there is a significant evolution of internal profiles after the first pericentre , which is consistent with a central dilution of abundances as expected from the discussion of the previous subsection . in agreement with observations @xcite and previous numerical works @xcite , the drop in (
o / h)@xmath33 is about 0.2 dex ( red points ) .
we also find that the final metallicity gradients are flatter than those in earlier stages of the interaction ( green points ) in agreement with ( * ? ? ?
* ; * ? ? ?
* and references therein ) . note that although the evolution of internal profiles seems to be not very sensitive to different orbital configurations in wet major mergers at low redshift , our results show some evidence that galaxies in coplanar direct orbits ( simii and simiv of fig . [ fit1 ] ) experience a larger flattening than those in coplanar retrograde ones ( simv in fig .
[ fit1 ] ) . : red and slope : green ) of the oxygen abundance radial profile , computed in a range of 3@xmath34 kpc for different simulations : i , ii , iv and v. arrows indicate the first and second pericentres.,width=317 ] as mentioned above , the external abundance profiles rise as the galaxy interaction proceeds .
this is a common feature to all simulations , hence , we will take simi again for illustration purposes .
we define the external region of the gaseous disc as everything outside @xmath36 kpc , which also denotes quite well the radius where the outer profiles depart from that defined by the central region ( fig .
[ leo - map ] ) .
then , we estimate the fraction of gas which has gotten into the external disc between two available time intervals , as well as the metallicity of this gas as a function of time . as can be seen from fig .
[ out]a , close to each pericentre , the external disc of interacting systems receives a significant fraction of gas , principally close to the second one ( @xmath37 at the second pericentre ) .
this gas comes from the inner regions of the disc as the arms get opened and distorted by tidal forces ( fig .
[ leo - map ] ) . as a consequence
, this gas tends to have high oxygen abundances ( fig .
[ out]b , solid line ) , which partially contribute , after the first pericentre , to increase the mean external abundances by @xmath38 dex ( fig .
[ out]b , dotted line ) .
this result is in agreement with @xcite , who find that tidal tails transport metal - enriched gas from the galaxy inner regions to the outskirts . on the other hand ,
there is new star formation activity triggered in the external regions , on the spiral arms as can be seen from fig .
[ leo - map ] .
these new stars will also contribute with new sn ii to the enrichment of the external ism .
therefore , the evolution of the external abundances are produced by the combined effects of these two processes : i ) the _ in situ _ sf activity and ii ) the gas - rich material coming from inner regions as the interaction takes place and the spiral arms get distorted .
kpc ) as a function of time .
b ) oxygen abundance of the gas getting into this external region ( solid line ) and the mean external abundances ( dotted line ) as a function of time .
arrows indicate the first and second pericentres .
these calculations are shown for one galaxy member of simi as an example.,width=317 ]
in previous sections , we analyse wet low - redshift interactions of galaxies with around 20 per cent of the disc in the form of gas . here
, we will discuss the metallicity evolution in interacting systems when the initial amount of gas is increased up to 50 per cent ( simvi ) .
this scenario is suitable for studying wet galaxy encounters at high redshift . for this simulation
, we have chosen the same orbital configuration of simii , but with a modified initial distribution of baryons in order to start with more gas - rich systems ( cf .
table [ tab1 ] ) .
we have also adopted the same star formation and sn parameters as for simii .
the initial abundance gradient has been adjusted to be consistent with the observations of the mass - metallicity relation at high redshift .
particularly , we used the observations of @xcite at @xmath39 . hence , the initial zero point is lower and the gradient steeper than in the case of less gas - rich interactions ( see section 2 ) .
[ gasrich1]a shows the gas inflows for one of the galaxy members in simvi . in this case , the amount of gas that falls into the central region is more significant than in its less gas - rich counterpart throughout the whole simulated time interval . particularly after the second pericentre
, the gas inflow roughly duplicates that of the less gas - rich interacting galaxies . the comparison of the sfr evolution of more gas - rich galaxies ( fig .
[ gasrich1]b ) with that of their less gas - rich counterparts ( fig .
[ general]c ) shows that long before the first close passage , the sfr of the more gas - rich system is more active ( around 30 m@xmath40yr ) than that of their less gas - rich counterpart .
this is mainly due to the higher fraction of gas available for star formation .
previous works @xcite have suggested that gas - rich galaxies might experience greater instability on the gaseous disc which can lead to an intense star formation activity via fragmentation and clump formation . in order to check this behaviour in our simulations ,
we compare the evolution of gaseous discs for galaxies in simi and simvi . fig .
[ gasxy ] shows projected gas distributions of one member of the interacting system in simi ( upper panels ) and in simvi ( lower panels ) , computed for different snapshots during the interaction . from these figures
it is clear that , while the gaseous disc in the less gas - rich simulation evolves more smoothly , in the more gas - rich system it has highly fragmented and formed high density clumps where the gas satisfies the condition to form stars in agreement with previous works .
we checked that the jeans mass is resolved by at least 4 gas particle so that fragmentation can be numerically described .
as it can be appreciated in the first low panel , the more gas - rich system shows high level of disk fragmentation well before the first close passage at 0.3 gyr , when the two interacting galaxies are separated by more than 100 kpc and consistent with the peak in the sfr ( see fig . [
gasrich1]b ) .
this suggests that the internal fragmentation of gas - rich discs is principally driven by their own self - gravity . to prove that this early strong sf activity is not induced by the interaction , we run an isolated galaxy case with the same star formation and sn parameters .
the clump formation is also present in our isolated galaxy run .
previous works have shown that galaxy interactions reinforce the level of disc fragmentation for mergers at low and high redshifts ( * ? ? ?
* ; * ? ? ?
* respectively ) .
the detailed analysis of this issue will be part of a future paper focused on clump formation in isolated and interacting disc galaxies .
after the intense sf activity reported during the early stage of interaction , the sfr in the more gas - rich galaxies begins to decline with two subsequent maxima shortly after the first and second pericentres , respectively ( fig .
[ gasrich1]b ) .
the induced star formation activity during the close encounters is less important than that expected considering the strong gas inflows experienced by these galaxies at these stages of evolution ( fig .
[ gasrich1]a ) .
we investigate which mechanism could be operating against the triggering action of gas inflows to inhibit the sf activity .
we analysed the thermodynamical properties of the gas , finding that its internal energy ( dotted line of fig .
[ gasrich1]b ) , increases continuously fed by the sn feedback .
this process is responsible of heating up and blowing away an important fraction of gas with an increasing efficiency as the galaxy approaches the second pericentre when it loses @xmath41 of its initial gas mass ( @xmath42 is lost at the first pericentre and more than @xmath43 at the merging phase ) . as a consequence ,
the level of star formation activity is not only directly correlated with the strength of the gas inflows but it is also modulated by the sn feedback which is , in this simulation , more important than in the less gas - rich interactions because of the larger amount of available gas to feed star formation . we investigate the chemical evolution of the ism by analysing the o / h abundances and @xmath0-enhancements as in the less gas - rich interactions .
[ gasrich1]c shows the evolution of oxygen abundance of the infalling gas and of the central region . from this figure
we can see that the central metallicity of the gas increases steadily up to the first pericentre .
we can also appreciate that the gas inflows , driven into the central regions during the first main starburst , are increasingly oxygen - rich as a consequence of the on - going important star formation activity .
this is a consequence of the first early starburst which is not present in less gas - rich interactions .
however , from the first to the second pericentre , the low metallicity of the infalling gas produces the oxygen dilution of the central region just as in wet local interactions ( fig .
[ general]b ) .
the different levels of enrichment of these gas inflows are due to their different origins : high - metallicity inflows involved gas located nearby the central region while lower metallicity inflows come from more external ones . since , our discs have initially metallicity gradients , these different origins imply different levels of enrichment . to visualize this , in fig .
[ edades ] ( upper panel ) we show the minimum , maximum , and the median location on the disc plane of those gas particles which will be immediately incorporated into the gas inflows . ] . before the first pericentre most of the gas that falls into the central regions comes from about @xmath44 kpc and hence ,
is principally driven by internal dynamics while during the interactions the tidally driven gas inflows are able to transport material inward from further away regions of the disc as can be seen in fig .
[ edades ] ( upper panel ) . in fig .
[ edades ] ( lower panel ) we show the stellar age gradients as a function of time .
we can see sharp drops in the profiles as signatures of a discontinuous sf distribution along the disc , reflecting the clumpy sf activity .
profiles in blue , red , and green show the evolution of the sf distribution before the first pericentre . as shown by these profiles , the strong sf activity reported during this period ( fig .
[ gasrich1]b ) is located not only in the central region but also extended in the outskirt of the galaxy , contributing to enrich the ism .
as the interaction goes on , the main star formation activity takes place principally in the central regions which gets rejuvenated . as a consequence of the contribution of a low - metallicity gas inflows during the close interaction ( between the two first pericentres ) the central o / h abundance decreases by @xmath38 dex ( fig . [ gasrich1]c , dotted line )
the fact that the central oxygen abundance decreases more slowly than that of the infalling gas , at this stage of the interaction , is a consequence of the sf activity being more concentrated in the central region , continuously supplying new chemical elements via snii events .
after the second pericentre , we observe a second sfr enhancement driven by new gas inflow ( figs .
[ gasrich1]a , b ) .
this sf activity is very centralized in the nuclear region of the galaxy as shown in fig .
[ edades ] ( lower panel ) , producing the increase of the central o / h abundance ( fig .
[ gasrich1]c , dotted line ) .
finally , we also investigate the @xmath0-enhancement of wet encounters at high redshift .
[ gasrich1]d shows that the initial [ o / fe ] value is consistent with the @xmath0-enhancement observed in less gas - rich interactions ( fig .
[ general]d ) .
the subsequent evolution of the @xmath0-enhancement shows that immediately after the second pericentre and simultaneously to the sfr enhancement , the central [ o / fe ] ( fig .
[ gasrich1]d , dotted line ) exhibits a sharp increase produced by the important increase of snii feedback associated to the more active central sf . after this @xmath0-enhancement
, we find a significant drop in the central [ o / fe ] values of around 0.8 dex .
this drop is caused by the injection of fe elements into the ism by snia , whose progenitors formed @xmath45 1 gyr before , during the main starbursts . as a function of time .
b ) evolution of star formation rate ( solid line ) and of the mean internal energy @xmath46 of the gas ( dotted line ) .
c ) oxygen abundance of the infalling gas ( solid line ) and the mean central oxygen abundance ( dotted line ) , computed within @xmath25 .
d ) @xmath0-enhancement of the infalling gas ( solid line ) and it mean central value ( dotted line ) .
arrows indicate the location of first and second pericentres.,height=434 ] + .,title="fig:",height=207 ] .,title="fig:",height=207 ] in this section , we quantify the effect of wet interactions at high redshift on the gas - phase metallicity profiles . similarly to the case of wet interactions in the local universe , the evolution of metallicity profiles in the more gas - rich interacting galaxies show a clear flattening and the central dilution of o / h abundances .
we quantify the evolution of the metallicity profiles by performing linear regressions .
[ perfiles ] shows the ( o / h)@xmath33 and slopes as a function of time ( red and green points , respectively ) . as expected , we find that the evolution of ( o / h)@xmath33 matches that of the central abundances computed within @xmath25 ( see dotted line of fig .
[ gasrich1]c ) .
we can see an increase of the ( o / h)@xmath33 in around 0.6 dex during the approaching phase while after the first pericentre , the decrease reflects the impact of the low - metallicity inflows driven by the interactions .
these low - metallicity inflows produce a decrease of @xmath47 dex in the oxygen central abundance while in less gas - rich counterparts , the decrease is of @xmath48 dex towards the final merging stage ( fig . [ fit1 ] ) .
the evolution of the gradients shows an approximately constant slope until the first pericentre where the profiles get more negative as a result of the oxygen carried into the central region by the early metal - rich inflows .
however , from the first pericentre there is a continuous mean flattening of the metallicity profiles which gets even slightly positive around the second pericentre where again , it starts to be get steeper toward more negative values as more oxygen is pumped in by snii produced during the second tidally - induced starburst ( see fig .
[ gasrich1]b ) . :
red and slope : green ) of the oxygen abundance profiles in the more gas - rich encounter ( simvi ) computed in a range of 3@xmath34 kpc .
arrows indicate the first and second pericentres.,width=245 ]
several interesting aspects remain to be understood or even acknowledged in galaxy interactions and their impact on chemical properties .
numerical simulations are , indeed , the most suitable tool to study them .
different authors with different techniques and levels of complexity have made important contributions ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) to this area over the last years .
all works agree to show the central abundance dilution due to low - metallicity inflows triggered during the interaction .
our simulations allow us to contribute further to explore the role played by the induced sf and sn feedback in modulating the strength of this metallicity dilution and the evolution of the metallicity gradients .
particularly , we run simiii which only follows the dynamics of the gas with fixed metallicity assigned according to an initial abundance profile .
no new star formation activity is permitted in simiii . as it can be appreciated by comparing the evolution of the central oxygen abundances in simii and simiii in fig .
[ feedback ] , neglecting the subsequent star formation and chemical enrichment leads to an overestimation of the metallicity dilution ( and flattening of the gradients ) .
one of the effects of star formation is the generation of new sn ii which contributes with new chemical elements and also pumps energy into the ism . the strongest the gas inflows , the strongest the starbursts and consequently , the impact of sn feedback which can heat up and blow away significant fractions of the remaining enriched gas , regulating the subsequent star formation activity and modulating the chemical enrichment . in simiv
we use a lower energy input for sn event and thus , reducing the impact of energy feedback . as a consequence
, a more intense star formation can develop producing a higher chemical enrichment than in simii . from fig .
[ feedback ] we can see that the central oxygen abundance is very similar to that of simii until the first pericentre while after that , the mean oxygen abundance in simiv is larger since the impact of the mass - loaded outflows are weaker and are not so efficient at transporting material outside the central region and at inhibiting the star formation activity ( the effects on the star formation rate , @xmath0-enhancement and metallicity profiles can be visualized comparing results from simii and simiv in figs .
[ alpha1 ] , [ alpha3 ] , and [ fit1 ] ) .
hence , it is important to model the on - going star formation activity in a self - consistent way with energy and chemical sn feedback in order to have a more comprehensive picture of the evolution of the metallicity and gradients during galaxy interactions . as a consequence of this complex interplay of various physical processes , we find that some a priory obvious correlations are actually difficult to detect .
one example of this is the use of @xmath0-enhancement to clock the interaction as we have already discussed in section 3.1 .
another expected correlation is that between the strength of the central starbursts and the metallicity dilution . in fig .
[ sfroh ] we plot the slopes of the metallicity gradients and the central oxygen abundances as a function of star formation rate for wet low - redshift interactions .
we can see that globally there is a correlation so that the highest the star formation rate , the strongest the metallicity dilution and the shallower the abundance profiles .
however , the relations depend strongly on the orbital parameters .
simi and simv have weaker starbursts than simii but nevertheless , they experienced metallicity dilution in the central regions due to the tidally - induced gas inflows . hence , according to our findings , it might be difficult to establish this correlation observationally where galaxy pairs can be observed at different stages of evolution and with a variety of orbital parameters . regarding the gas - richness , from fig.[perfiles ]
we can see that , although there is a global trend in diluting central metallicity and in flattening the mean gradients , there are also fluctuations which correlate with the triggering of inflows and the injection of fresh oxygen by the new sn ii . nevertheless , the mean metallicity dilution is stronger than the less gas - rich counterpart ( fig .
[ fit1 ] ) .
hence , the gas - richness is another key factor .
depending on the astrophysical properties and orbital parameters , galaxies in pairs can be located differently on the mean mass - metallicity relation depending on the stage of the interaction at which they are observed .
this effect might be particularly relevant for the study of high redshift galaxies and could be at the origin of the apparent increase with redshift of the scatter in the mass - metallicity relation .
it could also explain the presence of outliers in the local mass - metallicity relation .
for example , more gas - rich systems observed just as they are first approaching might be dominated by the massively and clumpy star formation activity taking place at all radii , before the interaction can actually trigger larger scale inflows capable to transport low - metallicity gas into the central region .
hence , these galaxies might show an excess of oxygen ( e.g. * ? ? ?
* ) . ) of one galaxy in simii ( black full line ) , simiii (
orange dashed line ) , and sim iv ( red dash - dotted line ) .
the relative distance between the centre of mass of the simulated galaxies is also plotted ( green line ) .
, width=317 ]
by using hydrodynamical simulations of interacting pre - prepared galaxies of comparable masses , we study the chemical evolution of the interstellar medium during the interactions .
our simulations include self - consistently , sn feedback and a detailed chemical model , which allows us to follow the evolution of the gas - phase metallicity from the initial to the final stages of an interaction .
these simulations constitute a unique set to study the effects of interactions on chemical properties .
we analyse the different chemical response to galaxy interactions , varying orbital configurations , strength of energy sn feedback and initial gas fractions in the galaxy discs .
particularly , we focused our study on wet interactions in the local universe . for comparison
, we also analysed a more gas - rich interaction in order to reproduce wet mergers typical of high redshift .
our main results can be summarized as follows : \1 ) we find that a low - metallicity gas inflow developed from the first close passage dilutes the central oxygen abundance and triggers starbursts at the first and second pericentre .
we find that the amplitude of the central oxygen dilution of close galaxy pairs is fairly consistent with observations .
this dilution is observed with a large scatter for interacting systems separated less by than 30 kpc .
the amplitude varies according to orbital configurations and phases of interactions . during the interactions the metallicity gradients get quite shallower with fluctuations , such as
sometimes they are a slightly positive gradient .
both the release of new chemical elements and energy by snii , produced during the tidally - induced starburst , modulates the strength of the central dilution and the metallicity gradients .
this behaviour seems in good agreements with what it is observed for galaxy pairs ( observations by * ? ? ?
* ; * ? ? ?
we note that it is not possible from these simulations to test the claims of @xcite since our boundary conditions are not consistent with a cosmological scenario where gas accretion from the intergalactic medium can fuel the galaxies .
\2 ) we find an increase of the interstellar [ o / fe ] values , observed after the first pericentre as suggested by previous works .
this @xmath0-enhancement does not only reflect the tidally - induced star formation activity but is also affected by galactic outflows and the abundances of the gas inflows . as a consequence ,
we show that the @xmath0-enhancement can only indicate the proximity to the first pericentre and in those encounters with non coplanar retrograde orbits .
after the first pericentre , gas inflows might dilute the central [ o / fe ] values , erasing the trace left by the star formation activity , and restarting the clock to date the interaction .
we also notice that the period of time where this enhancement could be observed is quite short @xmath49 gyr ) .
\3 ) more gas - rich interactions show massive and clumpy star formation on the disc from early stages of the interaction . during the interaction , from the first pericentre to the final stages ,
we find similar behaviours as in the case of the less gas - rich runs with the particularities imprinted by the more gas availability .
gas - richness has an impact on the strength of low - metallicity gas inflows which in the simulated interaction , are able to dilute the central abundances by @xmath50 dex , a factor of 2.5 larger than in its less gas - rich counterpart .
there are important variations in the central abundances and gradients as the interaction of more gas - rich galaxies proceeds due to the competition between various processes , such as the injection of new chemical elements after starbursts , the action of outflows which blow away part of the enriched material , and the gas dynamics of clumpy discs .
\4 ) we find a correlations between the metallicity gradient , the central chemical abundance and the strength of the star formation activity .
however , the slope and scatter of these correlations depend on orbital parameters , sn outflows and gas - richness .
hence , it is not a simple relation which can be easily unveiled observationally .
the metallicity gradients are affected by the strength of the central nuclear starbursts and the extended star formation detected in the outer regions of the galaxies .
star formation in clumps are also found to play a role in modulating the metallicity gradients .
a detailed analysis of disc fragmentation as a function of gas richness will be carried out in a forthcoming paper .
our results suggest that metallicity evolution is closely related to gas dynamics during the encounters . denying processes that affects the gas dynamics and its level of enrichment such as star formation and sn feedback ,
can contribute either to overestimate or underestimate the effects on the central abundance dilution or metallicity gradients .
these findings show that there are still many issues to be addressed in galaxy interactions and their role in chemical evolution .
we thank the referee , f. bournaud , for a constructive report .
pbt thanks isima 2010 for the hospitality and challenging environment which allowed part of this work to be done .
this work was partially supported by the consejo nacional de investigaciones cientficas y tcnicas ( pip 2009/0305 ) , agencia nacional de promocin cientfica y tcnica ( pict 32342 ( 2005 ) ) , max planck 245 ( 2006 ) , the agence nationale de la recherche ( anr-08-blan-0274 - 01 ) . | we analyse a set of galaxy interactions performed by using a self - consistent chemo - hydrodynamical model which includes star formation , supernova feedback and chemical evolution . in agreement with previous works
, we find that tidally - induced low - metallicity gas inflows dilute the central oxygen abundance and contribute to the flattening of the metallicity gradients . the tidally - induced inflows trigger starbursts which increase the impact of sn ii feedback injecting new chemical elements and driving galactic winds which modulate the metallicity distribution .
although @xmath0-enhancement in the central regions is detected as a result of the induced starbursts in agreement with previous works , our simulations suggest that this parameter can only provide a timing of the first pericentre mainly for non - retrograde encounters . in order to reproduce wet major mergers at low and high redshifts , we have run simulations with respectively 20 and 50 percent of the disc in form of gas .
we find that the more gas - rich encounters behave similarly to the less rich ones , between the first and second pericentre where low - metallicity gas inflows are triggered .
however , the higher strength of the inflows triggered in the more gas - rich interactions produces larger metal dilutions factors which are afterward modulated by the new chemical production by supernova .
we find that the more gas - rich interaction develops violent and clumpy star formation triggered by local instabilities all over the disc before the first pericentre , so that if these galaxies were observed at these early stages where no important tidally - induced inflows have been able to develop yet , they would tend to show an excess of oxygen .
we find a global mean correlation of both the central abundances and the gradients with the strength of the star formation activity . however , the correlations are affected by orbital parameters , gas inflows and outflows , suggesting that it might be difficult to determine it from observations .
overall , our findings show that a consistent description of the gas dynamics and stellar evolution along the interactions is necessary to assess their effects on the chemical properties of the interstellar medium .
[ firstpage ] galaxies : formation , galaxies : evolution , galaxies : interactions , galaxies : abundances . |
the thermal conductance of a classical harmonic lattice can be found analytically .
our starting point is the procedure of refs .
we consider the limiting cases of the single - coordinate hamiltonian , eq .
( [ eq : h ] ) , for a lattice of length @xmath42 .
one already saw that the low- ( l ) and high- ( h ) temperature limits can be approximated by a harmonic hamiltonian of the form @xmath11,\ ] ] where @xmath43 . to simplify the analysis ,
the lattice is coupled to two heat reservoirs at the first and last sites , which gives the equations of motion @xmath44.\end{aligned}\ ] ] we choose the spectrum of the dissipation to be ohmic , @xmath45 , with coupling @xmath46 , and the noise to be white noise , @xmath47 , with @xmath48 the low and high reservoir temperatures .
this form for the reservoirs will satisfy the fluctuation - dissipation theorem .
the resulting equations of motion are @xmath49.\end{aligned}\ ] ] the solution for the coordinates has the form @xmath50 where @xmath51 is a vector of length @xmath42 with the first and last components being @xmath52 and the rest being zero .
it represents the coupling of the reservoirs to the ends of the lattice .
the @xmath53 matrix @xmath54 encodes the solution . here
@xmath55 , @xmath56 , and @xmath57 and @xmath58 otherwise .
the heat current flowing into the lattice is @xmath59\dot{y}_{1}(t)\rangle$ ] , where the average is over the noise . setting @xmath60 ,
the heat current becomes @xmath61 where @xmath62 is the temperature difference of the reservoirs , @xmath63 is the cofactor of @xmath64 , and @xmath65 is the determinant of the submatrix of @xmath66 from the @xmath67-th row ( column ) to the @xmath68-th row ( column ) .
it follows that @xmath69 and @xmath70 .
the elements of @xmath71 are given by @xmath72 , where @xmath73 we look for the eigenvalues of @xmath74 of the form @xmath75 with real @xmath76 because those eigenvalues correspond to propagating modes .
this requirement imposes that @xmath77 . after changing variables from @xmath78
to @xmath76 that satisfy this constraint , the final expression ( for an infinite lattice ( @xmath79 ) ) is @xmath80}.\ ] ] this gives , explicitly for the low and high temperature thermal conductance @xmath81 , @xmath82,\label{eq : fullkappa}\ ] ] with @xmath83 with these expressions one can explicitly find the thermal conductance ratio @xmath18 .
we have verified that for reservoirs contacted to a single site on each end , the thermal conductance from our numerical simulations agree with our analytic formula to within @xmath84% , which we attribute to finite size effects in the numerical simulations .
we can take various limiting forms of these equations . if we define the prefactor as @xmath85 and a dimensionless reservoir coupling as @xmath86 the expressions for the conductance become @xmath87 , \label{eq : tcond}\ ] ] with @xmath88 the appropriate limiting forms for our case are the following .
when the high temperature harmonic limit has no onsite potential , then the heat conductance becomes @xmath89.\ ] ] for a low temperature limit that has a much greater onsite term than the nearest neighbor coupling , i.e. , @xmath90 , the heat conductance becomes @xmath91 which also assumes that the dimensionless coupling to the reservoirs is @xmath92 . for strong coupling to the reservoirs , the ratio becomes @xmath93 this is the analytic expression we use within the article .
the strong coupling limit gives the extreme value of @xmath18 .
now we show that the characteristic frequency in the pbd model is lowered as @xmath24 crosses @xmath13 from below . for the low temperature hamiltonian ,
the corresponding equation of motion is @xmath94\}.\ ] ] from the ansatz @xmath95 , one obtains the phonon spectrum as @xmath96.\ ] ] thus , the frequency band of phonons is @xmath97/m}$ ] . for the high temperature hamiltonian ,
the equation of motion is @xmath98\ ] ] and the phonon spectrum is @xmath99.\ ] ] the frequency band of phonons is @xmath100 .
the two limiting hamiltonians are both harmonic and the characteristic frequency is indeed lowered , and similar considerations apply for other models . in the low temperature limit , the onsite potential stiffens the dna compared to the high temperature limit , which results in the raising of the phonon spectrum of the low temperature limit compared to the high temperature . in the latter ,
as well , the nearest neighbor coupling drops from @xmath101 to @xmath102 shrinking the bandwidth
. this trade - off is responsible for the change in thermal conductance across the transition . if the drop in nearest neighbor coupling is small , then the softening will dominate , and the heat conductance will increase because these softened modes can conduct heat more effectively .
to study the dynamics of the dna out of equilibrium we solve numerically the langevin equation , which describes the dynamics of a hamiltonian system in the presence of thermal baths .
the langevin equation is given by @xmath103 where @xmath104 and @xmath105 are the potentials described in eq .
( [ eq : h ] ) .
the dna strand is split into three regions , the two ends , each of length @xmath106 , serve as the langevin thermal reservoirs at temperatures @xmath107 and @xmath108 .
this means that the friction term @xmath109 only operates for @xmath67 within the thermal reservoirs .
the fluctuating term @xmath110 is gaussian white noise which obeys the fluctuation - dissipation relation @xmath111 for the low and high temperature reservoirs , respectively .
the middle region of the length @xmath112 is the free dna strand , which is driven out of equilibrium by the langevin reservoirs when @xmath113 .
the parameters for the simulations are @xmath114 and @xmath115 , and the pbd model parameters are @xmath116 , @xmath117 , @xmath118 , @xmath119 , @xmath120 , and @xmath121 @xcite .
the strand is homogeneous dna except for the random forces and friction in the langevin regions .
the equations of motion are integrated with the fourth - order runge - kutta method .
the temperature profile of the strand is evaluated by defining the local temperature at site @xmath67 as @xmath122 , where the average is over time .
the local heat current is given by @xmath123 .
the simulations are performed long enough to allow the system to reach its steady state .
the simulations of the heat capacity were performed by connecting langevin reservoirs of identical temperature to every site in the chain and evaluating the average energy per site by @xmath124 , where the averaging is performed over time and all the sites of the chain .
then , the dependence of average energy per site on temperature , @xmath125 , obtained by scanning through a set of temperature points , was numerically differentiated to yield the heat capacity .
we also performed simulations on the jb model @xcite , which has the same onsite potential as the pbd model and uses the nearest neighbor potential @xmath126 the parameters in this interaction are @xmath127 , @xmath128 , and @xmath129 .
thus , the low temperature limit is @xmath130 , with @xmath131 .
the coupling @xmath132 is not substantially different from the pbd model , which has @xmath133 .
however , numerical simulations indicate that in the high temperature limit the heat conductance converges to the conductance of a harmonic hamiltonian with nearest neighbor interaction @xmath134 . in other words ,
the quantity @xmath135 is sufficiently large to suppress the second term in eq .
( [ eq : jb ] ) .
thus , the high temperature effective nearest neighbor coupling is @xmath136 , which is substantially smaller than the low temperature coupling , and also substantially smaller than the high temperature coupling of the pbd model . | dna denaturation has been a subject of intense study due to its relationship to dna transcription and its fundamental importance as a nonlinear structural transition .
many aspects of this phenomenon , however , remain poorly understood .
existing models fit quite well with experimental results on the fraction of unbound base pairs versus temperature , but yield incorrect results for other essential quantities , such as the base pair fluctuation timescales . here
, we demonstrate that nanoscale thermal transport can serve as a sensitive probe of the underlying microscopic physics responsible for the dynamics of dna denaturation .
specifically , we show that the heat transport properties of dna are altered significantly as it denatures , and this alteration encodes detailed information on the dynamics of thermal fluctuations and their interaction along the strand .
this finding allows for the discrimination between models of dna denaturation and will help shed new light on the nonlinear vibrational dynamics of this important molecule . #
1#1 # 1|#1 # 1#1| # 1#2|#1#2| # 1#2#1|#2 # 1#2 # 1#2 # 1#2 besides being the `` molecule of life '' or perhaps because of it dna lives at a unique position where its double - stranded structure can unravel into two single strands , i.e. , it denaturates or `` melts '' , via changes in conditions such as temperature or ionic concentration @xcite .
local melting of dna due to thermal fluctuations , which can occur well below the denaturation temperature , is thought to play a major role in the formation of the transcription bubble @xcite .
the denaturation of dna proceeds via the thermal breakage ( dissociation ) of base pairs and a nonlinear interaction between base pairs sharpens this transition via cooperative effects in base pair unbinding @xcite . as a result , the accurate description of mechanisms of dna denaturation requires one to not only correctly account for local thermal fluctuations but also how these fluctuations interact along the strand .
these processes are typically probed _ indirectly _ through the measurement of the dna melting curve the fraction of unbound base pairs versus temperature at equilibrium thus hindering the understanding of the denaturation mechanisms . however , thermal fluctuations and their interaction along the strand are the same processes that occur in thermal transport , which , as we shall see , can give an independent and more direct assessment of the proposed dna denaturation mechanisms . during the past decade ,
there has been tremendous progress in measuring the thermal and electronic properties of molecular systems @xcite , including a recent experiment on the heat conduction of inhomogeneous dna - gold nanocomposites @xcite .
however , except for few studies relevant to double - stranded dna @xcite , no existing analysis of thermal transport takes into account dna s large structural fluctuations that eventually result in its denaturation .
schematic of dna between two heat reservoirs that probe its structure via an energy current.,width=264 ] in this paper , the non - equilibrium behavior of single - coordinate nonlinear models of dna are analyzed as heat is driven through dna via two thermal reservoirs , as shown schematically in fig .
[ fig : schematic ] .
the preeminent single - coordinate model for thermally driven denaturation of dna is the peyrard - bishop - dauxois ( pbd ) model @xcite , which is believed to capture the main physical processes behind dna denaturation @xcite . by first deriving an analytic expression for the thermal conductance , @xmath0 ,
we show that the pbd model predicts a substantial jump in @xmath0 upon dna denaturation .
that is , the conductance of dna is predicted to actually _ increase _ , despite the fact that the dna becomes disordered as it is broken into two single strands .
this change is a delicate balance of the model parameters and the mechanisms that it represents .
indeed , we show that another related model @xcite , which also can describe the statistical properties of dna denaturation , gives qualitatively different non - equilibrium behavior , predicting a drop in @xmath0 .
thus , by driving dna out of equilibrium , the basic physical aspects of denaturation thermal fluctuations and their coupling along the molecule can be probed .
we begin with a description of the pbd model @xcite .
this model , as well as other single - coordinate models for dna , reduces the dna strand to a one - dimensional , classical lattice with each base pair represented by a single coordinate the collective stretching of the hydrogen bonds in a watson - crick pair .
the pbd hamiltonian takes on the form @xmath1,\label{eq : h}\ ] ] where @xmath2 represents the stretching coordinate within the @xmath3 base pair .
an on - site morse potential , @xmath4 , describes the harmonic binding of the base pair at small @xmath2 with the angular frequency @xmath5 , and unbinding at @xmath6 with the finite dissociation energy @xmath7 .
the pbd nearest - neighbor potential is given by @xmath8 the pbd model has been quite successful in characterizing statistical properties of dna near the denaturation transition @xcite .
however , since it was not derived from a microscopic hamiltonian , one has to validate the model by comparison with experimental results . in this regard
, there is a considerable ambiguity in the results for short dna strands @xcite as well as the timescales for the opening of bubbles at room temperature @xcite .
thermal transport , as we show below , can determine some of the main features single - coordinate models should display ( beyond equilibrium denaturation curves ) and help settle the above discrepancies .
we start our analysis by examining the behavior of the pbd model in the low ( @xmath9 ) and high ( @xmath10 ) temperature limits , where it becomes harmonic with the effective hamiltonian @xmath11,\label{eq : hhl}\ ] ] with @xmath12 .
numerical studies show that the full pbd model , described by the hamiltonian in eq .
( [ eq : h ] ) , undergoes a sharp transition corresponding to the denaturation of dna @xcite . above the transition temperature , @xmath13 ,
the base pairs become unbound , i.e. , the on - site potential approaches a constant leading to @xmath14 .
the nearest - neighbor potential becomes harmonic with @xmath15 . in the low temperature limit
, the strand will again have harmonic behavior with the parameters of the effective hamiltonian @xmath16 and @xmath17 ( the latter reflects the larger nearest - neighbor coupling of the low temperature strand ) . even at low - temperatures ,
the nonlinearity of the potential is important , but the harmonic limit gives the correct physical interpretation of the numerical results on the full model we present later .
we introduce the concept of the thermal conductance ratio , @xmath18 , as the ratio of the conductance between the high and low temperature phases @xmath19 . for ohmic heat reservoirs
strongly coupled to the first and last site of the strand , it can be calculated to be ( see appendix [ app : analytics ] and refs .
@xcite for details ) @xmath20 which uses the lowest order term in @xmath21 for @xmath22 ( see eq .
( [ eq : tcond ] ) in appendix [ app : analytics ] ) . since eq .
( [ eq : r ] ) is for strong coupling , it represents an extreme value for @xmath18 . for the pbd model with the original parameters from ref .
@xcite , the corresponding ratio is @xmath23 @xcite .
that is , the transition from low to high @xmath24 results in a substantial jump in @xmath0 due to the weakening of the on - site confining potential , i.e. , the base pair dissociation , as @xmath24 crosses @xmath13 from below .
the removal of the on - site potential softens the modes of the strand , making them more efficient in the transport of heat .
this is more transparent if one examines the low temperature conductance , @xmath25 where @xmath26 is the on - site frequency and we have taken @xmath21 to be small .
equation ( [ eq : kl ] ) shows that the conductance increases with decreasing frequency .
we note that the pbd model give a novel phenomenon compared to other transitions .
for example , in the ice - water transition at 273 k , the conductance decreases , rather than increases , with a ratio of @xmath27 @xcite .
this is mainly due to the breaking of crystal structure across the transition that reduces the coupling between molecules .
another example is single - crystal c@xmath28 , which exhibits a face - centered - cubic to simple - cubic structural transition at 260 k @xcite .
the ratio for this transition can be estimated from the experimental data to be @xmath29 .
the thermal conductance ratio is thus important in identifying the underlying mechanisms of a structural transition .
the balance of the softening of the lattice and the reduction of the nearest neighbor coupling determines both the magnitude of the conductance change , as well as whether it increases or decreases .
different proposals for the nonlinear forms of the model will give different limiting harmonic forms , changing the conductance behavior across the temperature range not only quantitatively , but also qualitatively .
for example , another single - coordinate model of dna is that of joyeux and buyukdagli ( jb ) in ref .
@xcite .
it uses the same on - site potential but employs an alternative nearest - neighbor potential that , although giving a very similar melting curve , results in a different high temperature harmonic hamiltonian ( see appendix [ app : analytics ] ) . in the strong - coupling limit
, @xmath18 changes as @xmath30 that is , the pbd model gives qualitatively distinct behavior from the jb model because the latter looses virtually all of its nearest neighbor coupling at high temperature , i.e. , it looks more like the ice - to - water transition in character .
this qualitative conclusion does not change even when using different parameter sets that have been developed @xcite .
therefore , while they both agree well with some denaturation experiments , the conductance ratio can be used to directly probe the physics contained within their respective interaction potentials . while temperatures much higher than the transition temperature are beyond the validity of the models , the opposing predictions persist within a range of temperatures , including around the denaturation transition .
in addition , having the ends of the dna `` clamped '' together on each of the heat reservoirs one way to measure its thermal transport properties will help the fluctuations described by these single - coordinate models to persist to higher temperatures .
this strongly supports using the heat conduction as a test to determine basic aspects of dna denaturation that need to be represented in microscopic models . to gain further insight into the non - equilibrium behavior of dna ,
we solve for the full dynamics of the pbd model ( eq . ( [ eq : h ] ) with the appropriate stacking interaction ) by numerically integrating the equations of motion of the dna between two langevin reservoirs ( see appendix [ app : numerics ] for details on numerical simulations ) . in fig .
[ fig : thermalcurrent](a ) , we plot the thermal current as a function of the average reservoir temperature , while keeping the temperature difference constant at a value significantly lower than the width of the phase transition so that the entire strand is always in one phase . thermal transport across the denaturation transition .
( a ) thermal current through dna as a function of average temperature , @xmath31 , calculated while keeping the temperature difference @xmath32 constant .
the thermal current displays a steep rise as the temperature crosses the transition temperature @xmath33 .
low and high temperature simulations of the harmonic strands defined by @xmath34 and @xmath35 are given by blue and red lines , respectively .
the numerical results on the full pbd model approach the harmonic limits at sufficiently low and high temperatures .
( b ) logarithmic scale plot of the thermal current . the convergence to the harmonic limit , as well as an non - monotonic dependence of the thermal current on temperature at low temperatures and a sharp rise near the transition are visible .
the latter indicates that dna may be suitable to function as a molecular thermal switch .
( c ) dependence of equilibrium heat capacity , @xmath36 , ( normalized per length of the strand ) on temperature showing the transition at @xmath13 and the convergence to the harmonic value @xmath37 .
the convergence to the harmonic value at high temperature is much more rapid than a low temperature , indicating that the anharmonicity of the morse potential is still important even at low temperature.,width=302 ] as seen fig .
[ fig : thermalcurrent ] , the ability of the dna to conduct heat increases substantially after it becomes denaturated according to the pbd model , thus confirming our analytical prediction . *
* further , the conductivity at @xmath38 is 0.18 pw / k smaller than that in ref .
@xcite .
the difference is due to different interactions taken into account , which result in an absence of any signatures of denaturation up to 350 k in ref .
@xcite .
thus , their results do not incorporate the effect of strong nonlinearities due to denaturation / local melting . *
* we have also calculated the thermal current for purely harmonic strands with the parameters corresponding to the different phases ( as discussed below eq .
( [ eq : hhl ] ) ) .
the thermal conductance ratio for the harmonic limits is found numerically to be @xmath39 , which is lower than our analytical result due to the more realistic reservoir implemented in our simulations ( see appendix [ app : numerics ] ) .
figure [ fig : thermalcurrent](c ) plots the heat capacity @xmath36 ( per base pair ) as a function of temperature , indicating the denaturation transition and the approach to harmonicity ( where @xmath40 is always unity ) .
the nonlinear model , however , gives much richer behavior .
while at high temperature there is convergence of the nonlinear model to its harmonic limit , at low temperatures a non - monoticity appears , more clearly seen in fig .
[ fig : thermalcurrent](b ) where we plot the thermal current as a function of temperature on a logarithmic scale .
the non - monotonicity , reflected in the drop of @xmath0 in the temperature range @xmath41 k , is mainly due to phonon scattering off the anharmonic on - site potential .
this signature conductance versus temperature indicates the strength of the nonlinearity in dna dynamics , which we will investigate further in a future contribution . although the single - coordinate models we study here were proposed to study dna behavior near the denaturation transition , such a nonlinear signature should still appear at low temperature , implying that a measurement of dna thermal transport could also serve to define the temperature range of validity of the pbd model .
it also gives rise to a three orders of magnitude increase of the thermal conductance , indicating that this characteristic of structural phase transitions can be useful in developing heattronic devices such as thermal switches and thermal transistors @xcite . in the above discussion we have considered only homogeneous dna strands .
it is clear that heterogeneity will give rise to interesting effects in the conductance .
alternating sequences of at and gc pairs , for instance , will tend to suppress the low temperature conductance due to the mismatch between parameters of base pair binding potentials , while the high temperature conductance will remain roughly the same .
this will result in a substantial increase in @xmath18 @xcite , which will also occur due to random or genomic dna .
differences between at and gc pairs gives an additional experimental test of this class of models @xcite .
to conclude , we have investigated the out - of - equilibrium properties of single - coordinate models of dna in order to elucidate the nonlinear nature of denaturation .
the balance of lattice softening and reduced coupling in different models leads to qualitatively distinct predictions for the thermal conductance ratio the experimental measurement of which will illuminate these fundamental aspects of dna and thus should shed light on the underlying mechanisms of dna denaturation . due to the nature of the effect
, experiments will be robust with respect to heterogeneity , defects , and contacts , unlike the electrical counterpart @xcite .
recent experiments on the heat conduction of nanotubes @xcite and dna - gold nanocomposites @xcite give a potential route to developing the experimental setup .
however , it will be necessary to perform the experiment in conditions of high humidity so the dna retains its b - form @xcite . such an experiment
could also be done on double- and single- stranded dna separately , which would remove uncertainties due to the lack of solvent .
fluorescence experiments to detect local openings of the dna strand @xcite can further complement the thermal transport measurements .
the thermal conductance ratio may also be useful in understanding other structural transitions , such as nanotube collapse .
finally , such experiments will pave the road to creating molecular thermal switches and open new avenues in understanding and controlling thermal flow at the nanoscale .
the authors would like to thank a. r. bishop , k. rasmussen and b. alexandrov for valuable discussions .
this research is supported by the u.s .
department of energy through the lanl / ldrd program .
k.a.v . also acknowledges the support provided by cnls . |
the response of physical systems to a change of external conditions is of eminent importance in physics .
in particular the dependence of expectation values on temperature , the particle density , the space region , the imposed boundary conditions or external fields has been widely studied @xcite .
nevertheless , many properties of such systems are poorly understood .
the massless thirring model @xcite , which is among the simplest interacting field theories , has already led to considerable confusion about its thermodynamic properties in the literature @xcite .
the reason is two - fold : firstly , the computation of the fermionic determinant in the presence of a chemical potential and/or non - trivial boundary conditions is delicate , because the eigenvalues of the dirac operator are generically complex . in section
@xmath4 we propose a regularization scheme via analytic continuation .
we argue that the so - obtained determinant , which differs from previous results @xcite , leads to the correct equation of state.the second complication originates in the infrared - sector .
an elegant infrared regularization , which is particularly well suited for the study of thermodynamic properties , is to quantize the model on a torus .
harmonic contributions to the current arise then naturally and taking them into account turns out to be crucial for a correct quantization . in particular
the so - obtained results differ from those gotten earlier @xcite using bosonization .
this is explained in section @xmath5 . on another front
there has been much effort to quantize self - interacting field theories in a background gravitational field @xcite .
for example , one is interested whether a black hole still emits thermal radiation when self - interaction is included . due to general arguments by gibbons and perry @xcite this question
is intimately connected with the universality of the second law of thermodynamics . the thirring model ( including the gauged version of it )
is still solvable in curved space - time and we can study its properties in a background gravitational field .
this provides us in particular with an elegant approach to the study of its conformal structure : correlation functions with current- and stress - tensor insertions , which are gotten by functional differentiation with respect to the gauge- and gravitational fields , contain the necessary information to characterize the underlying symmetry algebras . to familiarize the reader with our approach
we first rederive the conformal structure of the original thirring model in section @xmath6 .
we then show how a non - minimal coupling to gravity leads in a natural way to a modification of the conformal structure .
in particular , very much as for a free scalar field the central charge in the fermionic formulation of the thirring model is not unique .
furthermore , we find that the equivalence between finite size scaling and central charge of the virasoro algebra holds only for a particular treatment of the zero - mode sector in which a charge at infinity is generated automatically .
this charge combines in a non - trivial way with the weyl - anomaly of the determinant of the fluctuation operators to reconcile the equivalence of the finite size scaling and the central charge . for certain values of the non - minimal coupling
we obtain minimal models from interacting fermions .
this is the subject of sections 3.4 and 3.5 . the gauged thirring model , which contains the schwinger model ( @xmath7 ) as a particular limit , is no longer conformally invariant but has a mass gap : the photon acquires a mass @xmath8 via the schwinger mechanism .
it possesses a non - trivial vacuum structure which promotes it to an attractive toy model to mimic the complex vacuum structure in @xmath9-d gauge theories . from our experience with the schwinger model @xcite , which is supposed to share certain aspects with one - flavour @xmath10 @xcite
, we expect that gauge fields with winding numbers are responsible for the non - vanishing chiral condensate and in particular its temperature dependence .
configurations with windings , so called instantons , exist only for finite volumes and minimize the euclidean action .
they lead to chirality violating vacuum expectation values .
for example , a non - zero chiral condensate develops which only for high temperature and large curvature vanishes exponentially . since for particular choices of the coupling constants the model reduces to well known and well studied exactly soluble models there are many earlier works which are related to ours .
some of them concentrated more on the gauge sector and investigated the renormalization of the electric charge by the four - fermi interaction @xcite or the non - trivial vacuum structure in the schwinger model @xcite .
others concentrated on the un - gauged conformal sector .
freedman and pilch calculated the partition function of the un - gauged thirring model on arbitrary riemann surfaces @xcite .
we do not agree with their result and in particular show that there is no holomorphic factorization for general fermionic boundary conditions .
also we deviate from destri and devega @xcite which investigated the un - gauged model on the cylinder with twisted boundary conditions .
we comment on these discrepancies in section @xmath5.section @xmath11 contains introductory material and in particular the classical structure of the model .
other papers are dealing with different aspects of certain limiting cases of the model considered here .
in particular in @xcite , the thermodynamics of the thirring model has been studied and the hawking radiation has been derived in @xcite .
the equivalence of the massive thirring model and the sine - gordon model in curved space has been shown in @xcite .
partition functions for scalar fields with twisted boundary conditions have been computed in @xcite and more recently in @xcite .
the gauged thirring model in curved space - time has the lagrangian where the gamma - matrices in curved space - time are related to the ones in minkowski space - time as @xmath12 , @xmath13 is the coordinate- and gauge covariant derivative and @xmath14 is the electromagnetic field strength . the gravitational field @xmath15 ( or rather the @xmath11-bein @xmath16 , since the theory contains fermions ) is treated as classical background field , whereas the photons @xmath17 and fermions @xmath18 will be quantized .
the classical theory is invariant under @xmath19 gauge- and axial transformations and correspondingly possesses conserved vector and axial - vector currents here @xmath20 denotes the totally antisymmetric tensor .
in fact , the conservation laws together with the relation ( [ t3 ] ) between the vector- and axial currents imply that the currents are free fields which is the reason that accounts for the solubility of the model @xcite , even in the presence of gauge- and gravitational fields .
of course , for any gauge invariant regularization the axial current possesses an anomalous divergence in the quantized model .
thus the normal @xmath21 ward identities in the un - gauged thirring model @xcite become anomalous when the fermions couple to a gauge field .
the solution of the equations of motion is most easily presented by introducing auxiliary scalar- and pseudo - scalar fields , in terms of which the action takes the form note that for later use we have allowed for different couplings of the fermionic currents to the scalar- and pseudo - scalar auxiliary fields @xmath22 and @xmath23 , respectively .
the original thirring model is recovered for @xmath24 , since then is classically and quantum - mechanically equivalent to ( [ t13 ] ) , after elimination of the multiplier field @xmath25 . by decomposing the gauge field similarly as the @xmath25-field as and choosing isothermal coordinates for which @xmath26 , the generalized dirac operator reads hence , if @xmath27 solves the free dirac equation in flat minkowski space time
, then solves the dirac equation of the interacting theory on curved space - time .
the vector currents are related as @xmath28 the same relation holds for the axial vector current .
diffeomorphism invariance then leads covariantly conserved energy - moment um tensor applying the variational identities in appendix a one obtains after a lengthy but straightforward computation @xmath29\nonumber\\ & & \quad+2 \nabla^\mu\phi\nabla^\nu\phi - g^{\mu\nu } \nabla^\al\phi\nabla_\al \phi \quad + \quad ( \phi \leftrightarrow \lam ) \label{t12}\\ & & \quad+\ha j^{\mu}\;(g_{{\sy 1}}\nabla^{\nu}\lam- g_{{\sy 2}}\eta^{\nu \al } \nabla_\al \phi)\quad+\quad ( \mu \leftrightarrow\nu)\nonumber\\ & & \quad + g_{{\sy 2 } } g^{\mu\nu } j^\al\eta_{\al\beta}\nabla^\beta\phi-2 g_{{\sy 2 } } j^\al\eta_\al\,^{(\mu}\nabla^{\nu)}\phi \ ; , \nonumber\end{aligned}\ ] ] where we have introduced the symmetrization @xmath30 .
the first two lines are just the energy momentum of the electromagnetic field , charged fermions and free neutral ( pseudo- ) scalars .
the remaining terms reflect the interaction between the fermionic and auxiliary fields . on shell @xmath31
is conserved as required by general covariance .
using the field equations for @xmath18 and @xmath22 its trace reads in particular for @xmath32 it vanishes , and the theory becomes weyl - invariant .
[ [ symplectic - structure ] ] symplectic structure : + + + + + + + + + + + + + + + + + + + + + in the presence of both fermions and bosons it is convenient to exploit the graded poisson structure @xcite @xmath33 the sum is over all fundamental fields @xmath34 in the theory .
the sign is minus if one or both of the fields @xmath35 and @xmath36 are bosonic ( even ) and it is plus if both are fermionic ( odd ) fields .
the momentum densities @xmath37 conjugate to the @xmath38-fields are given by functional left - derivatives .
a simple calculation yields the following momenta @xmath39 in the following sections we are lead to consider the _ euclidean version _ of the model .
then one must replace the lorentzian @xmath40 and @xmath41 by their euclidean counterparts .
for example , with our conventions the relation ( [ t3 ] ) becomes @xmath42 and the generalized dirac operator in euclidean space - time becomes ( see ( [ t5 ] ) for the definition of @xmath43 and @xmath44 ) , instead of ( [ t5 ] ) .
also , to recover the euclidean thirring model as particular limit of ( [ t1 ] ) we must set @xmath45 .
in this section we analyze the quantum theory corresponding to the classical action ( [ t1 ] ) without gauge fields , in flat and curved space - time .
the gauged model is then considered in the next section . here
we calculate the _ partition function , ground state energy , equation of state _ and determine the _ conformal structure _ of the un - gauged model . to allow for a non - vanishing u(1)-charge we couple this
conserved charge to a chemical potential @xmath46 .
for the finite temperature model the imaginary time must vary from zero to the inverse temperature @xmath47 and the bosonic and fermionic fields must obey periodic and anti - periodic boundary conditions , respectively .
we enclose the system in a spatial box with length @xmath48 to avoid infrared divergences .
we shall determine the dependence of the partition function and correlators on the metric .
this provides us with an alternative approach to the conformal structure and its relation to finite size - effects .
also , it enables us to study the effect of non - minimal coupling to gravity in section @xmath6 .
hence we allow for an arbitrary metric or @xmath11-bein @xmath49 with euclidean signature
. we can choose ( quasi ) isothermal coordinates and a lorentz frame such that where @xmath50 is the teichmueller parameter and @xmath51 the gravitational liouville field .
space - time is then a square of length @xmath48 and has volume @xmath52 .
we allow for the general twisted boundary conditions for the fermions the parameters @xmath53 and @xmath54 represent _ vectorial and chiral twists _ , respectively .
we could allow for twisted boundary conditions for the ( pseudo ) scalars as well @xcite , e.g. @xmath55 .
however , to recover the thirring model for equal couplings we must assume that these fields are periodic .
for @xmath56 , @xmath57 in which case @xmath58 , and for @xmath59 the partition function has the usual thermodynamical interpretation .
its logarithm is proportional to the free energy at temperature @xmath60 .
twisted boundary conditions as in ( [ t518 ] ) require some care in the fermionic path integral .
the subtleties are not related to the unavoidable ultra - violet divergences but to the transition from minkowski- to euclidean space - time . to see that more clearly let @xmath61 denote the space of fermionic fields in _ minkowski space - time _ with chirality @xmath62 . since both the commutation relations and the action do not connect @xmath63 and @xmath64 we can consistently impose different boundary conditions on @xmath63 and @xmath64 . on the other hand , in the _ euclidean path - integral _ for the generating functional the dirac operator @xmath65 exchanges the two chiral components of @xmath18 , i.e. @xmath66 .
thus , in contrast to the situation in minkowski space the two chiral sectors are related in the action . of course , the eigenvalue problem for @xmath67 is then not well defined .
this is the origin of the ambiguity in the definition of the determinant .
it is related to the ambiguities one encounters when one quantizes chiral fermions @xcite .
here we reformulate this problem in such a way that the determinant with chiral twists ( @xmath68 ) can be obtained by analytic continuation .
the resulting determinants do not factorize into ( anti- ) holomorphic pieces . in appendix
b we give further arguments in favour of our result by calculating the determinants in a different way .
let us now study the generating functional for fermions in an external gravitational and auxiliary field .
for that we observe that on the torus we must add a harmonic piece to the auxiliary fields to which the fermionic current couples in ( [ t1 ] ) .
more precisely , in the hodge - decomposition of @xmath25 in ( [ t1a ] ) contains a harmonic piece , more generally , allowing for arbitrary couplings of the various terms in ( [ help ] ) to the fermionic current , we are led to add a term @xmath69 to the action ( [ t1 ] ) .
note , that in isothermal coordinates , for which the metric has the form ( [ t517 ] ) , the harmonics @xmath70 are constant . the constant @xmath70 couple to the harmonic part of the current and
are needed to recover the thirring model in the limit @xmath71 .
also , we shall see that the harmonic degrees of freedom are essential to obtain the correct thermodynamic potential.finally we introduce a _ chemical potential _ for the conserved @xmath19 charge . in the euclidean functional approach
this is equivalent to coupling the fermions to a constant imaginary gauge potential @xmath72 @xcite.as a consequence of the above observations the scaling formula ( [ scal1 ] ) ( recall , that @xmath73 and @xmath74 when the electromagnetic interaction is switched off ) is modified to this scaling property will enable us to relate the fermionic determinants and green s functions of @xmath75 and @xmath76 .
the spin connection @xmath77 in ( [ t28 ] ) vanishes for our choice of the reference zweibein .
the dependence of @xmath76 on the chemical potential @xmath46 and the constant harmonic field @xmath70 can not be gotten by the anomaly equation @xcite .
it must be computed by direct methods .
for this we expand the fermionic field in a orthonormal basis of the hilbert space where @xmath78 are independent grassmann variables .
a basis is given by and the @xmath79 are the eigenvectors of @xmath80 . recall that @xmath53 and @xmath54 represent the vectorial- and chiral twists ( [ t518 ] ) respectively .
the @xmath81 and @xmath82 must obey the @xmath83 and @xmath84 boundary conditions , respectively .
these boundary conditions fix the admissible momenta @xmath85 in ( [ dett30 ] ) .
since the dirac operator maps @xmath86 into @xmath87 the @xmath88 must then obey the same boundary conditions as the @xmath89 .
thus @xmath90 is obtained from @xmath91 by exchanging @xmath92 and @xmath93 .
it follows then that with here we have introduced @xmath94 . to continue we recast the infinite product for the determinant in the form where @xmath95 is the inverse of the reference metric ( [ t517 ] ) and the logarithm of the product ( [ inf ] ) can in turn be written as the derivative at zero argument of a generalized zeta function .
indeed one easily verifies that for we have ( formally ) however @xmath96 is divergent for @xmath97 .
these divergences can be regularized as follows : we compute @xmath96 for @xmath98 and subsequently define its value for @xmath99 by analytic continuation.assume for the moment that @xmath100 is real or equivalently that there are no chiral twists @xmath101 and chemical potential @xmath46 .
then @xmath96 has a well defined analytic continuation to @xmath102 via a poisson resummation @xcite . indeed , writing @xmath96 as a mellin transform the generalized poisson resummation formula applied to the integrand in ( [ det5 ] ) yields after integration over @xmath103 the zero mode with @xmath104
is eliminated because for @xmath105 it does not contribute .
after this analytic continuation @xmath96 and @xmath106 are now regular at @xmath107 .
more precisely @xmath108 and here we made use of @xmath109\es \det[(i\hat\di)^2]$ ] , which follows from @xmath110.for complex @xmath100 the poisson resummation is not applicable and @xmath111 can not be calculated by direct means . to circumvent these difficulties
we note that the infinite sum ( [ det34 ] ) defining the @xmath112-function for @xmath113 is a mereomorphic function in @xmath114 .
thus we may first continue to @xmath102 for real @xmath100 and then continue the result to complex values . using the transformation properties of theta functions the resulting determinant can be written as this is the main result of this section.it can be shown that this determinant is _ gauge invariant _ , i.e. invariant under @xmath115 , but not invariant under chiral transformations , @xmath116 , as expected .
furthermore , it transforms covariantly under modular transformations @xmath117 and @xmath118 . in other words
, @xmath119 is invariant under modular transformations if at the same time the boundary conditions are transformed accordingly .
the exponential prefactor is needed for modular covariance and is not present in the literature @xcite .
it correlates the two chiral sectors and will have important consequences . in appendix
b we confirm ( [ t36 ] ) with operator methods .
the last step in the calculation of the fermionic generating functional is the inclusion of the local contributions to the auxiliary- and metric field , i.e. the dependence of the determinant on @xmath22 , @xmath23 and @xmath120 . for this
we introduce the one - parameter family of dirac operators we take the @xmath121-dependence of the metric as @xmath122 . with @xmath123 as defined in ( [ t28 ] ) , this family interpolates between @xmath76 and @xmath75 .
the determinant of the full dirac operator is then obtained by integrating the corresponding anomaly equation @xcite : where is the _ liouville action . _ in deriving this result we assumed that @xmath124 .
this constraint on the zero - mode of @xmath22 ( and similarly of @xmath23 ) will be discussed below .
actually , for our reference metric the ricci scalar @xmath125 vanishes and the liouville action simplifies to @xmath126 .
however , the above formulae hold for arbitrary reference metrics and arbitrary riemannian surfaces . furthermore , as expected for a gauge - invariant regularization , the function @xmath22 and thus the longitudinal part of @xmath25 does not appear in the determinant . to complete the calculation of the generating functional we need to know the fermionic green - functions @xmath127 .
using the scaling property of the dirac operator , eq .
( [ t539a ] ) , it is easy to see that in an arbitrary background field @xmath127 is related to @xmath128 by @xmath129 together with the relation ( [ t543 ] ) and the explicit form ( [ t35],[t36 ] ) for @xmath119 this yields the fermionic generating functional by using the scaling properties of the ricci - scalar and laplacian ( see appendix a ) the exponent can be written in a manifest diffeomorphism - invariant way as @xmath130 here we used that on the torus @xmath131 integrates to zero . on the sphere or higher genus surfaces
the last formula is modified.the integration over the auxiliary fields then leads to the full generating functional of the thirring model .
it contains all information about the thermodynamic- and conformal properties .
this is the subject of the next two sections . in this chapter
we derive the grand canonical potential , equation of state and ground state energy for the thirring model . for this
we need to compute the partition function where @xmath132 is the fermionic generating functional ( [ t46 ] ) and @xmath133 the bosonic action as it stands the partition function is still ill - defined unless we constrain the zero - modes artificially introduced in the hodge decomposition of @xmath25 in ( [ help ] ) .
the choice of the constraints is restricted by the symmetries of the system .
in particular translation invariance ( or rotation invariance on the sphere ) and covariance under modular transformations of the torus are symmetries which me may want to preserve by the zero - mode constraint .
the constraint measure ( and similarly for @xmath134 ) satisfies these requirements ( the normalization by the volume in the definition of @xmath135 is needed such that the constraints and hence the partition function are both dimensionless ) .
for example , one finds the dimensionless partition function for free bosons , where the prime indicates the omission of the zero - eigenvalue .
[ [ integration - over - the - harmonics ] ] integration over the harmonics : + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + there is no restriction on the harmonic parts of the auxiliary fields and the gaussian integral yields where @xmath136(\lambda)=\sum\limits_{n\in z^2 } \,e^{i\pi(n+u)\lambda ( n+u)+2\pi i ( n+u)w}\end{aligned}\ ] ] is the theta function with characteristics and covariance [ [ integration - over - lam - and - phi ] ] integration over @xmath22 and @xmath23 : + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + the integral over @xmath22 , subject to the @xmath137-constraint in ( [ 52 ] ) , merely contributes one inverse square - root of the primed determinant of @xmath138 to the partition function and so does the integration over @xmath23 .
in fact , to obtain the partition function of the thirring model we divide @xmath139 by the corresponding partition functions @xmath140 of the free bosons , eq .
( [ t53 ] ) . using ( [ td1 ] ) and ( [ t46 ] ) we obtain where we have also used the scaling formula for the primed determinant of @xmath141 @xcite with @xmath142 being the number of zero modes of the operator . on the torus @xmath143 and we find @xmath144 which produces the extra factor @xmath145 . in the thirring model limit @xmath146 and the square - root in ( [ t95 ] ) disappears .
[ [ zero - temperature - limit ] ] zero - temperature limit : + + + + + + + + + + + + + + + + + + + + + + + to investigate the thermodynamics of the model we assume space - time to be flat and that @xmath57 . then @xmath147 is the grand canonical potential .
first we analyze the low temperature limit of @xmath148 . for @xmath149
this yields the ground state energy .
we observe that for @xmath57 the covariance matrix @xmath150 in ( [ t94 ] ) simplifies to and has eigenvalues with corresponding eigenvectors also , the @xmath151 tensor ( see [ t17a ] ) and @xmath152 ( see [ t28 ] ) in ( [ t93 ] ) simplify to @xmath153 for @xmath154 the saddle point approximation to the gaussian sum ( [ td1 ] ) defining the theta - function becomes exact and therefore using that @xmath155 we find for the zero - temperature grand potential of the un - gauged model . here
the chemical potential and chiral twist enter only through the combination @xmath156 .
let us now discuss the potential in the various _ limiting cases_. [ [ i - no - chiral - twist - beta_1es-0-and - vanishing - chemical - potential ] ] i ) no chiral twist , @xmath157 , and vanishing chemical potential : then @xmath158 coincides with the _ ground state energy_. the minimum in ( [ t99 ] ) is attained for @xmath159 $ ] and we find only for anti - periodic boundary conditions , that is for @xmath160 , does this _ casimir energy _
coincide with the corresponding result for free fermions . for @xmath161
the casimir force is always attractive whereas for @xmath162 it can be attractive or repulsive , depending on the value of @xmath163 .
the result ( [ t100 ] ) is in agreement with the literature @xcite .
for example , it coincides with de vega s and destri s result if we make the identification @xmath164 and @xmath165 in formula ( 42 ) of that paper .
[ [ ii - small - twists - and - chemical - potential ] ] ii ) small twists and chemical potential : for small @xmath166 and @xmath46 the minimum is assumed for @xmath167 and the potential simplifies to and does not depend on the chemical potential . for vanishing @xmath168 the minimum of ( [ t99 ] ) is attained for @xmath169\mtxt{and } n_2=[\ha+\al_1+\beta_1+{\mu l\ov 2\pi}],\end{aligned}\ ] ] where @xmath170 $ ] denotes the biggest integer which is smaller or equal to @xmath171 .
this then leads to the following zero temperature potential for @xmath172 this reduces to the casimir energy for free fermions with left - right symmetric twists and agrees with the results in @xcite.note , however , that for @xmath173 we disagree with @xcite .
the difference is due to the second term on the right in ( [ t99 ] ) .
let us give two arguments in favour of our result : the discrepancy arises from the prefactor appearing in the fermionic determinant ( [ t36 ] ) . as discussed earlier
this prefactor implies the breakdown of holomorphic factorization , a property which has been presupposed in @xcite .
one can show that our results can be reproduced by starting with massive fermions and taking the limit @xmath174 ( see appendix b ) .
the second argument goes as follows : suppose that @xmath175 .
then ( [ t101 ] ) simplifies to for massless fermions the fermi energy is just @xmath46 and at @xmath176 all electron states with energies less then @xmath46 and all positron states with energies less then @xmath177 are filled .
the other states are empty . since @xmath178 is the expectation value of the electric charge in the presence of @xmath46
we conclude that it must jump if @xmath46 crosses an eigenvalue of the first quantized dirac hamiltonian @xmath179 . for vanishing twists the eigenvalues of @xmath179
are just @xmath180 . from ( [ t102 ] )
one sees by inspection that the electric charge @xmath181=2n\quad\mtxt{for } e_n\leq\mu < e_{n+1}\end{aligned}\ ] ] indeed jumps at these values of @xmath46 .
further observe , that in the thermodynamic limit @xmath182 the density @xmath183 reduces for @xmath184 to the standard result for free electrons .
[ [ equation - of - state ] ] equation of state : + + + + + + + + + + + + + + + + + + we wish to derive the equation of state for finite @xmath185 in the infinite volume limit @xmath182 .
this may be achieved by interchanging the roles played by @xmath48 and @xmath47 .
more precisely , using that @xmath136(\lambda)=\sqrt{\det(i\lambda^{-1})}\ ; e^{2\pi i w\cdot u}\;\theta\big[{-w\atop u}\big ] ( i\lambda^{-1})\end{aligned}\ ] ] we find in analogy with the low temperature limit , that for @xmath182 the pressure is given by @xmath186 here the minimum of the real part has to be taken .
again the minimization arises from the saddle point approximation to the theta function which becomes exact when @xmath187 . for small twists
the minimum is assumed for @xmath167 and then @xmath188 becomes independent on the chiral twist @xmath189 .
as we have interchanged the roles of the temporal and spatial twists this is consistent with the earlier result that for small twists @xmath148 is independent of @xmath190 . in particular , for @xmath191 , we find the following equation of state this result is consistent with the renormalization of the electric charge which is conjugate to the chemical potential .
it shows that the thermodynamics of the thirring model is not just that of free fermions as has been claimed in @xcite .
indeed , the zero point pressure is multiplied by a factor @xmath192 .
this modification arises from the coupling of the current to the harmonic fields .
it is missed if only the local part of the auxiliary field is considered , which is the case if one quantizes the model in minkowski space and then replaces the @xmath193-integral in the green functions by the matsubara sum .
this remark should also be taken seriously in four dimensions !
furthermore , we see that the pressure @xmath142 is real only for @xmath191 , which is consistent with the finite temperature boundary conditions . in the first part of this section
we derive the kac - moody and virasoro algebras of the model ( [ t1 ] ) without gauge - interaction and prepare the ground for an extension , containing in particular the minimal models , in the second part.recall ( [ t12a ] ) that for @xmath194 the theory reduces to a conformal field theory on flat minkowski space - time . to continue
it is convenient to introduce adapted light cone coordinates @xmath195 and the chiral components of the dirac spinor @xmath196 .
then after substituting the classical equations of motion depends only on @xmath197 and is therefore the chiral noether current . evaluating the poisson bracket of the symmetry generator @xmath198 with the different fields
yields the classical structure [ [ short - distance - expansions ] ] short distance expansions : + + + + + + + + + + + + + + + + + + + + + + + + + + let us now determine the quantum corrections to these classical results .
these are computed within the euclidean functional approach from the short - distance expansions of the relevant @xmath199 point functions .
we need not postulate kac - moody and virasoro algebras in advance as has been done in @xcite .
these structures are derive here . when comparing the classical with the quantum results one should keep in mind that the roles of @xmath200 and @xmath201 are interchanged when one switches from minkowski to euclidean space - time . in coordinates adapted to the holomorphic structure of the torus @xmath202 the dirac operator and
the corresponding greens function take the form @xmath203 where @xmath204 , and the chiral components of the energy momentum tensor and current are given by @xmath205 from the conformal ward identities we obtain the central charges and conformal weights directly from the correlation functions .
however , because on the flat torus the expectation value of @xmath206 is constant , we need to compute at least the @xmath207-point function to read off the conformal weights . as in the classical theory ( see ( [ t11 ] ) ) the symmetric energy momentum tensor measures the change of the effective action @xmath208 under arbitrary variations of the metric . on the torus
there are two independent contributions .
one being due to variations of the modular parameter @xmath121 and its conjugate @xmath209 which depend implicitly on the metric .
the other is due to the variations of terms which depend explicitly on the metric .
since the chiral component @xmath206 is gotten by contracting @xmath210 with @xmath211 it follows that @xmath212 when doing metric variations it is always understood that we take the flat space - time limit afterwards .
the @xmath209 variation is constant and may be discarded in the short distance expansion .
thus to analyze the algebraic structure we can work on any riemann surface .
this is not true for the finite size effects , which are global properties .
this aspect will be analyzed in section @xmath213for example , taking three metric variations of the curvature dependent part of @xmath214 with @xmath139 from ( [ t95 ] ) we find the following short distance expansions for the three point correlation function @xmath215 substituting this result into the ward identity ( [ ward ] ) we obtain the _ central charge _ and the conformal weight of the energy momentum tensor note that the the central charge as well as the conformal weight are independent of the couplings @xmath216 and @xmath217 .
the conformal weights of the fundamental fields are obtained by computing the fermionic two point function with stress tensor insertion @xmath218 since @xmath219 $ ] , its metric variation vanishes after the flat space - time limit has been taken .
the variation of @xmath220 can be found in appendix a. this yields thus we have reproduced the classical results supplemented by additional @xmath221 and @xmath222 dependent quantum corrections . in the thirring model limit @xmath223 , these terms add up to give the known anomalous dimension appearing in the thirring model @xcite . furthermore , from ( [ t507 ] ) we may derive a condition on the couplings @xmath224 if we insist on unitarity , i.e. on @xmath225 .
we find in particular for @xmath226 the conformal weights are positive for any real @xmath217.next we determine the _ kac - moody algebra _ of the @xmath19 currents .
to derive the correlation functions with current insertions we couple the fermions to an external vector field , that is consider the gauged model without maxwell term .
for example , @xmath227 the effective action with external vector field is then obtained by shifting the auxiliary fields in ( [ t28 ] ) as where @xmath228 and we have neglected the harmonic contribution to the external vector field , because it does not contribute to the short distance expansion .
the resulting effective action does not depend on @xmath229 due to gauge - invariance . to relate the variation w.r.t .
@xmath17 to that w.r.t .
we use @xmath231 is the transverse part of @xmath17 .
we obtain the following short distance expansion @xmath232 we read off the value @xmath233 of the _ central extension _ in the @xmath234-kac - moody algebra the precise @xmath235-dependence of @xmath233 ( which can of course be rescaled to unity by an appropriate redefinition of the current ) is related to a finite renormalization of the electric charge in the gauged thirring - model which we will discuss in section @xmath9.finally , from @xmath236 we obtain @xmath237 . to see how the left and right kac moody currents act on the fermionic fields we notice that after the integration over the auxiliary fields the @xmath35-dependence of the fermionic green function factorizes as @xmath238 where @xmath239 , @xmath240 and @xmath241 is the induced photon-mass ( see([t60 ] ) ) .
variation w.r.t . the @xmath242 field yields , after some algebraic manipulations , the @xmath19 charges we have used the convention where the electric charge @xmath243 is unity . in the thirring model limit
we can compare ( [ t110 ] ) with the results obtained in @xcite .
for that we need to rescale the currents such that the central extension ( [ t105 ] ) of the kac - moody algebra becomes unity @xmath244 .
it is then easy to see that we agree with furlan et al .
@xcite if we make the identification @xmath245 . [ [ non - minimal - coupling ] ] non - minimal coupling : + + + + + + + + + + + + + + + + + + + + + in section @xmath4 we have analyzed the fermionic determinant in the presence of twisted boundary conditions .
one may ask what happens if we introduce a local twist instead , that is which should be interpreted as a modification of the charge neutrality condition .
the computation of the fermionic determinant in the presence of such twists is similar to that for a weyl rescaling of the background metric ( [ t539a]-[t543 ] ) . integrating the corresponding anomaly equation
we find we will come back to the relation between the above determinant and charges at infinity at the end of this section . for the moment we use the analogy merely as a motivation to study the extension of the thirring model obtained by coupling the @xmath22-field non - minimally to the background geometry .
that is we consider the model ( [ t1 ] ) again without gauge - interaction but with an extra coupling @xmath246 then @xmath247 in ( [ t14 ] ) is modified , @xmath248 the corresponding modification of the classical conformal transformations ( [ help1 ] ) generated by the modified generator @xmath249 are whereas @xmath23 and @xmath250 remain primary fields , @xmath22 does not .
this is in fact needed for consistency . indeed , since @xmath18 is not a scalar under conformal transformations generated by @xmath251 , the term @xmath252 in the action is only conformally invariant if @xmath22 transforms inhomogeneously like a spin connection .
it may be surprising that the new symmetry transformations depend on the coupling constant @xmath253 which is not present in the flat space time lagrangian .
however , the same happens for example in @xmath9 dimensions , if one couples a scalar field conformally , that is non - minimally , to gravity .
although the lagrangian for the minimally and conformally coupled particles are the same on minkowski space - time , their energy momentum tensors are not .
the same happens for the conformally invariant non abelian toda theories which admit several energy momentum tensors and hence several conformal structures @xcite.the current still transforms as a primary with weight @xmath254 , but the energy momentum tensor acquires a classical central charge the corresponding commutators in the quantized theory with non - minimal coupling to gravity are calculated as explained for the minimally - coupled model .
one finds that the quantum corrections to ( [ help2 ] ) are identical to those of the minimally coupled model and thus are @xmath255-independent.to summarize , we have obtained the following virasoro @xmath256 kac - moody structure : central charge : kac - moody level and charges : @xmath257 conformal weights : here some comment about unitarity is in order .
it can be shown that with respect to the standard scalar product @xcite reflection - positivity holds for any real @xmath253 @xcite .
however with respect to this inner product the virasoro generators are not selfadjoint .
choosing an alternative scalar product @xcite for which they are selfadjoint , positivity does not hold in general for @xmath258 .
we give a more detailed discussion about unitary subspaces in section @xmath259 .
when quantizing a conformal field theory on a space - time with finite volume one introduces a length scale .
the presence of this length scale in turn breaks the conformal invariance and gives rise to finite size effects .
it has been conjectured @xcite that the finite size effects on a riemann surface are proportional to the central charge .
for example , when one stretches space time , @xmath260 , then the change of the effective action is proportional to @xmath114 : where @xmath261 is the euler number of the euclidean space time . in @xcite
this conjecture has been proven for a wide class of conformal field theories on spaces with boundaries .
the only important assumption has been that the regularization respects general covariance . in this subsection
we shall see that the equivalence does hold only for a particular zero - mode treatment , which differs from ( [ 52 ] ) .
the only global conformal transformations on the torus are translations which do not give rise to finite size effects .
also , the euler number vanishes and according to ( [ t111 ] ) the finite size effects are insensitive to the value of @xmath114 . for that reason we quantize the un - gauged model ( [ t1 ] ) on the sphere where the global conformal group is the moebius group . an effective method to compute finite size effects has been developed in @xcite .
it is based on the following observation : any conformal transformation @xmath262 is a composition of a diffeomorphism ( defined by the same @xmath263 ) and a compensating weyl transformation @xmath264 with @xmath265 therefore , choosing a diffeomorphism invariant regularization one has @xmath266 the change of the effective action under weyl rescaling is @xmath267)\ov \int{\cal d}(\lam\phi)\det ( i{\di}_{\hat g})\exp ( -s_b[\hat g])}\ ; , \end{aligned}\ ] ] where @xmath133 is the bosonic action ( [ t48 ] ) . since on the sphere
there are no harmonic vector fields the term @xmath268 in @xmath133 is not present . imposing the conditions ( [ 52 ] )
we obtain to evaluate ( [ t112 ] ) one introduces the @xmath254-parametric family of laplacians @xmath269 interpolating between @xmath270 and @xmath141 . integrating the corresponding anomaly equation @xcite we end up with consider now a dilatation @xmath271 .
then , the conformal angle is constant , @xmath272 , and @xmath273 .
then the first term in ( [ t114 ] ) vanishes and the finite size effect does not depend on @xmath274 .
it is given by @xmath275 and does not agree with ( [ t111 ] ) since @xmath114 in ( [ help3 ] ) depends on @xmath253 . on other riemannian surfaces one would find the same result .
note that the finite size scaling comes from the middle term @xmath276 in ( [ t114 ] ) which is topological in nature , while the short - distance behaviour of the energy - momentum correlators is controlled by the remaining two terms in ( [ t114 ] ) which are insensitive to the topology . in that sense finite size scaling and the central charge
are complementary .
there is a way to match the two results by adding the term @xmath277 to the effective action . with this new effective action
the short distance expansion of the energy - momentum correlators does not depend on @xmath253 any more and the corresponding central charge equals that obtained from the finite size scaling .
however such a term would correspond to a non - local counter term to be added to the regularized action . in this subsection
we show how the equivalence between the central charge and finite size scaling can be restored , provided the partition function is replaced by an average over un - normalized expectation values of charges at infinity .
in fact it turns out that the @xmath278-term , ie .
the non - minimal coupling to gravity , itself can be given the interpretation of a charge at infinity if the zero - mode constraints ( [ 52 ] ) is replaced by a non - translation invariant sum over charges at infinity .
the hint comes from inspecting the fermionic weights ( [ t107 ] ) , which shows that @xmath279 and @xmath280 have the same conformal weights .
we can therefore consistently put a charge at infinity with a corresponding modification of the charge neutrality condition .
the non - vanishing two - point function is now @xmath281 .
it s coincidence limit @xmath282 is again a primary field with conformal weight @xmath283.on the other hand , including a charge at infinity into the definition of the partition function we have to continue we need to determine the coincidence limit of the scalar greens function @xmath284 , i.e. to regularize the composite operator @xmath285 appearing in ( [ t591 ] ) .
the normal ordering prescription works well on the whole plane @xcite . on curved space
we must be more careful when renormalizing this operator .
the required wave function renormalization is not unique but it is very much restricted by the following requirements : first we take as reference system ( the denominator in ( [ t76 ] ) ) one with a minimal number of dynamical degrees of freedom since we do not want to loose information by our regularization .
second , the renormalized operator should have a well - defined infinite volume limit .
finally , the regularization should respect general covariance .
these requirements then force us to take as reference system the infinite plane with metric @xmath15 .
the flat metric @xmath286 is not permitted since it leads to a ill - defined expression for @xmath287 . with this choice the normal ordering in ( [ t76 ] ) is equivalent to replacing the massless green function in ( [ t591 ] ) by here @xmath288 denotes the geodesic distance between @xmath171 and @xmath289 .
the occurrence of the arbitrary mass scale @xmath46 comes from the ambiguities in the required ultra - violet regularization . on the @xmath11-sphere with constant ricci scalar @xmath131 we have @xmath290 + 1].\end{aligned}\ ] ] the expectation value @xmath291 then transforms under a constant rescaling @xmath292 as and therefore gives an extra contribution @xmath293 to the finite size scaling of the effective action . adding this contribution to ( [ t114 ] ) above we see that this is precisely the piece needed to restore equivalence with the central charge for any real or imaginary @xmath253.more
generally we can define the functional integral as an average over all possible charges at infinity : assume @xmath253 imaginary .
the ( un - normalized ) expectation values are then given by here @xmath53 denotes the @xmath19-charge of the operator @xmath294 .
in particular the partition function on @xmath295 is @xmath296\;:e^{-8\pi g_{{\sy 3 } } \lam(\xi_0)}:\ ; \;e^{-s_b[\lam ^\prime]},\end{aligned}\ ] ] where @xmath297 is the zero mode and @xmath298 the excited modes of @xmath299 .
the middle term in the above integrand is the zero - mode part of @xmath133 .
the zero - mode integration yields a delta function @xmath300 and thus the @xmath278-term itself acquires the interpretation of a charge at infinity , due to the presence of the zero mode .
the extra charges @xmath301 assure the charge neutrality of the partition function . for the general n - point function ( [ t555 ] )
the @xmath297- integration yields @xmath302 where the sum of the @xmath19-charges of the operators in ( [ t555 ] ) enters .
in particular , for neutral states , for which ( @xmath303 ) , @xmath233 must be zero and no extra charge at infinity is present.finally , using @xmath304 the averaging over all possible charges can also be written as it is easy to verify that if the action has translation invariance in the target space , then the constraints ( [ t666 ] ) and ( [ 52 ] ) are equivalent and the correlation functions do not depend on the chosen base - point @xmath305 . however , in the present case ( [ t666 ] ) clearly breaks translation invariance ( or rotation invariance on @xmath295 ) and the zero - modes constraints are inequivalent .
although we have assumed an imaginary @xmath253 , our results apply for any @xmath253 .
for particular values we recover the ( unitary ) minimal models , provided screening charges @xcite are included for the @xmath306-point function with @xmath307 .
in particular for @xmath308 and @xmath309 we obtain the ising model with @xmath310 .
in this section we extend the model by gauging the global @xmath19-symmetry .
contrary to what one might think , many aspects of the gauged model are actually simpler as compared to the ungauged model .
in particular the thermodynamical properties are independent of external conditions like chemical potentials and twisted boundary conditions .
the reason is that the model is closely related to the schwinger model , for which the spectrum consist solely of a neutral , massive particle . on the other hand ,
the gauge interaction complicates the analysis , because the @xmath19- bundle over the torus allows for gauge field configurations with winding number , so called instantons .
these , in turn , imply fermionic zero - modes which trigger a chiral symmetry breaking and therefore a non - vanishing condensate .
this is the subject of the second part of this section . in the first part
we discuss the partition function to which only topologically trivial configurations contribute . to see how the fermionic generating functional ( [ t46 ] ) is modified , we decompose a general gauge potential on a torus as
where the last @xmath207 terms correspond ( as for the auxiliary field @xmath25 ) to the hodge decomposition of the single valued part of @xmath35 in a given topological sector , that is the harmonic- , exact- and co - exact pieces .
the role of the toron fields @xmath311 has recently been emphasized within the canonical approach @xcite . in the hamiltonian formulation
they are quantum mechanical degrees of freedom which are needed for an understanding of the infrared sector in gauge theories .
also , in @xcite it has been argued that the @xmath312-phases of hot pure yang - mills theories @xcite should correspond to the same physical state if the toron fields are taken into account .
the first term in ( [ t520 ] ) is an _ instanton potential _ which gives rise to a non - vanishing quantized flux .
as noted above configurations with non - vanishing flux do not contribute to the partition function due to the associated fermionic zero modes .
we can therefore assume @xmath313 for the moment .
the fermionic generating functional is obtained from ( [ t36 ] ) by simply shifting @xmath314 which leads to with @xmath315 . to compute the _ partition function _
we must switch off the sources @xmath316 and @xmath317 in ( [ t636 ] ) so that where now note that we have kept the non - minimal coupling of the @xmath22-field to gravity as in section @xmath6 . since @xmath133 and the fermionic determinants are both gauge invariant and thus independent of the pure gauge mode @xmath318 in ( [ t520 ] ) , it is natural to change variables from @xmath17 to @xmath319 .
this transformation is one to one , provided in contrast to the auxiliary harmonic fields @xmath70 in ( [ help ] ) , the toron fields @xmath320 and @xmath321 with integer @xmath322 are to be identified , due to gauge invariance @xcite .
the measures are related as in expectation values of gauge invariant and thus @xmath318-independent operators the @xmath318-integration cancels against the normalization .
this simply expresses the fact that in @xmath323 the ghosts decouple in the lorentz gauge.as we shall see shortly it is advantageous to integrate first over the toron fields . by using the series representation of the theta functions one computes
since the result appears always together with the @xmath316-function factor in ( [ t46 ] ) it is convenient to introduce @xmath324 in the following expressions .
the result ( [ t55 ] ) does not depend on the @xmath179-field and hence the @xmath179-integration in ( [ t54 ] ) becomes gaussian and yields a factor @xmath325 so that where we inserted the explicit expression ( [ t50 ] ) for the jacobian .
now we see why we did well integrating over the toron fields first .
it has washed out the dependence on the boundary conditions and chemical potential in ( [ t55 ] ) .
the integral over @xmath22 , subject to the condition in ( [ 52 ] ) , decouples completely apart from the non - minimal coupling to gravity which modifies the liouville factor and yields one inverse square - root of the determinant of @xmath138 in ( [ t56 ] ) .
thus where we have used ( [ t58 ] ) .
the @xmath23_-integration _ in contrast , leads to a finite renormalization of the dynamically generated photon mass plays the same role as the @xmath326-mass in @xmath10 @xcite .
the determinant obtained from the @xmath327 integration factorizes as @xmath328 this factorization property is not obvious since all determinants must be regulated .
but it holds for commuting operators and in the zeta - function scheme .
then the partition function simplifies to @xmath329 we can go further by using the scaling formula for the determinant of @xmath141 @xcite and the known result for the determinant of @xmath270 @xcite which together yield thus we obtain the following partition function for the general model ( [ t1 ] ) on curved spaces : again we have factored out the partition function @xmath330 for free auxiliary fields .
the formula ( [ t63 ] ) shows explicitly that in the topologically trivial sector the theory is equivalent to a theory of free massless and massive bosons with mass @xmath241 , even in curved space - time @xcite.the appearance of @xmath241 in ( [ t60 ] ) should be interpreted as _ renormalization of the electric charge _ induced by the interaction of the auxiliary fields with the fermions . after summing over all fermion - loops
this leads to an effective coupling between the photons and the @xmath23-field and in turn to a modified effective mass for the photons in ( [ t60 ] ) . in the limit @xmath331 this mass tends to the well - known schwinger model result , @xmath332 @xcite .
we have already mentioned that the chemical potential coupled to the electric charge has completely disappeared from the partition function .
this does not come as a surprise since the only particle in the gauged thirring model is a neutral boson .
this has no charge which may couple to the chemical potential .
also , if the partition function depended on @xmath46 then the expectation value of the charge would not vanish , in contradiction to the integrated gauss law .
the result ( [ t63 ] ) provides therefore another test for our result ( [ t36 ] ) for the fermionic determinants with chemical potential .
the final result is also independent of the chiral and non - chiral twists .
the normal twists have been wiped out by the toron integration .
in fact the chiral twists are equivalent to a chemical potential and therefore the above remarks concerning the chemical potential apply here as well .
did we assume holomorphic factorization for the fermionic determinant @xcite then the partition function would depend on the chiral twists.we conclude this subsection by giving the explicit formula for the partition function on the flat torus
. the zeta - function regularized massive determinant is expressed by @xmath333 where and @xmath334 is the inner product taken with the reference metric , and the sum is over all @xmath335 with the origin excluded . for @xmath336 ,
in which case the partition function has the usual thermodynamical interpretation , the result reduces to one derived previously by ambjorn and wolfram @xcite .
in addition , if @xmath48 approaches infinity we recover a result in @xcite .
the free energy for @xmath337 and on flat space simplifies then to @xmath338 with @xmath111 from ( [ t64 ] ) and the particular choice for the parameters .
we pointed out in section @xmath11 that for @xmath339 the classical theory ( [ t1 ] ) reduces to the gauged thirring model .
the same is true for the quantized theory on the torus if in addition we set @xmath340 .
more precisely , the hubbard - stratonovich transform @xcite of the thirring model is just the derivative coupling model ( [ t1 ] ) with identical couplings . in the process of showing that we shall arrive at the bosonization formulae for the gauged thirring model on the curved torus .
we shall see that only the non - harmonic part of the fermion current can naively be bosonized and that for this part the rules of the un - gauged model on flat space time @xcite need just be covariantized . for
that we calculate the partition function ( [ t54 ] ) in a different order .
first we integrate over the auxiliary fields . to understand the role of @xmath22 and @xmath23 we introduce sources for them .
thus we study the generating functional for the correlators of the auxiliary fields @xmath341=\int \cd ( \lam\phi h\psi a_\mu ) e^{\,-s + \int\sqrt{g}\big[\xi\lam+\zeta \phi\big]}.\end{aligned}\ ] ] here @xmath342\end{aligned}\ ] ] is the action of the full theory .
@xmath75 is the dirac operator in ( [ t28 ] ) with all couplings set equal and @xmath133 the bosonic action ( [ t648 ] ) .
since @xmath22 and @xmath23 integrate to zero ( see ( [ 52 ] ) ) we may assume the same property to hold for the sources .
the integration over the auxiliary fields is gaussian and yields where is the action of the _ gauged thirring model _ on curved space - time and comes from the integration over the auxiliary fields.let us first consider the partition function , that is set the sources to zero . comparing ( [ t65 ] ) with ( [ t60 ] ) and using ( [ t62 ] )
we easily find where @xmath343 is the mean field ( see ( [ 52 ] ) ) and we used ( [ t50 ] ) and ( [ t58 ] ) .
the action for the neutral scalar field @xmath344 is found to be @xmath345 since ( [ t68 ] ) holds for any @xmath346 ( and thus for the non - harmonic part of any @xmath17 , because of gauge - invariance ) we read off the following _ bosonization rules _ : where prime denotes the non - harmonic part of the currents .
thus , only the non - harmonic parts of the currents can be bosonized in terms of a single valued scalar field . to bosonize their harmonic parts
one would have to allow for a scalar field @xmath347 with winding numbers . on the infinite plane
the harmonic part is not present and we may leave out the primes in ( [ t69 ] ) .
if we further assume space time to be flat we recover the well - known bosonization rules in @xcite .
what we have shown then , is that for the gauged model on curved space time the bosonization rules are just the flat ones properly covariantized and with the omission of the zero - modes .
since ( [ t68 ] ) holds for any gauge field the current correlators in the thirring model are correctly reproduced by the bosonization rules ( [ t69 ] ) . to see that more clearly we calculate the two - point functions of the auxiliary fields in the thirring model ( [ t65]-[t67 ] ) .
for that we differentiate ( [ t65 ] ) ( @xmath346 is treated as external field ) with respect to the sources and find where @xmath348 is the free massless green - function in curved space - time and the integrations are over the variables @xmath349 and @xmath350 with the invariant measure on the curved torus . here @xmath351 are vacuum expectation values in the thirring model ( [ t66 ] ) .
alternatively we can calculate these expectation values from ( [ t56 ] ) and ( [ t59 ] ) , where the fermionic integration has been performed and find @xmath352 comparing this with the result ( [ t70 ] ) we see at once that @xmath353 these correlators express the gauge invariance and the axial anomaly @xmath354 in the gauged thirring model .
they can be correctly reproduced with the bosonization rules ( [ t69 ] ) .
they are not reproduced with the ones derived for the un - gauged model @xcite .
the chiral condensate is an order parameter for the chiral symmetry breaking .
however , on the torus its expectation value , whose temperature- and curvature dependence we will here compute would vanish if topologically non - trivial gauge field configurations were absent .
there is a useful classification of the gauge configurations corresponding to the number of fermionic zero modes they give rise to .
if we let @xmath355 , where @xmath356 counts the number of zero - modes with positive / negative chirality , then we have which establishes a relation between the number of fermionic zero modes ( or , more precisely the number of zero modes with positive chirality minus the number of those with negative chirality ) and the first chern character of the bundle . also from ( [ indexa ] ) one immediately concludes that the flux must be quantized in integer multiples of @xmath357 .
this is really a consequence of the single valuedness of the fermionic wave function ( cocycle condition ) . recalling the decomposition ( [ t520 ] ) of the gauge field we now concentrate on the first term @xmath358 which is the _ instanton potential _ giving rise to a non - vanishing quantized flux @xmath359 .
since @xmath11-dimensional gauge theories are not scale or weyl invariant , as @xmath9-dimensional ones are , the instantons on a conformally flat space - time are not identical to the flat ones @xcite .
as representative in the @xmath233-instanton sector we choose the , up to gauge transformations , _ unique absolute minimum _ of the maxwell action in a given topological sector .
it has field strength @xmath360 .
the corresponding potential can be chosen as is the instanton potential on the flat torus with the same flux but field strength @xmath361 .
the function @xmath261 is then determined ( up to a constant ) by the solution of this equation is given by where is the green - function for @xmath362 . in deriving ( [ t23 ] )
we have used that @xmath363 which follows from the spectral resolution ( [ t24 ] ) for the green function in which the constant zero mode @xmath364 of @xmath141 is missing .
our choice for the instanton potential ( [ t21 ] ) corresponds to a particular trivializations of the @xmath19-bundle over the torus @xcite . in other words ,
the gauge potentials and fermion fields at @xmath365 and @xmath366 are necessarily related by a _
nontrivial gauge transformation _ with winding numbers for the choice ( [ t21 ] )
we find @xmath367 note that @xmath35 is still periodic in @xmath368 with period @xmath48 and @xmath18 still obeys the first boundary condition in ( [ t518 ] ) . to calculate the fermionic _
zero modes _ we use the square of the dirac operator in a pure instanton and harmonic background ( @xmath369 ) on the flat torus ( [ dsq ] ) simplifies to in other words , @xmath370 is the same in the left- and right - handed sectors , up to the constant @xmath371 .
furthermore this operator commutes with the time translations which leads to the following ansatz for the zero - modes @xmath372 where we have assumed @xmath373 .
the choice of @xmath374 is dictated by the time - like boundary conditions in ( [ t518 ] ) .
inserting this ansatz into the zero mode equation @xmath375 yields @xmath376 this is just the differential equation for the ground state of a generalized harmonic oscillator to which it reduces for @xmath377 .
the solution is given by @xmath378.\end{aligned}\ ] ] these functions do not obey the boundary condition ( [ t26a ] ) , but the correct eigenmodes can be constructed as superpositions of them . for
that we observe that @xmath379 so that the sums where @xmath380 , obey the boundary conditions and thus are the @xmath233 required zero - modes . indeed , since @xmath381 in non - negative there are no zero modes with negative chirality because of ( [ t138 ] ) . with ( [ indexa ] )
we conclude then that there are exactly @xmath233 zero modes with positive chirality .
modes with different @xmath142 in ( [ t39 ] ) are orthogonal to each other and the overall factor normalized them to one , so that the system ( [ t39 ] ) forms an orthonormal basis of the zero - mode subspace .
for @xmath382 the zero - modes are the same if one replaces @xmath383 by @xmath384.to compute the fermionic determinant in a given topological sector we again introduce the one - parameter family of dirac operators which interpolates between @xmath76 and @xmath75 , similarly as in ( [ t539a ] ) .
but now @xmath385 with @xmath43 and @xmath44 from ( [ t5 ] ) , contains an instanton contribution .
also note that @xmath76 contains the instanton part @xmath386 . to compute it s determinant
we observe that the simple form ( [ t138 ] ) of @xmath387 allows one to reconstruct its spectrum completely @xcite : @xmath388 the corresponding determinant is @xcite to relate the determinants of @xmath76 to that of @xmath75 we again integrate the anomaly equation , which now reads where , due to the fermionic zero - modes , the projector onto the zero - mode - subspace , appeared . for the deformed operator @xmath389 the first seeley - dewitt coefficient is integrating w.r.t .
@xmath121 @xcite one ends up with the following formula for the determinant in arbitrary background gravitational and gauge fields : in deriving this result we used that @xmath390 .
now we are ready to compute the chiral condensate @xmath391 . observing that the fermionic green s function anti - commutes with
@xmath80 one sees at once that only configuration supporting one fermionic zero - mode with positive chirality contribute to the chiral condensate @xmath392\,e^{-s_b},\end{aligned}\ ] ] where @xmath393 .
earlier we have seen that these are the gauge fields with flux @xmath394 or instanton number @xmath395 .
thus the condensate becomes where @xmath396 stands for the exponentials in ( [ t43 ] ) .
first we integrate over the toron field @xmath103 .
the @xmath103-dependence enters only through the zero mode and more specifically @xmath397 in ( [ t39 ] ) with @xmath398 . using the series representation for the theta functions one finds note that the result does not depend on the chemical potential similarly as in our calculation of the partition function . to continue we observe that the term @xmath399 in @xmath396 vanishes because of our conditions ( [ t49 ] ) and ( [ 52 ] ) on the fields @xmath230 and @xmath23 .
furthermore @xmath400=s_b[k=0]+{2\pi^2\ov e^2v}$ ] .
the remaining functional integrals are performed similarly as those leading to the partition function and we end up with the following formula for the condensate the expectation value is evaluated with @xmath401.\end{aligned}\ ] ] a formal calculation of the resulting gaussian integrals yields where and @xmath402 are the massive and massless green - functions .
here we encounter ultra - violet divergences since @xmath284 is logarithmically divergent when @xmath171 tends to @xmath289 . to extract a finite answer we need to renormalize the operator @xmath403 as explained in section @xmath259 .
this wave function renormalization is equivalent to the renormalization of the fermion field in the thirring model and thus is very much expected @xcite . the flat green s function on the torus @xmath404(0,\tau)\big|^2=-\frac{1}{4\pi } \big|\frac{1}{\eta(\tau)}e^{i\pi\tau ( \xi^0/l)^2}\theta_1 \big({\tau \xi^0+\xi^1\ov l},\tau\big)\big|^2,\end{aligned}\ ] ] where @xmath405 , possesses the logarithmic short distance singularity furthermore @xmath406 with the prescription explained in section @xmath259 we find that on the flat torus @xmath407 has now the finite coincidence limit to determine the chiral condensate we also need to determine @xmath408 on the diagonal . in a first step
we shall obtain it for the flat torus .
its @xmath51-dependence is then determined in a second step . for @xmath56 and @xmath409
the green function @xmath410 has been computed in @xcite .
the generalization to arbitrary @xmath121 is found to be where we introduced the dimensionless constant @xmath411 and the functions we used the abbreviations substituting ( [ t79 ] ) and ( [ t78 ] ) into ( [ t74 ] ) with @xmath56 we obtain the following _ exact formula for the chiral condensate _ on the torus with flat metric @xmath412 : where we used that on the flat torus @xmath413 and @xmath414 .
furthermore , we identified @xmath46 with the natural mass scale @xmath241 of the theory . to extract the finite temperature behaviour of the chiral condensate we take @xmath57 where @xmath415 is the inverse temperature . in the thermodynamic limit @xmath187
. then @xmath416 , @xmath417 and the expression for the chiral condensate simplifies to using @xmath418 where @xmath419 is the euler number , we obtain the _ zero temperature limit _ for temperatures large compared to the induced photon mass @xmath43 vanishes .
thus we obtain the _ high temperature behaviour _ it is instructive to discuss the various limiting cases . for all @xmath420 , i.e. the schwinger model limit , the exact result ( [ t83 ] ) simplifies to where now @xmath421 is the induced photon mass in the schwinger model .
this formula for the temperature dependence of the chiral condensate in @xmath7 agrees with the earlier results in @xcite .
next we wish to investigate how the self interaction of the fermions affect the breaking . for large coupling @xmath217 and fixed temperature the exponent in ( [ t83 ] ) vanishes so that @xmath422 hence , for very large current - current coupling the chiral condensate vanishes . or in other words , the electromagnetic interaction which is responsible for the chiral condensate , is shielded by the pseudo scalar - fermion interaction.for intermediate temperature and coupling @xmath217 we must retreat to numerical evaluations of the sums defining the chiral condensate in ( [ t83 ] ) .
the numerical results are depicted in fig . 1 fig .
1 : the chiral condensate as a function of the temperature and the coupling constant @xmath217 .
@xmath423 @xmath424 @xmath425\ ] ] how does the gravitational field affect the chiral condensate ? to answer this question we need to know the massive green s function , entering in ( [ ta74 ] ) , for non - trivial gravitational fields ( for simplicity we assume @xmath176 ) .
let us first consider a space with constant negative curvature .
then @xmath426 has been computed explicitly in @xcite .
here we only need its short distance expansion , given by @xmath427 where @xmath428 and @xmath429 is the digamma function . substituting ( [ gmr ] ) into ( [ ta74 ] )
we end up with the exact formula for the _ chiral condensate for constant curvature _ the asymptotic expansions for _ large - and small curvature _
are easily worked out inserting the corresponding expansions for the digamma function @xcite .
we find and hence the chiral condensate decays exponentially for large curvature analogous to the high temperature behaviour .
however , the pseudo - scalars do not suppress the effect of the curvature in contrast to ( [ t85 ] ) . comparing the exponentials in ( [ bigr ] ) to ( [ t85 ] )
we may define the curvature induced effective temperature as @xmath430 in passing we note that if we compare the prefactors , rather than the exponentials , we would write @xmath431 the latter identification actually coincides ( up to factor of @xmath11 ) with the hawking temperature of free scalars in de sitter space @xcite .
the correct identification involves the ( dynamical ) mass of the gauge field and is therefore not universal . from this observation
we learn that the temperature associated with curvature depends on the matter content .
note finally that the non - minimal coupling ( @xmath253 ) has no effect on the chiral condensate . in fig .
2 we have plotted the chiral condensate for arbitrary constant values of the curvature .
2 : the chiral condensate as a function of the curvature and the coupling constant @xmath217 .
@xmath423 @xmath432 @xmath433\ ] ] for gravitational backgrounds with non - constant curvature we have to refer to perturbative methods for the calculation of the massive green s function .
again we only need the short distance expansion of @xmath426 . for geodesic distances @xmath434 small
compared to @xmath435 the massive green s function may be approximated by the seeley - dewitt expansion @xcite @xmath436 where @xmath437 is the hankel function of the second kind and order zero .
in particular @xmath438 \mtxt{for } z\to 0.\ ] ] inserting ( [ h0 ] ) into ( [ ta74 ] ) we end up with the following expansion for the _ chiral condensate in an arbitrary background _
@xmath439,\label{chirs}\ ] ] where we have used that @xmath440 .
the first order contribution involves @xmath441 and reproduces the asymptotic behaviour ( [ smallr ] ) .
higher order contributions must be taken into account to uncover the effect of variable curvature .
for this one has to substitute is the corresponding seeley dewitt coefficients @xmath442 into ( [ chirs ] ) . these have been computed up to @xmath443 @xcite .
in this paper we have elaborated on various features of the thirring model as well as some of its extensions .
in particular we found the dependence of the partition function on the chemical potential and the non - trivial boundary conditions for the fermions on the torus .
for that a careful analysis of fermionic determinants has been crucial .
we have found that the familiar chiral anomaly of the uv - regularized two point function is also seen in the ir - sector as a breakdown of holomorphic factorization .
this fact , which has not been properly taken into account previously , together with the presence of harmonic contributions to the current , leads to a modification of the equation of state due to the current - current interaction .
we believe that our results could also be obtained in the bosonized theory , provided the usual bosonization rules are modified to include scalar fields with winding numbers , i.e. scalar fields with values in a compactified target space.furthermore , we have deformed the conformal structure by allowing for different couplings in the transversal- and the longitudinal parts of the current - current interaction . this does not change the virasoro- and kac - moody algebra , but modifies the conformal weights of the primaries and in particular of the fermionic fields .
not all values of the coupling constants belong to physical theories , since positivity of the scalar product imposes certain restrictions on them .
our approach allows also for a non - minimal coupling of the longitudinal sector to gravity .
while such a coupling may seem to be ad - hoc we gave some arguments that it might arise naturally when quantizing fermions in presence of a a background charge .
we find that the central charge of the virasoro algebra is sensitive to the non - minimal coupling .
in particular @xmath444 occurs for certain values of the coupling constant .
however , we have not been able to derive constraints on this extra coupling without referring to the result by friedan , qiu and shenker .
we believe that an independent derivation of their result within a fermionic model would be most welcome .
we have also established that the central charge controls the finite size effects only for a particular treatment of the zero - modes of the auxiliary fields which is equivalent to an average over charges at infinity.finally we have considered the _ gauged thirring model _ in curved space - time .
we find that the partition function is independent of vectorial as well as chiral twists and the chemical potential .
this result , which technically is due to the harmonic contributions to the gauge - fields , is in fact expected as a consequence of gauss s law .
furthermore , using the ( probably not so obvious ) factorization property of the zeta - function regularized determinants of commuting operators we find that the partition function can be expressed completely in terms of a single massive scalar field .
the gauged thirring model shows a chiral symmetry breaking which originates in the existence of fermionic zero - modes and thus in configurations with winding number ( instantons ) .
we have obtained explicit expressions for these instantons as well as the expectation value of the chiral condensate as a function of temperature and curvature .
the condensate is exponentially suppressed for high temperatures and/or big curvature which is interpreted as an almost restoration of the chiral symmetry under these extreme conditions . although temperature and curvature have qualitatively the same effect they can not be identified . in particular the identification with the hawking temperature for free scalar fields in de sitter space does not hold in the present situation .
it follows from general arguments that the chiral symmetry can not be restored for any finite temperature or curvature so an exponential suppression is most we can expect .
in fact , it has been argued earlier , that the axial @xmath19-symmetry in @xmath9 dimensional @xmath10 also shows an almost restoration as a function of the temperature @xcite .
our results on the curvature dependence could motivate a corresponding investigation in @xmath10 .
finally we note that the chiral condensate is linearly suppressed for large current - current couplings .
[ [ acknowledgments ] ] acknowledgments : + + + + + + + + + + + + + + + + this work has been partially supported by the swiss national science foundation and the eth - zrich , where part of the research leading to the present results has been done . we wish to thank arne dettki for his collaboration at the beginning of this work and j. frhlich , k. gawedzky , d. oconnor and c. nash for helpful discussions .
our conventions for the metric and curvature agree with those of birrell and davies @xcite .
we use the chiral representation @xmath445 for flat space with lorentzian signature and @xmath446 in euclidean space .
furthermore @xmath447.in what follows we derive some variational formulae used in the text .
here @xmath448 denotes the space - time and lorentz covariant derivative.using the definition of the christoffel symbols it is straightforward to show that for some formulae related to the variation of the tetrad let us refer to @xcite in addition we have when performing the variation of curvature dependent expressions we have used the identities depending on the topology of space - time , the reference curvature @xmath449 may be different from zero . in this case
it is not possible to express the conformal angle @xmath51 in terms of the curvature scalar .
nevertheless , to perform variations of @xmath450-dependent expressions , the identity proves to be useful.taking the variations of the equations for the scalar and fermionic greens functions we may derive ( up to contact terms ) the following variational formulae @xmath451 here all arguments and derivatives which are not made explicit in the integral refer to the coordinate @xmath349 over which is integrated .
finally , we need the following formula for the variation of the inverse laplacian where @xmath452 is the volume of space - time and @xmath123 an arbitrary function . to prove this identity we note that for @xmath453 we have @xmath454 varying this equation yields @xmath455 which may be inverted to give varying the identity @xmath456 allows to replace the last term of ( [ b.11 ] ) to obtain the required result ( [ b.10 ] ) .
in this appendix we compute the partition function for massive dirac fermions in the canonical formalism . in the limit @xmath174
we confirm the result ( [ t36 ] ) for the fermionic determinant with chemical potential in chapter 3 . for massive fermions one can not consistently impose chirally twisted boundary conditions .
however , from the explicit eigenvalues ( [ t31b ] ) one sees at once that the chiral twist @xmath166 and the chemical potential are equivalent . one can easily verify that this equivalence holds also for massless fermions in the canonical approach and that @xmath457 .
let us therefore compute the partition function where the last @xmath114-number term represents the infinite vacuum contribution which must be regularized . to do that we employ the zeta function regularization .
that is we define the zeta - function for @xmath113 by the sum @xmath464 which in turn defines an analytic function on the whole complex @xmath434-plane up to a simple pole at @xmath465 .
the analytic continuation is constructed by a poisson resummation and @xmath466 , @xmath467 .
taking the mellin transform of ( [ c7b ] ) we find @xmath468 @xmath43 diverges at @xmath469 since the kelvin function @xmath470 for small @xmath471 .
it follows that the @xmath472 term in ( [ c7a ] ) diverges .
this divergence is regularized by subtracting the ground state energy of the infinite volume system . indeed , because of the exponential decay of the bessel function for large arguments , only the @xmath473 term contributes for infinite volume .
so we find for the regularized sum now we perform the limit @xmath474 . only the most singular term in
the expansion of the bessel function contributes , hence @xmath475 where @xmath476 is the lommel function @xcite . in particular for @xmath477 this function is @xmath478 so that finally let us now compute the partition function . using ( [ c12 ] ) we easily find @xmath479=q^{[\al_1 ^ 2-{1\ov 12}]}\nonumber\\ & = & \prod_{n>[\ha+\al_1]}^\infty ( 1+q^{(n-\ha-\al_1)}e^{\beta\mu})\prod_{n>-[\ha+\al_1]}^\infty ( 1+q^{(n-\ha+\al_1)}e^{\beta\mu})\cdot\nonumber\\ & & \;\prod_{n>[\ha-\al_1]}^\infty ( 1+q^{(n-\ha+\al_1)}e^{-\beta\mu})\prod_{n>-[\ha-\al_1]}^\infty ( 1+q^{(n-\ha-\al_1)}e^{-\beta\mu})\nonumber\\ & = & { 1\over |\eta(\tau)|^2 } \theta\big[{-\al_1\atop i\mu { \beta\ov 2\pi}}\big](0,\tau)\ \bar\theta\big [ { -\al_1 \atop -i\mu { \beta\ov 2\pi } } \big](0,\tau),\label{c.13}\end{aligned}\ ] ] where we have used the product representation of the theta functions in the last identity and that @xmath480
. a non - vanishing chiral twist @xmath166 can now be included by shifting the chemical potential .
thus we have confirmed the formula ( [ t36 ] ) in the text . note that for @xmath481 the zero - temperature limit of the grand potential is not equal to the vacuum expectation value of @xmath482 . for @xmath481
all states up to the @xmath46-dependent fermi energy are filled .
for example , for @xmath483 in the limit @xmath154 , @xmath148 reduces to the expectation value of @xmath484 in the one - electron state .
99 see e.g. w. dittrich and m. reuter , `` effective lagrangians in qed '' , lecture notes in physics , springer , heidelberg , 1984 and references therein .
w.e . thirring , _ ann .
( ny ) _ * 3 * ( 1958 ) , 91 ; v. glaser , nuovo cim . * 9 * , ( 1958 ) , 1007 .
h. yokota , prog .
. phys . * 77 * ( 1987 ) , 1450 .
freedman and k. pilch , phys .
213b * ( 1988 ) , 331 ; _ ann .
( ny ) _ * 192 * ( 1989 ) , 331 ; s. wu , comm .
phys . * 124 * ( 1989 ) , 133 .
c. destri and j.j .
devega , phys . lett . *
223b * ( 1989 ) , 365 .
birrell and p.c.w .
davies , `` quantum fields in curved space '' , cambridge univ . press , 1982 .
gibbons and m.j .
perry , proc .
london * a 358 * ( 1978 ) , 467 .
i. sachs and a. wipf , helv .
acta * 65 * ( 1992 ) , 653 .
h. leutwyler and v. smilga , phys . rev .
* d46 * ( 1992 ) , 5607 .
k. johnson , nuovo cim .
* 20 * ( 1964 ) , 773 . c. jayewardena , helv .
acta * 61 * ( 1988 ) , 636 ; h. joos , nucl . phys . * b17 * ( proc . suppl . ) ( 1990 ) , 704 ; helv .
acta * 63 * ( 1990 ) , 670 ; k.d .
rothe and j.a .
swieca , _ ann .
phys ( n.y . ) _ * 117 * ( 1979 ) , 382 ; j. verbaarschot et.al , nucl . phys .
* 425 * ( 1994 ) , 553 .
birrell and p.c.w .
davies , phys . rev .
* d18 * ( 1978 ) , 4408 .
eboli , phys . rev .
* d36 * ( 1987 ) , 2408 .
k. gawedzky , `` conformal field theory '' , to appear in birkhuser , basel .
c. nash and d. oconnor , _ preprint _ dias - stp-95 - 24 .
b. klaiber , `` lecture notes in phys .
xa '' , gordon and breach , new york , 1968 .
r. casalbuoni , nuovo cim . * a33 * ( 1976 ) , 115 l. alvarez - gaume and e. witten , nucl . phys .
* b234 * ( 1983 ) , 269 ; h. leutwyler , phys . lett .
* 153b * ( 1985 ) , 65 ; l. alvarez - gaume et .
al , comm .
* 112 * ( 1987 ) , 503 ; e. dhoker and d.h .
phong , nucl . phys .
* b278 * ( 1 986 ) , 226 ; k. harada and i. tsutsui , phys . lett . * b183 * ( 1987 ) , 311 .
a. actor , fortschritte der phys .
* 35 * ( 1987 ) , 793 ; k. tsokos , phys . lett .
bf 157b ( 1985 ) , 413 .
s. blau , m. visser and a. wipf , int .
phys . * a4 * ( 1989 ) , 1467 .
s. blau , m. visser and a. wipf , int .
phys . * a6 * ( 1991 ) , 5409 .
see e.g. a.m. polyakov , `` gauge fields and strings , harwood academic publishers , 1987 . c. wiesendanger and a. wipf ,
_ ann . of phys .
( ny ) _ * 233 * ( 1994 ) , * 233 * , 125 . c. kiefer and a. wipf , _ ann . of phys .
( ny ) _ * 236 * ( 1994 ) , 241 ; s. iso and h. murayama , progr .
* 84 * ( 1990 ) , 142 .
d. lst and s. theisen , ' ' lectures on string theory `` , lecture notes in physics , springer , 1989 .
p. furlan , g.m .
sotkov and i.t .
todorov , riv .
nuovo cim .
* 12 * ( 1989 ) , 1 .
l. oraifeartaigh and a. wipf , phys . lett .
* 251b * ( 1990 ) , 361 .
j. glimm and a. jaffe , `` quantum physics '' , second edition , springer , new york , 1987 .
i. sachs , phd .
thesis , eth - no . 10728
( 1994 ) , _ unpublished_. j.l .
cardy , _ in _ ' ' fields , strings and statistical mechanics
, les houches , 1988 .
a. dettki and a. wipf , nucl
b377 * ( 1992 ) , 252 .
s. coleman , phys . rev .
* d11 * ( 1975 ) , 2088 .
da silva , m. gomes and r. kberle , phys .
* d34 * ( 1986 ) , 504 ; m. gomes and a.j , da silva , phys .
rev . * d34 * ( 1986 ) , 3916 . see eg .
t. banks , _ in _ proceedings of the santa fe tasi-87 , vol .
ii , 1987 .
f. lenz , h.w.l .
naus and m. thies , _ ann . of phys .
( ny ) _ * 233 * ( 1994 ) , 17 .
smilga , _ ann . of phys .
( ny ) _ * 234 * ( 1994 ) , 1 . l. susskind , phys
. rev . * d20 * ( 1979 ) , 2610 ; n. weiss , phys . rev .
* d24 * ( 1981 ) , 475 .
j. schwinger , phys . rev .
* 128 * ( 1962 ) , 2425 .
j.ambjorn and s. wolfram , _ ann . of phys .
( ny ) _ * 147 * ( 1983 ) , 1 .
j. hubbard , phys .
lett . * 3 * ( 1958 ) , 77 ; r.l .
stratonovich , sov .
phys . dokl .
* 2 * ( 1958 ) , 416 .
c. gattringer and e. seiler , _ ann . of phys .
( ny ) _ * 233 * ( 1994 ) , 97 .
i. sachs and a. wipf , phys .
b326 * ( 1994 ) , 105 .
ketov and o. lechtenfeld , phys .
b353 * ( 1995 ) , 463 .
bunch and p.c.w davies , proc .
* a360 * ( 1978 ) , 117 .
m. abramowitz and i.s .
stegun , `` handbook of mathematical functions '' , dover publications inc .
ny , 1972 .
christensen , phys .
rev . * d14 * ( 1976 ) , 2490 . i.g
. avramidi , phys .
lett . * b238 * ( 1990 ) , 92 .
h. leutwyler and s. mallik , z. phys .
* c33 * ( 1986 ) , 205 .
shuryak , phys . lett . * 79b * ( 1978 ) , 135 ; r.d .
pisarski and l.g .
yaffe , phys .
lett . , * 97b * ( 1980 ) , 110 .
i.s . gradshteyn and i.m .
ryzhik , `` table of integrals , series and products '' , academic , london , 1980 . | the thirring model and various generalizations of it are analyzed in detail .
the four - fermi interaction modifies the equation of state .
chemical potentials and twisted boundary conditions both result in complex fermionic determinants which are analyzed . the non - minimal coupling to gravity does deform the conformal algebra which in particular contains the minimal models .
we compute the central charges , conformal weights and finite size effects.for the gauged model we derive the partition functions and the explicit expression for the chiral condensate at finite temperature and curvature .
the bosonization in compact curved space - times is also investigated .
# 1 = cmr10 scaled4 = cmss10 at 5pt = cmbx12 scaled4 = cmr10 scaled1 = cmbx9 = cmmi9 = cmti9 = cmr9 = cmss10 at 9pt = cmsl9 = cmsy9 = cmtt9 = cmss10 at 7pt = cmbx6 = cmmi6 = cmr6 = cmss10 at 6pt = cmsy6 = cmr5 at 7pt = cmr5 = cmr5 at 10pt = cmr10 scaled4 = cmbx7 scaled2 = cmbx5 scaled2 = cmbx10 scaled2 = cmmib10 = cmmib10 = cmmib10 scaled2 = cmsy7 scaled2 = cmsy5 scaled2 = cmsy10 scaled2 = cmex10 scaled1 = cmex10 scaled2 = cmbx7 scaled1 = cmbx5 scaled1 = cmbx10 scaled1 = cmmib10 scaled 833 = cmmib10 scaled 600 = cmmib10 scaled1 = cmsy7 scaled1 = cmsy5 scaled1 = cmsy10 scaled1 = cmb10 = cmsy10 scaled1 = cmb10 = cmss10 = cmbx7 scaled3 = cmbx5 scaled3 = cmbx10 scaled3 = cmmib10
scaled3 = cmsy7 scaled3 = cmsy5 scaled3 = cmsy10 scaled3 = cmex10 scaled3 = = = 111 -.5 in epsf # 1#2 dias - stp-95 - 19 + fsuj tpi 4/95 + july 1995 i. sachs@xmath0 and a. wipf@xmath1 + @xmath2dublin institute for advanced studies + 10 burlington road , d 4 , ireland + @xmath3theoretisch - physikalisches institut , friedrich - schiller - universitt , + max - wien - platz 1 , 07743 jena , germany + < |
the two - point correlator function of a classical dynamical system can be illustrated as a double sum over periodic orbits @xmath0 where @xmath1 abbreviates the average over the energy and over a small time window , @xmath2 denotes the heisenberg time and @xmath3 , @xmath4 , and @xmath5 are the amplitude , the action , and the period of the orbit @xmath6 , respectively . as one is interested in the semiclassical limit @xmath7
, it is expected that only orbit pairs @xmath8 such that @xmath9 or small .
formulated in more mathematical terms , for a classical chaotic dynamical system the problem is to determine the periodic orbit pairs @xmath10 such that @xmath4 is close to @xmath11 , and then calculate .
this was first considered by sieber and richter @xcite who predicted that a given periodic orbit with a small - angle self - crossing in configuration space will admit a partner orbit with almost the same action .
the original orbit and its partner are called a sieber - richter pair . in phase space ,
a sieber - richter pair contains a region where two stretches of each orbit are almost mutually time - reversed and one addresses this region as a _ @xmath12-encounter _ or , more strictly , a _ @xmath12-antiparallel encounter _ ; the ` 2 ' stands for two orbit stretches which are close in configuration space , and ` antiparallel ' means that the two stretches have opposite directions .
the accuracy of sieber / richter s prediction was completely proven by huynh / kunze in @xcite . in that paper
the authors considered the geodesic flow on compact factors of the hyperbolic plane .
it was shown in @xcite that a @xmath13-periodic orbit of the geodesic flow crossing itself in configuration space at a time @xmath14 has @xmath15-partner orbit and the action difference between them is approximately equal @xmath16 with the estimated error @xmath17 , where @xmath18 is the crossing angle .
periodic orbits with @xmath19-parallel encounters was investigated by mller et al . in @xcite .
we speak of an _ @xmath19-encounter _ when @xmath19 stretches of a periodic orbit are mutually close to each other up to time reversal .
in other words , all the @xmath19 stretches must intersect a small poincar section .
mller et al . used combinatorics to count the number of partner orbits and provided an approximation for the action difference , but a construction of partner orbits and an error bound of the approximation had not been derived .
then , huynh @xcite continued considering the hyperbolic dynamical system in @xcite to deal with the technically more involved higher - order encounters .
the author proved that a given periodic orbit including an @xmath19-parallel encounter has @xmath20 partner orbits , constructed partner orbits and gave estimates for the action differences between orbit pairs .
furthermore , mathematical definitions for ` encounters ' , ` partner orbits ' , etc . were also arrived in @xcite . in the case of @xmath19-antiparallel encounter with general @xmath19 ,
the problem is a very complicated and it is still open . in this paper , we only consider the problem for @xmath21 .
we prove that a periodic orbit with @xmath12-antiparallel encounter has a partner orbit .
if the space is compact , then the partner is unique .
then we apply the result to prove sieber / richter s prediction and derive a better estimate for the action difference .
the paper is organized as follows . in section 2
we introduce some necessary background materials .
then in section 3 we consider periodic orbits with 2-antiparallel encounters .
we prove that a periodic orbit with a 2-antiparallel encounter has a partner orbit .
then we apply this result to reprove the accuracy of sieber / richter s prediction in @xcite .
we consider the geodesic flow on factor @xmath22 , where @xmath23 is the hyperbolic plane endowed with the hyperbolic metric @xmath24 and @xmath25 is a discrete subgroup of the projective lie group @xmath26 .
the group @xmath27 acts transitively on @xmath28 by mbius transformations @xmath29 . if the action is free ( of fixed points ) , then the factor @xmath30 has a riemann surface structure .
such a surface is a closed riemann surface of genus at least @xmath12 and has the hyperbolic plane @xmath28 as the universal covering . if the space @xmath30 is compact , then all the elements in @xmath31 are hyperbolic , i.e. , @xmath32 for @xmath33 .
the geodesic flow @xmath34 on the unit tangent bundle @xmath35 goes along the unit speed geodesics on @xmath30 . on the other hand ,
the unit tangent bundle @xmath36 is isometric to the quotient space @xmath37 , which is the system of right co - sets of @xmath38 in @xmath27 , by an isometry @xmath39 .
then the geodesic flow @xmath40 can be equivalently expressed as the natural ` quotient flow ' @xmath41 on @xmath42 associated to the flow @xmath43 on @xmath44 by the conjugate relation @xmath45 here @xmath46 denotes the equivalence class obtained from the matrix @xmath47 .
there are some more advantages to work on @xmath42 rather than on @xmath35 .
one can calculate explicitly the stable and unstable manifolds at a point @xmath48 to be @xmath49 where @xmath50 denote the equivalence classes obtained from @xmath51 .
if the space is compact , the flow @xmath52 is hyperbolic .
general references for this section are @xcite , and these works may be consulted for the proofs to all results which are stated above . for @xmath53 ,
denote by @xmath54 the equivalence class obtained from @xmath55 .
[ adpi ] ( a ) the following relations hold for @xmath56 : @xmath57 ( b let @xmath58\in { \rm psl}(2 , \r)$ ] for @xmath59 .
if @xmath60 , then @xmath61 for @xmath62 * proof : * ( a ) in @xmath63 we calculate @xmath64 which upon projection yields the first one .
the argument is analogous for the others .
\(b ) let @xmath65 be given by ( [ tsu - def ] ) . to begin with , @xmath66 if @xmath67 , then @xmath68 , @xmath69 , @xmath70 , and @xmath71 , using that @xmath72 .
thus @xmath73 and @xmath74 .
if @xmath75 , then @xmath76 , @xmath77 , @xmath78 , and @xmath79 , and hence @xmath80 which yields once again that @xmath81 .
@xmath82 we recall the definition of poincar sections , the shadowing lemma and the anosov closing lemma from @xcite . [ poindn1 ] let @xmath83 and @xmath84 .
the _ poincar section _ of radius @xmath85 at @xmath86 is @xmath87 where @xmath88 is such that @xmath89 ( see figure [ pss](a ) ) .
see figure [ pss](b)&(c ) for an illustration of the next two results .
[ shadlemii ] if @xmath84 , @xmath90 , and @xmath91 , then @xmath92 and @xmath93-\infty , 0];\ ] ] recall @xmath94 , for any @xmath95 such that @xmath96 . [ anosov1 ] suppose that @xmath970 , \frac{1}{4}[$ ] , @xmath83 , @xmath98 , and @xmath99
. let @xmath89 and @xmath100 for @xmath101 .
then there are @xmath102 and @xmath103 so that @xmath104.\ ] ] furthermore , @xmath105 and @xmath106 see figure [ se](a ) for an illustration for the next result .
[ selfcthm ] suppose that all elements of @xmath31 are hyperbolic and let @xmath1070,\pi[$ ] , and @xmath108 be given .
the orbit of @xmath109 under the geodesic flow @xmath110 crosses itself in configuration space at the time @xmath111 , at the angle @xmath112 , and creates a loop of length @xmath19 if and only if @xmath113 holds for any @xmath114 .
furthermore , @xmath115 the following definitions can be found in @xcite .
[ tr ] the _ time reversal map _
@xmath116 is defined by @xmath117 where @xmath118 is the equivalence class of the matrix @xmath119 .
using lemma [ adpi](a ) , we have @xmath120 [ partnerdf1 ] let @xmath84 be given .
two given @xmath13-periodic orbit @xmath121 and @xmath122-periodic orbit @xmath123 of the flow @xmath52 are called an _
@xmath85-orbit pair _ if there are @xmath124 and two decompositions of @xmath125 $ ] and @xmath126:$ ] @xmath127 and @xmath128 , and a permutation @xmath129 such that for each @xmath130 , either @xmath131\ ] ] or @xmath132\ ] ] holds for some @xmath133 and @xmath134 . then @xmath123 is called an _
@xmath85-partner orbit _ of @xmath121 and vice versa .
roughly speaking , two periodic orbits are an @xmath85-orbit pair if they are @xmath85-close to each other in configuration space , not for the whole time , since otherwise they would be identical , but they decompose to the same number of parts and any part of one orbit is @xmath85-close to some part of the other .
let @xmath84 and @xmath135 be given .
we say that a periodic orbit @xmath121 has an _ @xmath136-encounter _ if there are @xmath83 , @xmath137 such that for each @xmath138 , @xmath139 if either @xmath140 holds for all @xmath141 or @xmath142 holds for all @xmath143 then the encounter is called _ parallel encounter _ ; otherwise it is called _ antiparallel encounter _
( see figure [ se](b ) ) .
in this section we only consider 2-antiparallel encounters .
it is impossible to reconnect the ports in 2-parallel encounter to get a new ( genuine ) partner orbit ; but in the case of antiparallel encounter , we have the following result .
[ 2anti ] suppose that all the elements in @xmath31 are hyperbolic and let @xmath84 . if a periodic orbit @xmath121 of the flow @xmath52 on @xmath144 with period @xmath145 has a @xmath146-antiparallel encounter , then it has a partner
. furthermore , let @xmath147 , @xmath89 and @xmath148 for @xmath149 ; @xmath150 .
then the partner is @xmath151-partner with @xmath152 and the action difference between the orbit pair satisfies @xmath153 where @xmath122 is the period of the partner .
* proof : * let @xmath89 and write @xmath154 with @xmath155 and set @xmath156 . then by the assumption @xmath157 or @xmath158 due to lemma [ adpi](a ) .
this implies that @xmath159 . by the shadowing lemma ( theorem [ shadlemii ] ) , @xmath160 and @xmath161-\infty,0].\ ] ] putting @xmath162 , we claim that @xmath163 .
indeed , @xmath164 with @xmath165 apply the anosov closing lemma ( theorem [ anosov1 ] ) to obtain @xmath166 such that @xmath167,@xmath168 and @xmath169.\ ] ] for @xmath170 $ ] , it follows from and that @xmath171 similarly , for @xmath172 $ ] , it follows from and that @xmath173 we can easily check that the partner is a @xmath151-partner orbit in the sense of definition [ partnerdf1 ] .
@xmath82 a periodic orbit with a small - angle self - crossing has 2 almost mutually time - revered stretches .
this means that the orbit crosses the poincar section of a point in this orbit and theorem [ 2anti ] may be applied .
[ existthm ] if a periodic orbit of the geodesic flow @xmath110 on @xmath35 with the period @xmath180 crosses itself in configuration space at a time @xmath1810,t[$ ] and at an angle @xmath112 such that @xmath182 for @xmath183 , then it has a @xmath184-partner orbit .
furthermore , @xmath176 for the period of the partner and the action difference satisfies @xmath185 * proof : * let the orbit of @xmath186 be @xmath13-periodic ( @xmath13 is the prime period ) and such that it has a self - crossing of angle @xmath112 in configuration space at the time @xmath1870 , t[$ ] , i.e. , we have @xmath188 where @xmath189 ; see figure [ mainthm ] .
in addition , assume that @xmath190 with @xmath183
. then in particular @xmath191 holds .
set @xmath192 and @xmath193 .
this follows from theorem [ selfcthm ] that @xmath194 with some @xmath195 such that @xmath96 and @xmath196 .
we only consider the first case , the latter is similar .
write @xmath197 and @xmath198 ; recall the notation @xmath199 from definition [ tr ] . then @xmath200 we write @xmath201 where @xmath202 by ( [ sinphi/2 ] )
, we have @xmath203 then @xmath204 and @xmath205 due to @xmath206 for @xmath207 .
denote @xmath208 .
this leads to @xmath209 using .
we are in a position to apply theorem [ 2anti ] to have @xmath210 and @xmath211 such that @xmath167 and the orbit of @xmath212 is @xmath151-partner of the orbit of @xmath86 , where @xmath151 is determined later .
note that @xmath213 implies @xmath176 owing to remark [ ttrm ] .
furthermore , using and @xmath214 yields @xmath215 and hence @xmath216 which is [ actdiff ] .
finally , @xmath217 by and ; and therefore the partner is a @xmath184-partner orbit . @xmath82
\(i ) recall from @xcite that the partner is a @xmath15-partner and the action difference satisfies @xmath218 this means that the orbits in a sieber - richter pair are estimated closer in this paper and the estimate of the action difference is also better .
\(ii ) as mentioned in @xcite , the partner orbit is avoided crossing .
this means that it does not cross itself in encounter area .
conversely , a periodic orbit with 2 stretches almost mutually time - reversed and avoiding crossing has a partner orbit with a small - angle self - crossing in encounter area . indeed ,
since the two stretches are almost mutually time - reversed , there are @xmath89 and @xmath219 on that orbit such that @xmath86 and @xmath220 are very close . using lemma [ adpi ] we write @xmath221 with @xmath222 for some small @xmath85 . then @xmath223 implies that @xmath224 for @xmath225 and we can apply theorem [ 2anti ] to obtain a partner orbit for the original one . , phd thesis , universitt kln 2014 partner orbits and action differences on compact factors of the hyperbolic plane .
ii : higher - order encounters , _ physica d _ * 314 * , 35 - 53 ( 2016 ) partner orbits and action differences on compact factors of the hyperbolic plane .
i : sieber - richter pairs , _ nonlinearity _ * 28 * , 593 - 623 ( 2015 ) , cambridge university press , cambridge - new york 1995 , phd thesis , universitt duisburg - essen 2005 semiclassical foundation of universality in quantum chaos , _ phys .
lett._*93 * , 014103 ( 2004 ) periodic - orbit theory of universality in quantum chaos , _ phys .
e_*72 * , 046207 ( 2005 ) semiclassical approach to spectral correlation functions , in _ hyperbolic geometry and applications in quantum chaos and cosmology _ , eds .
bolte j. & steiner f. , london math .
lns 397 , cambridge university press , cambridge - new york 2012 , pp . | in this paper , we consider the geodesic flow on factors of the hyperbolic plane .
we prove that a periodic orbit including a 2-antiparallel encounter has a partner orbit .
we construct the partner orbit and give an estimate for the action different between the orbit pair .
then we apply the result to reprove the accuracy of sieber / richter s prediction in [ sieber and richter , 2001 ] .
* keywords : * geodesic flow , partner orbit , sieber - richter pair , 2-antiparallel encounter |
let us consider the electrons moving on a surface of the sphere of radius @xmath2 . in the presence of the uniform magnetic field @xmath3 we choose the gauge of the vector potential as a vector product @xmath4 .
the hamiltonian of the system is given by @xmath5 where @xmath6 is the ( effective ) mass of an electron , @xmath7 its charge ; we have set @xmath8 and omitted the trivial term connected with the spin of the particle .
we assume that the total number of particles @xmath9 ( with one projection of spin ) is fixed and defines the value of the chemical potential @xmath10 and the areal density @xmath11 .
the confining potential @xmath12 at @xmath13 and @xmath14 otherwise .
we focus our attention on the limit @xmath15 , when the variables of the problem are separated .
the radial component @xmath16 of the wave function is a solution of the shrdinger equation with the quantum well potential .
henceforth we ignore the radial component and put @xmath17 in the remaining angular part of the hamiltonian @xmath18 ; it can be done if @xmath10 lies below the first excited level of @xmath16 , which in turn means @xmath19 .
we choose the direction of the field @xmath20 along the @xmath21axis and as a north pole of the sphere ( @xmath22 ) and look for the eigenfunctions to the shrdinger equation @xmath23 in the form @xmath24 . defining the dimensionless energy @xmath25 and setting @xmath26
, we write @xmath27 s = 0 .
\label{sphdifeq}\ ] ] we introduced here the important dimensionless parameter @xmath28 with the magnetic flux quantum @xmath29 t@xmath30m@xmath31 , the magnetic length @xmath32 , the cyclotron frequency @xmath33 .
note that for a sphere of radius @xmath34 nm one has @xmath35 at the field @xmath36 t .
the equation ( [ sphdifeq ] ) is known as the spheroidal differential equation and was extensively studied previously .
@xcite the solutions to it are given by the oblate ( angular ) spheroidal functions @xmath37 with the corresponding eigenvalues @xmath38 .
it is known that spheroidal functions belong to the simplest class of special functions which are not essentially hypergeometric ones . for the spheroidal functions
there are no recurrence relations , generating function representations etc .
, which are characteristic for the classical special functions .
the spectrum @xmath38 is found as the eigenvalues of the infinite matrices . for different sets of functions orthogonal on the interval @xmath39 ,
these matrices are reduced to the tridiagonal form and could be further analyzed within the chain fraction formalism or by explicit ( numerical ) diagonalization .
@xcite perhaps , the mostly known quantum - mechanical application of the spheroidal functions is the problem of an electron in two - center coulomb potential ( h@xmath40 molecule ) .
let us first concentrate on the case of the weak field . in the absence of the field ,
the spectrum is that of a free rotator model and the solutions to ( [ sphdifeq ] ) are the associated legendre polynomials @xcite : @xmath41 according to ( [ zerob ] ) the wave functions are the spherical harmonics @xmath42 and are normalized on the surface of the sphere , @xmath43 one develops the perturbation theory around the initial wave functions ( [ zerob ] ) as long as @xmath44 .
@xcite the energies and wave functions in this case are given by series in @xmath45 : @xmath46 + o\left[\frac{p^2}l\right ] , \label{ene - weak } \\
s_{lm}(p,\eta ) & & \propto p_l^m(\eta ) + p_{l\pm2}^m(\eta ) \ , o\left[{p^2}/l\right ] \label{wf - weak } \end{aligned}\ ] ] to clarify the criterion @xmath47 we consider the case when the field ceases to be small for electrons near the fermi level @xmath48 , mostly contributing to the physical properties .
this case corresponds to the following relation @xmath49 ^ 2 e^2 \sqrt{\nu } \sim 137 ^ 2 \omega_{pl}.\ ] ] hence the field can not be treated as a perturbation , if the energy of the magnetic field in the volume of the sphere is @xmath50 times larger than the characteristic plasma frequency . for densities @xmath51@xmath52 and @xmath34
nm it corresponds to the fields greater than 40 t. for the weak field regime it is interesting to observe the following property of the spectrum ( [ ene - weak ] ) . for simplicity
we consider the situation when the @xmath53th unperturbed level is completely filled and @xmath54th level is empty . linearizing the spectrum we have @xmath55 .
it is clear that at @xmath56 the state @xmath57 is energetically more favorable than the state @xmath58 , with the resulting change in the occupation of the levels .
this phenomenon is accompanied by the jumps in the static susceptibility .
consider the free energy @xmath59 .
then , apart from the pauli spin contribution , the magnetic ( differential ) susceptibility @xmath60 is given by @xmath61 ^ 2 n'_f(\varepsilon_{lm})\right]\ ] ] with the bohr magneton @xmath62 , the fermi function @xmath63 and @xmath64 its derivative .
the first term here gives the diamagnetic contribution @xmath65 and the second term is the paramagnetic one . at low temperatures @xmath66
we let @xmath67 and change @xmath68 . performing the integration
we get @xmath69 \label{susc0}\ ] ] where the summation is restricted by @xmath70 . from ( [ susc0 ] )
we see that @xmath71 exhibits the jumps at @xmath72 while @xmath73 in the formal limit @xmath74 .
this behavior is shown in the fig .
[ fig : susc ] , together with the quantity @xmath75 . at the other fillings ,
i.e. at the other values of @xmath10 the jumps of @xmath71 take place at other values of @xmath1 and the qualitative picture remains true .
the amplitude of the jumps is @xmath9 times larger than the pauli spin susceptibility @xmath76 ; this coherent effect vanishes if the coherence on the sphere is lost due to the finite quasiparticle lifetime ( see below ) .
let us discuss the properties of the electron green s function .
for the two points on the sphere @xmath77 and @xmath78 we define @xmath79 with @xmath80 . in the absence of the magnetic field @xmath81 and one can find ( see below ) an exact representation of @xmath82 through the legendre function @xmath83 where we introduced @xmath84 and the distance @xmath85 on the sphere : @xmath86 we immediately see from ( [ g0 ] ) the logarithmic singularity in the one - point correlator ( @xmath87 ) as it should be .
it is instructive to find out how the usual expressions for the planar geometry are recovered for the large radius of the sphere , @xmath88 ; in this case @xmath89 , while @xmath90 . in the limit @xmath91 and for @xmath92 one has in the main order of @xmath93 : @xcite @xmath94 the existence of two oscillating exponents @xmath95 corresponds to quantum coherence of two waves .
one propagates along the shortest way between two points and another wave goes along the longest way , turning round the sphere . in the theory of metals one considers @xmath96 , while the finite quasiparticle lifetime can be modelled by ascribing the imaginary part to @xmath97 .
we write in this sense @xmath98 with the mean free path @xmath99 . when @xmath100 , we see that the coherence breaks and the only surviving exponent in ( [ g0-asymp ] ) has the form @xmath101 , \label{g0-damped}\ ] ] which expression coincides with the usual findings .
@xcite it is possible to obtain the closed form of the green s function in the weak field regime .
we sketch the corresponding derivation below .
first we neglect the @xmath102 terms in the wave function ( [ wf - weak ] ) and write @xmath103 using ( [ ene - weak ] ) and dropping the terms containing @xmath104 , we have @xmath105^{-1 } - \left[a + \frac{pm}a + l+\frac12\right]^{-1}\ ] ] now we represent the appearing fraction as the taylor series @xmath106 and substitute @xmath107 by @xmath108 in the last exponent .
after that we can sum over @xmath107 in ( [ defg ] ) @xcite and represent the remaining sum over @xmath109 as the contour integral , to obtain @xmath110 \oint \frac{d\nu}{2\pi i } \frac{m_e}{2 \sin \pi\nu } \frac{p_{\nu}(-\cos\omega)}{a-\nu-1/2 } \nonumber \\ & = & \exp\left[i\frac{p}{a } \frac{\partial^2}{\partial a\partial \phi } \right ] g^{0}(\omega ) , \label{g - weak } \end{aligned}\ ] ] with @xmath111 given by ( [ g0 ] ) .
when @xmath91 the expression ( [ g - weak ] ) could be simplified at the above condition @xmath112 . in this case
we use ( [ g0-asymp ] ) and observe that , upon differentiating over @xmath113 , the main contribution of order of @xmath114 stems from the numerator @xmath115 of ( [ g0-asymp ] ) .
then one uses the identity @xmath116 = \frac1a\exp\left[z \frac{\partial}{\partial a } \right ] a $ ] and finds @xmath117 } { \sqrt{2\pi a\sin\omega } } \left [ \frac{e^{ia(\pi-\omega)-i\pi/4}}{\cos \pi(a+p\beta ) } \right .
\nonumber \\ & & \left .
+ \frac{e^{-ia(\pi-\omega)+i\pi/4}}{\cos \pi(a - p\beta ) } \right ] \label{g - weak2 } \end{aligned}\ ] ] with @xmath118 . as before ,
in the presence of damping @xmath119 , the coherence on the sphere breaks and the only surviving term in ( [ g - weak2 ] ) has the form @xmath120 in accordance with previous results ( see e.g. @xcite ) .
let us now turn to the case of strong fields , @xmath74 .
we define the integer number @xmath121 as follows : @xmath122 for even @xmath123 and @xmath124 for odd @xmath123 .
the value of @xmath125 has a simple meaning , it corresponds to the number of zeroes of wave function @xmath126 within the interval @xmath127 , by analogy with the known property of @xmath128 .
the spectrum of ( [ sphdifeq ] ) is then given by a series @xcite where plus ( minus ) sign corresponds to even ( odd ) values of @xmath133 .
@xcite the functions @xmath135 are found as a series in the laguerre polynomials @xmath136 ; in the main order of @xmath137 they can be written as we see that in the main order of @xmath1 the landau quantization takes place , i.e. the spectrum ( [ ene - strong ] ) is that of quantum oscillator with the cyclotron frequency being the energy quantum .
@xcite the convergence and hence the applicability of the series ( [ ene - strong ] ) is given by the condition @xmath139 . in its turn , it means @xcite that all @xmath125 zeroes of the approximate eigenfunction @xmath140 in ( [ wf - strong1 ] ) lie in the northern hemisphere , @xmath141 , as they should .
we notice the following important property of the spectrum ( [ ene - strong ] ) . for given non - positive @xmath107 the values of @xmath142 corresponding to @xmath143 and @xmath144 coincide .
this property and the form of the wave function ( [ wf - strong0 ] ) can be understood as follows . at @xmath145 the field - induced potential @xmath146 in ( [ sphdifeq ] ) localizes the particles near the poles of the sphere .
this form of two - well potential leads to the discussed degeneracy of energy levels , while the total wave function is given by a symmetrization ( [ wf - strong0 ] ) of the wave functions ( [ wf - strong1 ] ) related to each of the wells .
the possibility of quantum tunneling between the wells lifts the degeneracy and produces the exponentially small energy splitting ( @xmath147 ) between the states @xmath148 and @xmath149 ( [ wf - strong0 ] ) .
@xcite it would be tempting to obtain the green s function at @xmath74 in the closed form
. it can be done at the following simplifications .
first , we neglect the exponentially small splitting and notice that @xmath150 $ ] , i.e. the correlations within one hemisphere only survive .
now we leave only the leading terms ( [ wf - strong1 ] ) of the wave functions and @xmath151 terms in energies ( [ ene - strong ] ) . the necessity to restrict the summation in ( [ defg ] ) by the terms with @xmath152
could be modelled by inclusion of the cutoff factor @xmath153 with @xmath154 into ( [ defg ] ) .
we put @xmath155 for simplicity of writing and raise the denominator into the exponent , @xmath156 .
next we perform the summation over @xmath125 with the use of bilinear generating function for the laguerre polynomials @xcite : with the modified bessel function @xmath158 ; in our case @xmath159 , @xmath160 and @xmath161 . making use of the property @xmath162 and rescaling @xmath163
we arrive at the intermediate formula for @xmath164 : the summation over @xmath107 is now easily done ( @xmath166 ) and we obtain the green s function as a sum of two terms , referring to two hemispheres .
the `` northern '' term , corresponding to the above expression , is given by with the analogue of the distance on the sphere @xmath168 $ ] and that of the vector product @xmath169 .
the form of the `` southern '' term is obtained by changing @xmath170 and @xmath171 in these expressions .
the last integral is reduced for vanishing @xmath172 to the hypergeometric function @xmath173 and we obtain the `` northern '' term in the form : @xmath174 \psi\left[\frac12 -\frac\omega{\omega_c},1 ; \rho \right ] , \nonumber \\ & = & \frac{m_e}{4\pi } e^{iv-\rho/2 -\delta \omega/\omega_c } \sum_{n=0}^ { \infty}\frac{l_n(\rho)}{\omega/\omega_c - n-1/2}. \label{g - strong1 } \end{aligned}\ ] ] let us discuss the applicability of ( [ g - strong1 ] ) .
our derivation was straightforward until ( [ g - strong ] ) , while the last step demanded @xmath175 .
the incomplete restructuring of the spectrum into the landau level scheme is absent in the usual planar geometry , wherein one would put @xmath176 and the expression ( [ g - strong1 ] ) would be exact . in the spherical case we can not treat the higher levels with @xmath177 in an analytical way
, it is mimicked in ( [ g - strong1 ] ) by the appearance of the exponential cutoff at @xmath178 .
a subtler issue in justifying ( [ g - strong1 ] ) is the shift of the last integration in ( [ g - strong ] ) to the interval @xmath179 .
the integration over the remaining segments @xmath180 yields the factor @xmath181 , thus the contribution of these segments to the green s function has no poles in @xmath97 .
this contribution could be combined with the smooth ( real ) part of @xmath182 stemming from the consideration of the highest energy levels .
the finite value of the cutoff parameter @xmath172 in ( [ g - strong1 ] ) becomes important for the one - point correlation function , i.e. at @xmath184 and @xmath185 .
the residues of the green s function define the local density of states ( ldos ) by the relation @xmath186 /(2\pi i ) $ ] .
if we assume that @xmath125 lowest landau levels are filled by the electrons , then it follows from ( [ g - strong1 ] ) that the ldos is given by @xmath187 . in this case
the finite @xmath188 provides a smooth variation of ldos of the form @xmath189,\ ] ] which variation is absent in the usual planar geometry .
this result , however , may depend on the approximations made and requires further numerical investigation .
it should be stressed that at the intermediate fields , at @xmath190 and @xmath191 , the spectrum and wave functions are not principally reduced to closed expressions of hypergeometric type .
the numerical methods are indispensable here .
we calculated the evolution with @xmath1 of the energy levels with @xmath192 by diagonalizing the @xmath193 tridiagonal matrices in the basis of @xmath194 .
the results are shown in the fig.[fig : lines ] where we made a following convention . from the general property of eq.([sphdifeq ] ) it follows that @xmath195 with formally negative @xmath1 .
@xcite for better readability of the graph , it is possible then to show all the data , plotting only half of them with one sign of @xmath107 but in the formally extended region of @xmath1 .
one can see in the fig .
[ fig : lines ] that the degeneracy at @xmath196 is eventually changed by the landau levels formation at @xmath197 . in the intermediate region @xmath198 the absence of any structure in the levels scheme is noted .
the mesh of lines at the intermediate fields mimicks the chaotic behavior , although this is not so .
both in the quantum problem considered here and in its classical couterpart one has two variables @xmath199 and two integrals of motion , the energy and the projection of the angular momentum onto the field direction . to obtain chaos
, it suffices to break the rotation symmetry , @xmath200 .
the latter problem is , however , beyond the scope of this study . in conclusion
, we demonstrate the exact solution of the electron gas on the sphere in the magnetic field . in the limits of weak and strong fields
this solution is reduced to the hypergeometric functions and the observable quantities are found in the closed form . in the case of intermediate fields , the solution is not essentially hypergeometric function and the observables require further numerical analysis . the author thanks s.v .
maleyev , m.l.titov , s.l .
ginzburg , v.e .
bunakov , v.i .
savichev for helpful discussions .
this work was supported in part by the rfbr grant no .
96 - 02 - 18037-a and russian state program for statistical physics ( grant viii-2 ) .
j.meixner , f.w.schfke , _ mathieusche funktionen und sphroidfunktionen _ ,
( springer - verlag , berlin , 1954 ) ; j.meixner , f.w.schfke , g.wolf , _ mathieu functions and spheroidal functions and their mathematical foundations _ , ( springer - verlag , berlin , 1980 ) .
this property is seen in the low field limit , eq.([ene - weak ] ) , but is obscure in the expression for the strong fields , eq.([ene - strong ] ) .
the reason is that the function @xmath38 has cuts in the complex plane of @xmath1 , hence the analytical continuation @xmath201 is not straightforward for @xmath1 near the complex infinity . @xcite | we consider the electron gas moving on the surface of a sphere in a uniform magnetic field .
an exact solution of the problem is found in terms of oblate spheroidal functions , depending on the parameter @xmath0 , the number of flux quanta piercing the sphere .
the regimes of weak and strong fields are discussed , the green s functions are found for both limiting cases in the closed form . in the weak fields the magnetic susceptibility reveals a set of jumps at half - integer @xmath1 .
the strong field regime is characterized by the formation of landau levels and localization of the electron states near the poles of the sphere defined by a direction of the field .
the effects of coherence within the sphere are lost when its radius exceeds the mean - free path .
= 10000 2 the electronic properties of cylindrical and spherical carbon macromolecules @xcite have attracted much of theoretical interest last years .
a large part of these studies is devoted to the band structure calculations @xcite and to the effects of topology for the transport and mechanical properties of these nanostructures .
the main interest to the topological aspects is connected with the carbon nanotubes , where one can investigate the effective models , based on the band calculations and incorporating the particular geometry of the object .
@xcite at the same time , the problem of the topology appears to be a general one and may be studied independently from the physics of carbon materials .
the recent advances in technology @xcite let one think of a wider class of the spherical nanostructures , e.g. the spheres coated by metal films,@xcite whose properties may differ from those of planar objects .
in this paper we study the gas of electrons moving within a thin spherical layer in an applied magnetic field .
we find the exact solution of this problem in terms of oblate spheroidal functions .
this physical application of the theory of spheroidal functions is demonstrated apparently for the first time .
@xcite we show the jumps in the susceptibility of the system in `` weak '' fields and the localization of the electronic states in `` strong '' fields .
the last effect could be experimentally investigated for the hemispherical tips of nanotubes .
@xcite at the intermediate fields the sophisticated structure of the functions makes the analytical treatment impossible and the numerical methods should be used for the analysis of observable quantities . |
the galactic center ( gc ) , at an assumed distance of 7.9 kpc , is a complex region containing many compact and diffuse high - energy astrophysical sources ( see @xcite for a recent review ) .
the dynamics of the central few lightyears is dominated by what is believed to be a supermassive black hole , sagittarius a * ( sgr a * ) .
the gc has been identified as a point - like source of tev @xmath1-rays by a number of air erenkov telescopes . in our work
we concentrate exclusively on the data from hess @xcite as these provide the best constraints on the source properties .
the gc signal is coincident within @xmath2 of sgr a * , though with a centroid displaced roughly @xmath3 ( @xmath4 pc ) to the east of the gc @xcite .
the spectrum is a pure power law with photon index @xmath5 , and the total flux above 1 tev is @xmath6 m@xmath7 s@xmath8 @xcite . in a recent paper @xcite
, we examined a hadronic scenario for the creation of the @xmath0tev @xmath1-rays : a power - law population of protons is accelerated close to sgr a * , perhaps through the process of stochastic acceleration on the turbulent magnetic fields expected to pervade this inner region .
these then diffuse outwards through the strong magnetic fields of the largely evacuated stellar - wind region surrounding sgr a * before , finally , colliding with gas in the high molecular density clumps that make up the circumnuclear disk ( cnd ) . such collisions ( assumed to be exclusively proton on proton for simplicity ) generate neutral mesons that then decay into the @xmath1-rays which , by construction , we detect here at earth .
in the current work we self - consistently calculate the synchrotron emissivity of the secondary leptons that must also be produced in our scenario .
these come from the decay of the charged mesons created in the same p - p collisions that we posit explain the @xmath1-ray emission
though , in constrast to the case for @xmath1-ray emission , we find that , because of the extremely strong magnetic fields there ( and despite the relatively low ambient densities ) radio emission from the wind region is more important than that from the cnd . finally , we compare our predictions for radio emission with archival and new radio data on the region we have assembled .
we firstly summarize the results obtained by @xcite . in our calculation
we employed a realistic model of the density of matter through the gc arrived at by @xcite .
these authors computed the density distribution in this region caused by the interactions of stellar winds from the young stars surrounding sgr a*. in addition to the stellar wind gas , the volume also contains a high - density torus of molecular gas with an inner radius of 1.2 pc and a thickness of 1 pc representing the observed cnd . the average density in the model stellar - wind gas ( taken to be any region where @xmath9 @xmath10 ) is @xmath11 @xmath10 , while it is @xmath12 233,222 @xmath10 within the model molecular torus .
taking @xmath13 kev as the average temperature of the stellar - wind gas @xcite , 100 k for the temperature of the molecular torus ( @xcite and references therein ) , and assuming equipartition , the average field intensity is @xmath14 mg in the stellar - wind region and @xmath15 mg within the torus . in any particular cell
the magnetic field was assumed to be generated with an intensity that satisfies @xmath16 gauss @xmath17 @xmath10 for the cnd region and @xmath18 gauss @xmath19 @xmath10 in the stellar - wind gas .
our procedure for determining a physical description of the field direction at any position in the computational grid is described in @xcite .
computational resources limited our modeling to proton energies between @xmath20 and @xmath21 tev
. a total of 222,617 proton trajectories was calculated with energies uniformly distributed in this range .
these trajectories were split into @xmath22 energy bins ( @xmath23 , \ldots , [ 13.9 - 14.0]$ ] ) . from these data the steady state distribution of protons ( given in @xmath10 ev @xmath8 ) in each of the 7500 computational cells of the cnd and the 7022 cells of the stellar - wind region could be inferred .
as a proton random - walks its way through the the turbulent magnetic field of the gc it may collide with a low - energy proton in the ambient medium and produce pions via the reaction @xmath24 . here
@xmath25 , denoting hadron , is a proton or , if charge exchange occurs , a neutron and variable and energy - dependent multiplicities of neutral ( @xmath26 ) and charged ( @xmath27 ) pions are , in general , produced but electric charge must be conserved overall .
( in our calculations we also account for the sub - dominant contribution from charged and neutral kaons . )
neutral pions will subsequently decay into two photons and charged pions to electrons , positrons , and neutrinos .
in this letter we adopt the steady - state proton distributions within each modeled clump of cnd and stellar - wind region gas arrived at by @xcite .
we then use the techniques outlined at length in @xcite to calculate on the basis of ( i ) these steady state @xmath28 distributions , ( ii ) the ambient hydrogen number density , and ( iii ) clump magnetic fields the steady state ( processed ) distributions of secondary leptons within each clump and the resulting synchrotron emission from each clump .
to test our theoretical model we have obtained @xmath0ghz radio fluxes covering the region of the cnd defined by a @xmath29 rectangle with sides parallel to galactic longitude / latitude and with sgr a * at the center .
where possible , we have chosen interferometry data obtained with an array configuration that results in a beam smaller than this region of interest to ameliorate the problem of confusion with nearby strong radio sources ( sgr a east in particular ) . on the other hand , to ensure that the diffuse flux through this region that we are interested in
present on scales up to the size of the region is not integrated out , we also demand that the array configuration possess some antenna spacings that are suitably small . for the reasons just given we make use in our analysis of medium resolution ( @xmath30 ; @xcite ) 330 mhz vla data , similar resolution sumss 843 mhz data , and serendipitous atca observations of the gc at 1384 mhz and 2368 mhz .
these latter data were obtained with a somewhat lower resolution ( @xmath31 at 1384 mhz , and @xmath32 at 2368 mhz . ) which is nevertheless sensitive to emission on the size scales of the cnd . for each radio image
, we used a number of independent methods to obtain the flux inside the region of interest ( see @xcite for details ) . to obtain a conservative estimate on the error in the flux determinations , the standard deviation between all flux estimates at a particular frequency
was taken to be the rms error on the final flux quoted .
( errors due to gain and to image noise make a negligible contribution ) .
the quoted central value for the flux at a given frequency is obtained from smearing the beam to @xmath33 resolution , and reading the peak flux which provides the most reliable determination .
flux determinations are : @xmath34 jy at 330 mhz , @xmath35 jy at 843 mhz , @xmath36 jy at 1384 mhz , and @xmath37 jy at 2368 mhz .
these data are shown as the lower frequency data points in figure [ fig : cndradiospectrum ] ( b ) .
the data points at 843 , 1384 , and 2368 mhz define a non - thermal , power - law spectrum in frequency @xmath38 , indicating a steep , synchrotron - radiating electron population ( whether primary or secondary ) of @xmath39 .
the presence of non - thermal emission near sgr a west was first discussed by @xcite and was subsequently extensively investigated by @xcite .
we note that the spectrum we determined for the region is steeper than that found by these latter authors though the overall normalization at 1384 mhz is compatible .
our datum at 330 mhz falls below the @xmath40 power law ; this , however , is as expected as sgr a complex radio emission is strongly attenuated at around this frequency and below because of free - free absorption by thermal gas associated with the sgr a west structure as established by @xcite .
it is interesting to note that , according to these authors ( see also @xcite ) , the spectral index pertaining to sgr a east as a whole is approximately 1 with a corresponding average flux density at 1384 mhz of approximately 0.014 jy / arcsec@xmath41 .
this is roughly half of the flux density we measure from the cnd region . at frequencies higher than 2368 mhz
, we have obtained snapshot observations with atca at 4800 and 8640 mhz .
we measure 22 and 21 jy at 4800 and 8640 mhz over the same solid angle as for the lower frequency observations and compatible with the results obtained by @xcite at 4.9 ghz .
finally , in figure [ fig : cndradiospectrum ] ( b ) , we also show fluxes measured by @xcite at 84 and 230 ghz , @xcite at 43.25 ghz , and @xcite at 91 ghz .
these latter measurements are not perfectly matched in solid angle to the region of concern and should be taken as indicative only .
overall , they show that the spectrum of the cnd region has become dominated by thermal emission from sgr a west at @xmath42 4 ghz and do not provide a strong constraint on our secondary synchrotron emission model .
from the modeling described in @xcite , we have simulated data describing the steady - state proton distribution for the clumps of matter that make up the cnd and the stellar - wind region ( each of varying density , magnetic field strength , and position ) . synchrotron radiation in the ghz range by secondary leptons in the mean magnetic field of the cnd , 0.35 mg
, requires initiating parent @xmath28 s of energy @xmath43 ev and only @xmath44 ev in the 3 mg mean field of the wind region ( assuming a power - law @xmath28 spectrum of spectral index 2.3 ) .
such energies are unfortunately well below the 1 tev threshold of our proton propagation modeling .
we assume , then , that the steady - state proton distribution of any given computational cell is a pure power law ( in momentum ) with such power law fitted to match our modeled 1 - 100 tev proton spectra . despite our simulating more than @xmath45 proton trajectories , we must deal ( in our power - law fitting ) with empty energy bins in many clumps .
we address this issue by producing ( on the basis of our modeled , _ overall _ clump spectra ) parameterizations of the tev proton flux and fitted spectral index in terms of the clump radial separation from sgr a * and magnetic field strength . with these parameterizations in hand
we can determine , for any clump at given radius and with given magnetic field , the _ expected _ value for the @xmath28 flux in a given 10@xmath46-decade energy bin .
for a zero - entry bin , we replace the zero with this expected value divided by the free parameter @xmath47 .
we then use a mathematica routine to perform a @xmath48 fit in the parameter @xmath47 of the overall cnd tev @xmath1-ray spectrum to the hess data ( the @xmath1-ray emission by cells in the wind region only contributes at the @xmath0 0.1% level and is ignored here ) .
we find a minimum reduced @xmath48 value of 1.2 ( for 32 degrees of freedom ) at @xmath49 .
extension of individual clump spectra below tev using the power - law assumption then allows us to also make a prediction for the synchrotron radio emission due to secondary leptons in each cnd clump and , using the same @xmath47 value , in each cell of the wind region too ( our procedure here takes into account cooling by the ionization , bremsstrahlung and synchrotron processes to arrive at the steady - state electron and positron distributions within each clump ; see @xcite for details ) . the same extension to still lower proton energies allows us to also predict the lower - energy @xmath1-ray spectrum from the cnd ( and also the wind region ) .
the total @xmath1-ray and synchrotron radio emission curves for all cells in the cnd are shown as the solid ( blue ) curves in figures [ fig : cndradiospectrum ] ( a ) and ( b ) , respectively .
gamma - ray and radio emission by cells in the wind region is shown by the long - dashed ( purple ) curves in the same figures .
as may be seen , the @xmath1-ray normalized secondary synchrotron radio flux from the cnd directly accounts for @xmath0 10 % of the observed ghz radio emission but with a rather flatter spectrum than the observation data suggest . in contrast , the wind region is modeled to produce a radio spectrum consistent with observation but with around twice the observed normalization . given that this determination involves absolutely no fine tuning and is predicated on a fit of theoretical @xmath1-ray emission _ from a separate region _
( i.e. , the cnd clumps ) to the tev hess data , this is a remarkable level of agreement and we conclude that our secondary emission model can self - consistently account for both the @xmath1-ray emission seen from the direction of the gc and the total radio emission from the wind and cnd regions ( i.e. , a @xmath50 region centered in sgr a * ) . in figure
[ fig : cndradiospectrum ] ( a ) one notes that the cnd @xmath1-ray emission for the scenario above is roughly consistent with the level of emission seen from the egret source 3eg j1746 - 2851 @xcite at the two lowest energy data points around 70 mev .
we do not claim , however , that the cnd emission explains the origin of the observed @xmath1-rays .
in fact , recent studies have shown that the egret source excludes the gc at the @xmath51 confidence level @xcite . on the other hand ,
the position of the _ predicted _ cnd emission was certainly inside the wide field of view of the egret instrument s gc pointings .
the spectrum of 3eg j1746 - 2851 serves , then , as an upper limit to the allowed emission from the cnd . the ( blue ) solid curve would seem to be just excluded .
a fit to the hess data with the additional constraint that the extrapolated @xmath1-ray emission obey the constraint that it be unobserved by egret ( which we translate to the requirement that the predicted low energy gamma ray curve pass 2@xmath52 below the most constraining egret datum near 70 mev ) requires @xmath53 and the radio spectrum obtained in this case , shown as the ( red ) short - dashed curve in [ fig : cndradiospectrum ] ( b ) , is statistically compatible with the radio data .
the reduced @xmath48 for this case is rather bad , however : 2.2 for 32 degrees of freedom .
the cnd @xmath1-ray spectrum for this case is shown as the ( red ) short - dashed curve in [ fig : cndradiospectrum ] ( a ) . also shown in figures [ fig : cndradiospectrum ] ( a ) and ( b ) are ( yellow ) dot - dash curves that show , respectively , the predicted gamma - ray and radio emission for a single zone model that assumes a single power - law @xmath28 population ( @xmath54 ) tuned to match the hess emission .
with such a model one notes that , as first determined by @xcite , the low energy @xmath1-ray curve passes well below the egret points .
but the model predicts far too little synchrotron when compared with our full calculations that sum emission over each clump of the cnd and wind regions with its own @xmath28 spectrum , magnetic field and ambient hydrogen number density .
this underestimation probably arises from the assumption of a ` mean ' @xmath55 field within the single zone model that implictly neglects the ( i ) relative accumulation of protons into clumps of higher magnetic field strength ( where they are more strongly trapped than elsewhere ) and ( ii ) the greater synchrotron emissivity of secondaries created in such clumps . finally , we have reviewed data available from observations of the gc and the cnd and wind regions at other wavelengths to see whether these offer any further constraint on our model . given the steady - state positron production rate in our scenario
we predict a 511 kev @xmath1-ray production rate from electron - positron annihilation of @xmath56 yr@xmath8 , well inside the limit from integral observations ( @xmath57 yr@xmath8 out to an angular radius of 8@xmath58 : @xcite another possible constraint is offered by hard x - ray / soft @xmath1-ray observations . our model predicts diffuse secondary emission at energies of 10@xmath59 ev due to both the bremsstrahlung and , dominantly , synchrotron processes .
we have modeled such emission to compare it against the spectrum , at a comparable energy , of the integral / ibis source igrj17456 - 2901 @xcite ; we find ourselves well below the level of flux from this source .
there are several major conclusions we can draw from this study : \(1 ) the pre - existing data were not sufficient for a tight constraint on our predictions .
new data were acquired for this project .
the compilation of data demonstrates that the black hole - induced hadronic model for the tev gamma rays introduced in @xcite is certainly consistent with the broadband emission from this process at other wavelengths .
\(2 ) in fact , more strongly , we have also shown that , _ without fine - tuning , the hadronic process may actually explain essentially all the diffuse radio emission observed from the inner few pc of the galaxy_. \(3 ) the fact that in this model the protons energized by the black hole and ejected into the ism accumulate in the magnetic fields of the cnd , provides a spatial definition for where the secondaries will be active unlike , say , the primary lepton model , in which the electrons may be accelerated wherever shock fragments form in an expanding shell .
\(4 ) the propagation calculation of @xcite showed that diffusion processing by the gc magnetic field leads to a considerable spectral steepening of the injection spectrum , @xmath60 requiring in turn , that the spectrum of protons injected close to sgr a * must be very flat , @xmath61 , in order to supply the requisite @xmath62 spectrum of @xmath1-rays .
this is much flatter than the @xmath63 spectrum expected from first - order fermi acceleration .
such a hard spectrum might be created via the ( second - order fermi ) stochastic acceleration mechanism .
for instance , as determined by @xcite stochastic acceleration in a magnetically - dominated funnel close to the black hole could accelerate protons into a distribution as flat as @xmath64 , approaching the requisite hardness . the stochastic acceleration mechanism investigated by @xcite
may also be able to produce the required spectrum .
the question of whether a stochastic acceleration can be made to work in this context will be addressed elsewhere by the current authors .
\(5 ) finally , the hadronic scenario we have explored here may be a test bed of what is actually happening near the base of relativistic jets in more powerful sources , such as agns and quasars .
there we would have a more difficult time discerning the various processes , because the environment is dense and chaotic .
but in sgr a * , the environment is much more sedate , with lower density , and less activity .
sgr a * does not itself produce relativistic jets that we can see .
but that may simply be a consequence of the relative weakness of this process in this particular source .
it may be that something like the proton injection / acceleration scenario ( with subsequent propagation out to relatively large scales before final interaction ) posited in our earlier paper and whose phenomenological consequences we have explored here is happening on a much bigger scale in the more powerful agns .
the authors thank todor stanev for producing the simulated secondary particle spectra from pp collisions exploited in this paper and anne green for providing the sumss gc 843 mhz data in numerical form .
drb is supported by the university of arizona theoretical astrophysics program prize postdoctoral fellowship .
rmc gratefully acknowledges advice and assistance from ray protheroe .
rmc is supported at the university of adelaide by ray protheroe and ron ekers australian research council s discovery funding scheme grant ( project number dp0559991 ) .
this work was funded , in part , at the university of arizona by nsf grant ast-0402502 .
the work has made use of nasa s astrophysics data system abstract service .
fm is grateful to the university of melbourne for its support ( through a sir thomas lyle fellowship and a miegunyah fellowship ) .
-ray flux data points for the gc region with theoretical spectral curves .
the data points in the @xmath0 tev range show the hess source coincident with sgr a*. the five lower energy data points are from the egret source 3eg j1746 - 2851 .
the theoretical curves are : ( blue ) * solid * : @xmath1-ray spectrum from the entire cnd after @xmath48 fitting to the @xmath47 parameter _ to the hess data only _ ; ( red ) * short dashed * : @xmath1-ray spectrum from the entire cnd after @xmath48 fitting of the @xmath47 parameter to the hess data _ with the additional constraint that the egret upper limit is obeyed _ ; ( purple ) * long dashed * : @xmath1-ray spectrum from the wind region clumps assuming the same @xmath47 value ; ( yellow ) * dot - dash * : @xmath1-ray spectrum for the case of a single zone ( with @xmath65 mg and @xmath66 @xmath10 ) model and a single , power - law proton population ; ( green ) * dot - dot - dash * : projected glast sensitivity for a 5 @xmath52 detection of a point - source with a @xmath67 spectrum ( see the glast website : http://www-glast.stanford.edu ) .
* b. * ( * right * ) observed radio fluxes from the cnd and wind regions together with theoretical predictions for radio emission .
the theoretical radio curves are : ( blue ) * solid * : radio emission from the * cnd * region for the case that the theoretical cnd @xmath1-ray emission is optimized in @xmath47 to the hess data ; ( purple ) * long dashed * : radio emission from the * wind * region assuming the same @xmath47 value ; ( red ) * short dashed * : radio emission from the * wind * region for the case that the theoretical cnd @xmath1-ray emission is optimized in @xmath47 to the hess data _ with the additional constraint that the egret upper limit is obeyed _ ; ( green ) * dotted * curve : simple power law fit to 843 , 1384 , and 2368 mhz flux points.,title="fig:",scaledwidth=40.0% ] -ray flux data points for the gc region with theoretical spectral curves .
the data points in the @xmath0 tev range show the hess source coincident with sgr a*. the five lower energy data points are from the egret source 3eg j1746 - 2851 .
the theoretical curves are : ( blue ) * solid * : @xmath1-ray spectrum from the entire cnd after @xmath48 fitting to the @xmath47 parameter _ to the hess data only _ ; ( red ) * short dashed * : @xmath1-ray spectrum from the entire cnd after @xmath48 fitting of the @xmath47 parameter to the hess data _ with the additional constraint that the egret upper limit is obeyed _ ; ( purple ) * long dashed * : @xmath1-ray spectrum from the wind region clumps assuming the same @xmath47 value ; ( yellow ) * dot - dash * : @xmath1-ray spectrum for the case of a single zone ( with @xmath65 mg and @xmath66 @xmath10 ) model and a single , power - law proton population ; ( green ) * dot - dot - dash * : projected glast sensitivity for a 5 @xmath52 detection of a point - source with a @xmath67 spectrum ( see the glast website : http://www-glast.stanford.edu ) .
* b. * ( * right * ) observed radio fluxes from the cnd and wind regions together with theoretical predictions for radio emission .
the theoretical radio curves are : ( blue ) * solid * : radio emission from the * cnd * region for the case that the theoretical cnd @xmath1-ray emission is optimized in @xmath47 to the hess data ; ( purple ) * long dashed * : radio emission from the * wind * region assuming the same @xmath47 value ; ( red ) * short dashed * : radio emission from the * wind * region for the case that the theoretical cnd @xmath1-ray emission is optimized in @xmath47 to the hess data _ with the additional constraint that the egret upper limit is obeyed _ ; ( green ) * dotted * curve : simple power law fit to 843 , 1384 , and 2368 mhz flux points.,title="fig:",scaledwidth=40.0% ] | a point - like source of @xmath0tev @xmath1-rays has recently been seen towards the galactic center by hess and other air erenkov telescopes . in recent work @xcite
, we demonstrated that these @xmath1-rayscan be attributed to high - energy protons that ( i ) are accelerated close to the event horizon of the central black hole , sgr a * , ( ii ) diffuse out to @xmath0 pc scales , and ( iii ) finally interact to produce @xmath1-rays .
the same hadronic collision processes will necessarily lead to the creation of electrons and positrons . here
we calculate the synchrotron emissivity of these secondary leptons in the same magnetic field configuration through which the initiating protons have been propagated in our model .
we compare this emission with the observed @xmath0 ghz radio spectrum of the inner few pc region which we have assembled from archival data and new measurements we have made with the australia telescope compact array .
we find that our model predicts secondary synchrotron emission with a steep slope consistent with the observations but with an overall normalization that is too large by a factor of @xmath0 2 .
if we further constrain our theoretical @xmath1-ray curve to obey the implict egret upper limit on emission from this region we predict radio emission that is consistent with observations , i.e. , the hadronic model of gamma ray emission can , simultaneously and without fine - tuning , also explain essentially all the diffuse radio emission detected from the inner few pc of the galaxy . |
thanks to the introduction of the focusing capability , the x - ray astronomy is now providing data almost on all type of sources in the sky and the x - ray data are of common use in the astronomical community almost as much as the optical , near - infrared and radio data . however , this is true only in the soft - x - ray band , at energies between 0.1 and 10 kev . above 10 kev ,
i.e. in the so called hard x - ray band , so far we could use only passive collimators and coded mask . as a result
only a few hundred sources are known in the whole sky in the @xmath1 kev band .
however , it is in this band that the x - ray background has its peak , at @xmath2 kev and we still do not know which sources are making up most of this emission , a question of cosmological importance . to discover them we need an instrument with much higher sensitivity and imaging capability , to separate the various weak x - ray sources .
this band is also very important for the study of matter accretion around black holes and the particle acceleration mechanisms .
for these reasons , there is a clear need for a hard x - ray mission with a sensitivity and angular resolution similar to those of the current soft x - ray telescopes .
a hard x - ray focusing optics is needed to do this . with the emerging technology in mirror manufacturing , providing that one can achieve a very long focal length and uses multilayer reflecting coatings , it is now possible to develop such a mission . in particular simbol - x ,
a new mission currently under study by the french and italian agencies will have the capability to obtain x - ray images in the band @xmath3 kev with very good sensitivity and spatial resolution @xcite .
the simbol - x very long focal length of 20 m will be obtained by using a formation flying strategy ; two satellites will carry one the focusing mirrors and the other one the focal plane detector @xcite . while in order to meet the highly demanding image quality and effective area over the full @xmath4 kev energy band a mirror module of 100 nested shells is foreseen @xcite .
these shells will be nickel replicated and coated with a multilayer of more than 200 bilayers of platinum and carbon . here
we will describe the results that we obtained during the phase a study that has been completed in 2008 , during which we fabricated three engineering models .
m. below this value also the surface roughness of the nico - shell needs to be improved .
this problem will be further investigated during the phase b. ] the simbol - x mirrors will be electroformed ni shells with wolter i profile . the adopted technology has been successfully used for the gold coated x - ray mirrors of the beppo - sax , xmm - newton and jet - x / swift telescopes . this technology has been developed and consolidated in the past two decades in italy by the inaf - brera astronomical observatory in collaboration with the media lario technology company . for the simbol - x mirrors , a few important modifications of the process will be implemented : 1 ) the use of multilayer reflecting coatings , allowing us to obtain a larger fov and an operative range up to 80 kev and beyond ; 2 ) the ni walls will be a factor of two thinner than the xmm mirror shells , to maintain the weight as low as possible . with respect to the first point ,
once the gold - coated ni mirror shell has been replicated from the mandrel , the multilayer film will be sputtered on the internal surface of the shell by using a two - targets linear dc magnetron sputtering system @xcite .
this process has been developed at the sao - center for astrophysics ( cfa ) for monolithic pseudo cylindrical shells @xcite and now also at media lario where a multilayer coating facility has been developed and installed . to replicate mirror shells with the requested image quality we must start from mandrels with a very good surface roughness , even better than those produced for the xmm - newton mirrors .
in fact we need mandrels with a surface roughness of the order of 1 - 2 over the space wavelength up to a few microns . during phase
a we have developed four mandrels , three using the classical approach already used for the xmm - newton ones ( i.e. the mandrels are produced by grinding and then worked with a long lasting lapping procedure to reach the requested surface roughness ) and one using the diamond turning technique .
the latter approach produces mandrels with a better starting surface , therefore minimizing the lapping procedure .
the goal is to reduce the time for the lapping procedure by a factor of four .
the four mandrels have a diameter of 286 , 291 , 295 and 297 mm , respectively .
for all of them the surface roughness has been measured to be within the specification needed to meet the mission requirements , including the mandrel made with the diamond turning ( # 297 ) . as an example in fig .
1 we show the roughness profile that we measured with different metrological systems for mandrel # 291 .
25 arcsec , not very far away from the requirements .
see the text for more details .
, title="fig : " ] 25 arcsec , not very far away from the requirements .
see the text for more details .
, title="fig : " ] using mandrels # 286 , 291 and 295 we have replicated various mirror shells , using gold as release agent , and coated them with a multilayer of 95 tungsten and silicon ( w / si ) bilayers .
we used w / si because at the moment at the panter facility , where we performed the x - ray measurements , we can use x - ray only up to 50 kev and up to these energies w / si and pt / c have a similar behaviour , but the w / si is much cheaper . also these mandrels and shells
have a focal length of 10 m , once again because at the moment it is not possible to test mirrors with longer focal length at the panter facility of mpe , in munich .
it is already planned to upgrade this facility in order to test mirrors with focal length of 20 m and up to 100 kev .
we choose these diameters for the mandrels in order to test the same incident angles , in turn related to the reflectivity , that will have the largest simbol - x mirror shells with a focal length of 20 m. to test the integration procedure that we have developed for thin shells , we integrated these shells in three different engineering models , the first two with two shells ( # 295 & 291 ) and the third one with three shells ( # 286 , 291 & 295 ) ( we did not used mandrel # 297 because it was completed only at the end of this phase ) .
this procedure is based on the use of stiffening rings that are put around the shell and are removed after the shell has been glued on the spider .
we showed that this procedure meets the simbol - x requirements ( see basso et al .
these proceedings ) . for the first engineering model
the multilayer deposition has been performed at the cfa , while for the other two this has been performed at media lario . in fig .
2 , left and middle panels , we show two images taken at the panter facility at 0.93 kev and in the band 30 - 50 kev with the shell # 291 integrated in the second engineering model . as expected , at the higher energies the scattering is larger and the structure of the spider arms can be noted , still the image quality is extremely good and sharp up to 50 kev .
this is also shown in fig
. 2 right panel , where we plot the hew of this shell as a function of energy .
the hew below 10 kev is about 18 arcsec , while that one at 30 kev is @xmath5 25 arcsec , not very far away from the requirements .
the very slow increase of the hew from 1 to 45 kev , due to x - ray scattering superposing to the mirror figure error ( responsible for the low - energy hew ) , is probably due to a higher mirror surface roughness , with respect to that one required by the mission .
in fact in fig .
1 we can see that while the mandrel surface is very good , that one of the replicated ni shell is quite worse , in particular at the middle spatial wavelengths which are the ones influencing more the scattering .
this needs to be improved .
but we need to improve also at the higher spatial wavelengths and have a better mirror figure error and , therefore , a better hew also below 10 kev . to this end ,
media lario started a new program in order to electro - form shells in nickel - cobalt ( nico ) , a metal alloy that has better structural performances than the pure nickel , in particular it has a higher yield .
therefore these shells should be stiffer and more resistant to plastic deformation during the release process .
the third engineering model has been assembled with three shells already made in nico .
these shells , when measured on our optical bench , have shown an improvement of their intrinsic optical quality , therefore we would have expected an improvement of the hew as measured in the x - ray at the panter facility , at least at lower energies . however , due to an error during the integration process , some bolts were tightened too much , we degraded somewhat the image quality of the integrated shells .
therefore the values that we measured at the panter facility were only slightly better . in summary , we are confident that the use of the nico alloy will improve the intrinsic image quality of the mirror shells , but we do not have x - ray measurements to prove this , yet . however , from fig .
1 we can already see that the nico shells will expectedly have better optical performances .
note how the surface roughness measured for the nico shell ( color - continued lines ) is much better than that one of the pure ni - shell ( open square black points ) .
this is particularly true in the @xmath6 m spatial wavelength range .
however , below @xmath7 m the surface roughness of the two shells are very similar and quite worse with respect to the surface roughness of the mandrel .
therefore , we still need to improve over the release process in order to have better mirror roughness also at very low spatial wavelength , that have a stronger impact on the mirror performances at higher energies .
this is reflected also by the measurements of the effective area as a function of energy . in fig .
3 we compare the effective area measured at the panter facility for the ni - shell # 291 and the nico - shell # 286 , with the theoretical expectations . while for the ni - shell the measured values are @xmath5 10% lower than the predicted one below 10 kev , @xmath5 30% lower between 15 - 20 kev and @xmath5 50% lower at 30 kev ; for the nico - shell the measured values below 10 kev are in perfect agreement with the theoretical ones . from 15 to 20 kev
the nico - shell is still better than the ni - shell , although somewhat below the theoretical values . while above 25 kev also the nico - shell is well below the theoretical values .
in fact , for the nico - shell the measured values are more consistent with the theoretical model assuming a 8 roughness , instead of the requested 4 .
the phase a study for the realisation of the simbol - x mirror module has been completed , showing that the requirements of the mission can be meet . however , some further technological developments are necessary and will be investigated during phase b. the design of the mandrels has been optimized to reach very good longitudinal profiles and surface roughness able to provide an intrinsic hew of about 7 arcsec .
we showed that diamond turning technique can provide mandrels that have the surface quality that is needed by the mission the hew of each shells was not degraded in a significant way by the integration : the concept of the integration of thin shells with the stiffening rings works very well .
however , we need a better control on the behavior of the temporary structure during the integration process , in order to avoid mistakes as the one that occurred during the integration of the third engineering model .
we need a better control on the shell roughness degradation introduced by the electroforming and release processes , in particular at the spatial wavelength below @xmath8 m .
this aspect will be investigated and improved during phase b. ferrando p. , et al .
, _ spie proc . _ * 6266 * , 62660 ( 2006 ) la marle o. , _ spie proc . _
* 6686 * , 66860 ( 2007 ) pareschi g. , et al . , _ spie proc . _ * 7011 * , 70110n ( 2008 ) romaine s. , et al . , _ spie proc . _ * 5168 * , 65 ( 2006 ) | simbol - x will push grazing incidence imaging up to 80 kev , providing a strong improvement both in sensitivity and angular resolution compared to all instruments that have operated so far above 10 kev .
the superb hard x - ray imaging capability will be guaranteed by a mirror module of 100 electroformed nickel shells with a multilayer reflecting coating . here
we will describe the technogical development and solutions adopted for the fabrication of the mirror module , that must guarantee an half energy width ( hew ) better than 20 arcsec from 0.5 up to 30 kev and a goal of 40 arcsec at 60 kev . during the phase a , terminated at the end of 2008 ,
we have developed three engineering models with two , two and three shells , respectively .
the most critical aspects in the development of the simbol - x mirrors are _
i ) _ the production of the 100 mandrels with very good surface quality within the timeline of the mission , _ ii ) _ the replication of shells that must be very thin ( a factor of 2 thinner than those of xmm - newton ) and still have very good image quality up to 80 kev , _ iii ) _ the development of an integration process that allows us to integrate these very thin mirrors maintaining their intrinsic good image quality . the phase a study has shown that we can fabricate the mandrels with the needed quality and that we have developed a valid integration process .
the shells that we have produced so far have a quite good image quality , e.g. hew@xmath0 arcsec at 30 kev , and effective area .
however , we still need to make some improvements to reach the requirements
. we will briefly present these results and discuss the possible improvements that we will investigate during phase b. address = inaf - osservatorio astronomico di brera , via bianchi 46 , 23807 merate , italy address = inaf - osservatorio astronomico di brera , via bianchi 46 , 23807 merate , italy address = media lario technologies s.r.l .
, localit pascolo , 23842 bosisio parini , italy address = max - planck - institut fr extraterrestrische physik , giessenbachstr . 1 , 85748 garching , germany address = inaf - osservatorio astronomico di brera , via bianchi 46 , 23807 merate , italy , altaddress = media lario technologies s.r.l .
, localit pascolo , 23842 bosisio parini , italy address = inaf - osservatorio astronomico di brera , via bianchi 46 , 23807 merate , italy address = inaf - osservatorio astronomico di brera , via bianchi 46 , 23807 merate , italy address = inaf - osservatorio astronomico di brera , via bianchi 46 , 23807 merate , italy address = max - planck - institut fr extraterrestrische physik , giessenbachstr . 1 , 85748 garching , germany address = media lario technologies s.r.l . , localit pascolo , 23842 bosisio parini , italy address = harvard - smithsonian center for astrophysics , 60 garden street , cambridge , ma 02138 , usa address = max - planck - institut fr extraterrestrische physik , giessenbachstr . 1
, 85748 garching , germany address = media lario technologies s.r.l . ,
localit pascolo , 23842 bosisio parini , italy address = media lario technologies s.r.l .
, localit pascolo , 23842 bosisio parini , italy address = inaf - osservatorio astronomico di brera , via bianchi 46 , 23807 merate , italy address = harvard - smithsonian center for astrophysics , 60 garden street , cambridge , ma 02138 , usa address = inaf - osservatorio astronomico di brera , via bianchi 46 , 23807 merate , italy address = media lario technologies s.r.l . , localit pascolo , 23842 bosisio parini , italy address = media lario technologies s.r.l .
, localit pascolo , 23842 bosisio parini , italy |
a collection @xmath7 of @xmath1 hypersurfaces in @xmath2 decomposes @xmath2 into open cells of dimension @xmath8 and into ( relatively open ) faces whose dimension @xmath9 varies between 0 and @xmath10 .
each of these faces and cells is a maximal connected relatively open set contained in a fixed subset of @xmath7 and avoiding all other hypersurfaces .
we refer to @xmath11-dimensional faces as _ facets _ , to 0-dimensional faces as _ vertices _ , to 1-dimensional faces as _ edges _ , and to faces of dimension @xmath9 , where @xmath12 , simply as _
@xmath9-faces_. these cells and faces define a cell complex , called the _ arrangement _
@xmath13 of @xmath7 ( see @xcite ) .
the _ complexity _ of a cell in @xmath0 is defined to be the number of faces ( of all dimensions ) that are contained in the closure of the cell . for a set @xmath14
, the _ zone _ of @xmath15 is defined to be the set of all cells in @xmath0 that @xmath15 intersects .
the _ complexity _ of a zone is the sum of the complexities of the cells in the zone . for general hypersurfaces , studying the complexity of a zone in an arrangement
is closely related to the problem of analyzing the maximum possible complexity of a single cell ( see ( * ? ? ?
* section 5.3 ) ) . in this paper
we focus on the special case where @xmath7 is a collection of @xmath1 hyperplanes in @xmath2 . [
[ planar - arrangements . ] ] planar arrangements .
+ + + + + + + + + + + + + + + + + + + + it is a trivial fact that the number of edges or vertices of a single cell in an arrangement of @xmath1 lines in the plane is at most @xmath1 .
a linear bound on the complexity of a cell in an arrangement of halflines is also known ( see alevizos et al .
@xcite ) . on the other hand
, a result of wiernik and sharir @xcite on the lower envelopes of segments implies that a single cell in an arrangement of @xmath1 line segments in the plane can have @xmath16 vertices in the worst case , where @xmath17 is the inverse ackermann function . a matching upper
bound has been proved by pollack et al .
@xcite . to demonstrate the relation of these results to the zone problem ,
consider the following problem : let @xmath18 be a collection of @xmath1 lines in the plane and let @xmath15 be a given circle drawn in the plane .
what is the complexity of the zone of @xmath15 in @xmath19 ?
to solve this problem we use the following reduction ( following @xcite ) : for each intersection point of @xmath15 with a line @xmath20 , we split @xmath21 into two pieces and leave an arbitrarily small gap between these pieces . in this manner ,
all faces of the zone of @xmath15 are merged into one cell , now of an arrangement of ( at most @xmath22 ) lines , halflines and line segments ( see figure [ fig : zonecell ] ) .
the aforementioned results imply that the complexity of the zone in question is @xmath23 .
more precisely , the part of the zone outside the circle @xmath15 has complexity @xmath24 ( because , if we erase the portions of the lines inside the disk @xmath25 bounded by @xmath15 , the outer zone becomes a portion of a single cell in an arrangement of @xmath24 halflines and lines ) . on the other hand the complexity of the inner part of the zone ( which becomes a portion of a single cell in an arrangement of @xmath24 line segments
if we erase the portions of the lines outside @xmath25 ) is known only to be @xmath23 . whether this complexity bound ( of the inner part ) is tight , is a major open problem . in a line arrangement ( consisting all the cells between the two highlighted cycles ) .
right : turning the zone into a single cell.,scaledwidth=80.0% ] a major role in the derivation of these results is played by _ davenport - schinzel sequences _ , combinatorial structures that arise in the analysis of lower envelopes of collections of univariate functions .
analyzing the envelope of a collection of multivariate functions appears to be considerably harder than the univariate case and the known bounds on its complexity are not as sharp as in the planar case ( for more details see the book @xcite and the survey paper @xcite ) .
[ [ higher - dimensions . ] ] higher dimensions .
+ + + + + + + + + + + + + + + + + + a fundamental result on _ hyperplane _ arrangements is the ( standard ) zone theorem @xcite , in which @xmath15 is assumed to be a hyperplane distinct from those in @xmath7 .
it asserts that the zone of a hyperplane in an arrangement of @xmath1 hyperplanes in @xmath2 has complexity @xmath26 , where the constant of proportionality depends on @xmath8 . in @xcite ,
the zone theorem is extended to cases where @xmath15 is either an algebraic variety or the boundary of an arbitrary convex set .
it is shown in @xcite that the complexity of the zone in these settings is @xmath27 , where the constant of proportionality depends on @xmath8 and , in the former case , also on the degree of @xmath15 . in this paper
, we consider hyperplane arrangements in @xmath2 and the zone of the boundary of an arbitrary convex set in @xmath2 .
we show ( in theorem [ main ] ) that , similar to the planar case , taking into account only the outer part of the zone ( precise definition given below ) has complexity @xmath5 ( without the logarithmic factor ) . as in the planar case , whether this bound also holds ( in the worst case ) for the complexity of the inner part of the zone is still an open question . _ acknowledgment .
_ i would like to thank my advisor prof .
micha sharir for his help , advice , useful discussions and support .
let @xmath7 be a collection of @xmath1 hyperplanes in @xmath2 and let @xmath28 be a closed convex set in @xmath2 .
we denote the zone of @xmath29 in @xmath0 as @xmath30 , and define the _ outer complexity _ of @xmath3 ( with respect to @xmath28 ) to be @xmath31 where @xmath32 and @xmath33 is the number of @xmath34-faces on the boundary of the cell @xmath35 that are _ contained _ in @xmath36 .
we assume that @xmath7 and @xmath28 are in _ general position _ , i.e. , any @xmath37 hyperplanes in @xmath7 intersect in a @xmath38-flat , no @xmath39 hyperplanes have a point in common , and @xmath29 is not tangent to any flat formed by the intersection of any number @xmath37 hyperplanes in @xmath7 ( in particular , @xmath29 does not meet any vertex of @xmath0 ) .
this assumption involves no loss of generality .
this can be proved using a standard perturbation argument : displacing the hyperplanes of @xmath7 slightly will put @xmath28 and @xmath7 in general position , and can only increase the complexities of the outer part of the cells in @xmath3 ( see @xcite ) .
note that a @xmath9-face @xmath40 in @xmath0 lies in @xmath41 hyperplanes of @xmath7 and is on the boundary of @xmath42 cells of @xmath0 .
more than one of these cells can belong to @xmath3 , and thus the contribution of @xmath40 to @xmath43 can be larger than one .
it is therefore more convenient to consider pairs @xmath44 , where @xmath45 and @xmath40 is a @xmath9-face in @xmath0 that lies on the boundary of @xmath35 .
each such pair contributes at most one to the count ( it contributes when @xmath46 ) . following a similar notation in @xcite ,
we call such a pair a _ @xmath9-border_. [ main ] let @xmath7 be a collection of @xmath1 hyperplanes in @xmath2 , let @xmath47 be a closed convex set , and put @xmath30 .
then @xmath48 where @xmath49 is an absolute constant independent of @xmath8 . in the definition of the outer complexity of @xmath3 we ignore ,
not only faces that lie fully inside of @xmath28 , but also faces that intersect @xmath29 .
it is natural though , to include faces of the latter kind in the complexity of the outer zone , and it is not hard to handle such faces , and to show that their total number is @xmath5 . indeed , as is shown in ( * ? ? ?
* lemma 2.3 ) , the number of ( @xmath8-dimensional ) cells in @xmath3 is @xmath5 . to bound the number of @xmath9-faces that intersect @xmath29
, one can simply restrict @xmath0 to each of the @xmath9-flats , formed by the intersection of some @xmath41 hyperplanes of @xmath7 , and apply this bound on the number of cells to the restricted ( lower - dimensional ) arrangements . a more detailed discussion of this issue is given at the end of the analysis .
* proof of theorem [ main ] .
* for each @xmath50 , we put @xmath51 that is , @xmath52 is the number of @xmath34-borders @xmath44 such that @xmath35 is in @xmath3 and @xmath40 is ( an @xmath34-face on the boundary of @xmath35 which is ) contained in @xmath36 . by definition , @xmath53 and
so it is sufficient to bound each of the quantities @xmath52 . to bound @xmath57 ,
consider an edge @xmath58 contained in @xmath36 and incident to a cell in @xmath3 , and let @xmath59 be the line that supports @xmath58 .
we use the property that no such edge @xmath58 is incident to more than one cell in @xmath3 ( this is a special case of a more general property established in claim [ clm : nopop ] below ) , and thus the number of these edges is the same as the number of 1-borders we are interested in . we may assume that @xmath58 has at least one vertex , since the lines in @xmath7 are in general position , and we may assume that @xmath60 .
suppose first that @xmath58 has two vertices @xmath61 . for @xmath62 , we let @xmath63 be the second line ( besides @xmath64 ) defining the vertex @xmath65 and let @xmath66 denote the halfplane bounded by @xmath67 that contains @xmath58
. then there exists @xmath68 such that the halfline @xmath69 does not intersect @xmath29 ( and @xmath28 ) ; indeed , if this were not the case then by the convexity of @xmath28 the edge @xmath58 would have intersected @xmath28 as well ( in fact would be contained in @xmath28 ) , contrary to our assumption that @xmath70
. we charge @xmath58 to the pair @xmath71 .
if @xmath58 has only one vertex @xmath72 , we let @xmath21 be the second line ( besides @xmath64 ) defining @xmath72 and charge @xmath58 to @xmath73 ; here too , @xmath58 is a halfline which ( trivially ) does not intersect @xmath29 .
we claim that for every line @xmath74 , there are at most four pairs of the form @xmath75 charged by edges @xmath58 as described above ( that is , @xmath21 is a line of @xmath7 and @xmath76 is a halfline of another line that ends on @xmath21 ) .
more precisely , we will show that there are at most two pairs @xmath75 of this kind such that @xmath77 , for each of the two halfplanes @xmath78 bounded by @xmath21 . indeed , suppose that there were three such pairs @xmath79 , charged by three different edges as before and with @xmath80 , for @xmath81 .
let @xmath82 , for @xmath81 , be the line that supports @xmath83 and let @xmath84 be the halfplane bounded by @xmath85 that contains @xmath86 .
since the halflines lie on the same side of @xmath21 , their containing lines @xmath85 are all distinct . moreover , since the @xmath83 s are disjoint from @xmath28 , the halfplanes @xmath84 are well defined .
consider the intersection @xmath87 ( see figure [ fig : lem22 ] for an illustration ) . since @xmath49 is convex
, @xmath21 supports at most one edge on the boundary of @xmath49 and thus for some @xmath88 , the vertex @xmath89 is not of the boundary of @xmath49 .
let @xmath58 be the edge supported by @xmath90 incident to @xmath91 . by assumption @xmath58
is on the boundary of a cell @xmath92 .
but then , @xmath93 , and hence it can not intersect @xmath29 , which is a contradiction .
thus , @xmath94 finally , observe that @xmath95 since for every 0-border @xmath96 with @xmath97 , @xmath91 is incident to an edge @xmath58 on the boundary of @xmath35 such that either @xmath58 is fully contained in @xmath36 or @xmath58 intersects @xmath29 .
each line generates at most two segments of the latter kind , which can generate at most four @xmath98-borders .
together with inequality ( [ 2 ] ) , this gives @xmath99 as asserted .
we use induction on the dimension @xmath8 , where the induction base , for @xmath6 , has already been established in lemma [ lem : dim2 ] . for the induction step , fix @xmath101 and assume that ( [ 1 ] ) holds for all @xmath105 and @xmath106 .
[ lem : rec ] for each @xmath107 , we have @xmath108 where @xmath109 , viewed as a collection of @xmath110 hyperplanes ( of dimension @xmath111 ) in @xmath64 ( which is a translated and rotated copy of @xmath112 ) .
we defer the proof of lemma [ lem : rec ] for later and assume for the moment that the lemma holds .
together with the induction hypothesis ( applied with the parameters @xmath113 and @xmath110 ) , the lemma implies @xmath114 for @xmath34 fixed , we use induction on @xmath1 . for @xmath115
there is exactly one @xmath34-face @xmath40 in @xmath116 and at most @xmath117 @xmath34-borders @xmath44 with @xmath45 , thus ( [ 1 ] ) holds . for larger values of @xmath1 we have , by the induction hypothesis , @xmath118\\ & \leq 4(d - i)!\binom{n}{d - i}n^{i-1}\left(1-\frac{1}{n}\right)^{i-2}\\ & \leq 4(d - i)!\binom{n}{d - i}n^{i-1}\;\;\;\text{(because $ i\geq 2$).}\end{aligned}\ ] ]
this establishes the induction step and thus completes the proof of proposition [ prop : tauid ] .
@xmath100 we next provide the proof of lemma [ lem : rec ] .
we need the following simple property ( which is illustrated in figure [ fig : popfacet ] ) , whose planar case has already been used in the proof of lemma [ lem : dim2 ] .
let @xmath40 be a facet in @xmath0 such that @xmath119 and assume that @xmath40 is on the boundary of two distinct cells @xmath121 in @xmath3 .
observe first that @xmath122 is convex removal of the hyperplane supporting @xmath40 merges @xmath121 , and ( the missing ) @xmath40 into a single , necessarily convex , cell .
suppose for the sake of contradiction that @xmath40 does not intersect @xmath29 .
since @xmath119 this implies that @xmath46 .
now let @xmath123 and @xmath124 ; @xmath125 exist since @xmath121 are both in @xmath3 .
then the segment @xmath126 connecting the points @xmath127 and @xmath128 intersects @xmath129 , because @xmath130 and @xmath131 lie on different sides of it . but
@xmath122 is convex and so @xmath126 intersects @xmath40 ( because @xmath132 ) .
however , since @xmath28 is a convex ( and closed ) set , every segment connecting two points on its boundary must be contained in @xmath28 .
but we have shown that @xmath46 , which is a contradiction .
thus @xmath40 intersects @xmath29 .
@xmath100 .
for a convex set @xmath28 any facet @xmath40 that is incident to two cells @xmath133 must intersect @xmath29 . this property does not have to hold when @xmath28 is non - convex , as is indicated by the dashed line . ]
we note that the convexity of @xmath28 is crucial for claim [ clm : nopop ] to hold .
an instance where claim [ clm : nopop ] fails for a non - convex @xmath28 is depicted in figure [ fig : popfacet ] ( where the dashed line depicts the relevant portion of @xmath29 ) .
* proof of lemma [ lem : rec ] . * remove an arbitrary hyperplane @xmath64 from @xmath7 and consider @xmath134 .
our goal is to bound the increase in the number of @xmath34-borders in @xmath116 that _ are not contained in @xmath64 _ , over their number in @xmath135 . for this , let @xmath136 be an @xmath34-border in @xmath137 , such that @xmath138 is an @xmath34-face contained in @xmath36 . now reinsert @xmath64 into the arrangement .
we are only interested in the situation where @xmath136 is split by @xmath64 into two @xmath34-borders in @xmath139 and both of the cells of these borders are in @xmath3 , because only then will our count go up .
note that in this case @xmath140 is a facet in @xmath0 which lies on the boundary of the two zone cells @xmath141 and @xmath142 , where @xmath143 denote the two open halfspaces bounded by @xmath64 . by claim [ clm : nopop ]
, @xmath40 intersects @xmath29 ( this follows since @xmath144 so @xmath145 and this implies that @xmath40 meets @xmath36 ) .
thus @xmath146 , where @xmath147 , is an @xmath148-border in the @xmath11-dimensional arrangement @xmath149 ( within @xmath64 ) with respect to @xmath150 .
moreover , @xmath58 is contained in @xmath36 ( because @xmath151 is ) .
thus we may charge the split of @xmath136 to @xmath146 in a unique fashion . to sum up ,
if we consider only @xmath34-borders with @xmath34-faces not supported by @xmath64 , the insertion of @xmath64 into @xmath137 increases the count by at most @xmath152 . summing these bounds over all hyperplanes @xmath59 , and using the fact that the number of hyperplanes not supporting a specific @xmath34-face is @xmath153 ,
complete the proof of the lemma .
@xmath100 first , observe that @xmath157 indeed , let @xmath158 be a 1-border such that @xmath35 is a zone cell and @xmath58 is an edge on @xmath159 contained in @xmath36 .
obviously ( using the general position assumption ) , the vertices @xmath160 incident to @xmath58 are also inside @xmath36 , and thus the 0-borders @xmath161 contribute both to the count in @xmath162 ; we charge @xmath158 to them .
since every 0-border can be charged by at most @xmath8 1-borders , the inequality ( [ eq : tau1 ] ) follows . to bound @xmath162
, we will consider 3-faces @xmath163 in @xmath0 ( that bound cells of @xmath3 ) and use euler s formula to bound the number of vertices incident to @xmath163 in terms of the number of 2-faces incident to @xmath163 .
this is a standard technique , already used in @xcite , and it produces ( as we show below ) a bound on @xmath162 using the bound that we already have on @xmath164 .
however , since in our setting not all the 2-faces on the boundary of @xmath163 are necessarily contained in @xmath36 , and therefore not all of them necessarily contribute to the count in @xmath164 , one needs to be somewhat more careful using euler s formula . to address this technical issue , we separate our analysis into the following two cases .
let @xmath96 be a 0-border such that @xmath35 is a zone cell and @xmath91 is a vertex on @xmath159 contained in @xmath36 .
consider first the the case where there exists a 2-border @xmath44 , such that @xmath91 is incident to @xmath40 , and the 2-face @xmath40 intersects @xmath29 . in this case , pick such a 2-border @xmath44 and charge @xmath96 to the pair @xmath165 .
next , consider the unique plane @xmath166 that supports @xmath40 , formed by the intersection of some @xmath111 hyperplanes in @xmath7 .
then @xmath165 is a 0-border in the 2-dimensional arrangement @xmath167 , where @xmath168 is a collection of @xmath169 lines in @xmath166 .
moreover , @xmath40 is a zone cell in @xmath170 and @xmath91 is contained in @xmath171 . by lemma [ lem : dim2 ] , there are at most @xmath172 0-borders on @xmath166 of this form . going over all the @xmath173 planes ,
each formed by the intersection of some @xmath111 hyperplanes in @xmath7 , gives a bound of at most @xmath174 on the number of such pairs @xmath165 .
hence , the number of 0-borders @xmath96 in @xmath0 such that @xmath91 is contained in @xmath36 and is incident to a 2-face on the boundary of @xmath35 that intersects @xmath29 , is at most @xmath175 where the factor @xmath176 arises since every 2-face @xmath40 is incident to exactly this number of cells @xmath35 in @xmath0 and therefore @xmath165 can be charged by at most these many 0-borders of the form @xmath96 . in the complementary case , we consider 0-borders @xmath96 such that @xmath35 is a cell in @xmath3 and @xmath91 is a vertex of @xmath35 inside @xmath36 , such that all the 2-faces @xmath40 on the boundary of @xmath35 that are incident to @xmath91 , do not intersect @xmath29 ( in other words , all these 2-faces are fully contained in @xmath36 ) .
let @xmath177 be a 3-border with @xmath163 incident to @xmath91 , and denote by @xmath178 the unique 3-flat that supports @xmath163 , which is the intersection of some @xmath179 hyperplanes of @xmath7 .
we put @xmath180 , which is a collection of @xmath181 planes in @xmath178 .
we assign to each plane in @xmath182 an orientation , letting its positive side be the one that contains @xmath163 . then @xmath163 is a 3-dimensional convex polyhedron in @xmath178 , which is the intersection of the halfspaces , defined by the positive sides of the planes in @xmath182 .
we want to use euler s formula on @xmath163 , but wish to consider only facets of @xmath163 that are contained in @xmath36 .
let @xmath183 be the subset of planes @xmath184 such that @xmath64 supports a facet of @xmath163 that is contained in @xmath36 .
denote by @xmath185 the polyhedron in @xmath178 , formed by the intersection of the halfspaces defined by the positive sides of the planes in @xmath183 . by assumption
, @xmath91 is also a vertex of @xmath185 , because it is incident only to facets of @xmath163 contained in @xmath36 and thus we did not remove any of the hyperplanes that support it .
so the number of vertices of @xmath185 , using euler s formula , is at most twice the number of facets of @xmath185 .
but , by definition , this is exactly the number of facets of @xmath163 that are fully contained in @xmath36 .
( note that the facets of @xmath185 need not be fully contained in @xmath36 : each such facet @xmath186 is an expansion of a facet @xmath40 of @xmath163 ; while @xmath40 is fully contained in @xmath36 , @xmath186 does not have to be . ) finally , every 2-face @xmath40 on the boundary of @xmath35 is incident to exactly @xmath187 3-faces @xmath163 on the boundary of @xmath35 , and similarly , every vertex @xmath91 on the boundary of @xmath35 is incident to exactly @xmath188 such 3-faces @xmath163 .
therefore , the number of 0-borders @xmath96 , with @xmath91 and @xmath35 as above , is at most @xmath189 finally , returning to the proof of theorem [ main ] , we note that the constants of proportionality in the bounds for @xmath162 and @xmath57 are both at most some absolute constant ( in fact , they tend to 0 as @xmath8 increases ) .
this , and the bound in proposition [ prop : tauid ] , establish the bound in theorem [ main ] .
@xmath100 returning to the remark following the statement of theorem [ main ] , we complete the analysis to include also faces that intersect @xmath29 .
we fix @xmath191 , and a @xmath9-flat @xmath166 formed by the intersection of some @xmath41 hyperplanes of @xmath7 .
every @xmath9-face of @xmath3 that is contained in @xmath166 and intersects @xmath29 is on the boundary of exactly @xmath42 cells of @xmath3 , and is a cell in the zone of @xmath192 in the @xmath9-dimensional arrangement @xmath193 .
since the number of such cells is @xmath194 , where @xmath195 is an absolute constant independent of @xmath8 , and the number of @xmath9-flats @xmath166 is @xmath196 , the overall number of faces of @xmath3 that intersect @xmath29 is @xmath197 in conclusion , we obtain the following strengthening of theorem [ main ] . the overall number of faces of the cells of the zone of the boundary of an arbitrary convex set @xmath28 in an arrangement of @xmath1 hyperplane in @xmath2 , that are either fully contained in , or intersect @xmath36 , is at most @xmath198 , for some absolute constant @xmath199 , independent of @xmath8 and of @xmath28 . 9 agarwal , p.k . and sharir , m. , davenport - schinzel sequences and their geometric applications , in _ handbook of computational geometry _ , j .-
sack and j. urrutia editors , elsevier science b.v . ,
north holland , 1999 : 147 .
alevizos , p. , boissonnat , j.d and preparata , f.p .
, an optimal algorithm for the boundary of a cell in a union of rays , _ algorithmica _
5(14 ) ( 1990 ) : 573590 .
aronov , b. , pellegrini , m. and sharir , m. , on the zone of a surface in a hyperplane arrangement , _ discrete comput .
geom . _ 9(1 ) ( 1993 ) : 177186 .
edelsbrunner , h. , seidel , r. and sharir , m. , on the zone theorem for hyperplane arrangements , _ siam j. comput . _ 22(2 ) ( 1993 ) : 418429 .
edelsbrunner , h. , _ algorithms in combinatorial geometry _ , springer - verlag , berlin heidelberg , 1987 .
grunbaum , b. , _ convex polytopes _ , wiley , new york , 1967 .
orlik , p. and terao , h. , _ arrangements of hyperplanes _ , springer - verlag , berlin heidelberg , 1992 .
pollack , r. , sharir , m. and sifrony , s. , separating two simple polygons by a sequence of translations , _ discrete comput .
geom . _ 3(1 ) ( 1988 ) , 123136 .
sharir , m. and agarwal , p.k . , _ davenport - schinzel sequences and their geometric applications _ , cambridge university press , cambridge - new york - melbourne , 1995 .
wiernik , a. and sharir , m. , planar realizations of nonlinear davenport - schinzel sequences by segments , _ discrete comput .
geom . _ 3(1 ) ( 1988 ) : 1547 . | we consider an arrangement @xmath0 of @xmath1 hyperplanes in @xmath2 and the zone @xmath3 in @xmath0 of the boundary of an arbitrary convex set in @xmath2 in such an arrangement .
we show that , whereas the combinatorial complexity of @xmath3 is known only to be @xmath4 @xcite , the outer part of the zone has complexity @xmath5 ( without the logarithmic factor ) .
whether this bound also holds for the complexity of the inner part of the zone is still an open question ( even for @xmath6 ) . |
the discovery of the expansion of the universe and the concept of a singularity from which the universe emerged ( big bang ) has an interesting history @xcite . in this introduction
, we provide a brief review of the early development of modern cosmology , describing very elementary ( yet fundamental ) cosmological models that will play an important role in our study . in 1917 , einstein @xcite applied the equations of general relativity to cosmology , assuming that the universe is homogeneous and isotropic . on the ground of cultural and philosophical beliefs , he considered a static universe . in order to accommodate for such a solution , he had to introduce a cosmological constant @xmath26 in the equations of general relativity , so as to obtain a static solution @xmath27 . in einstein
s belief , the newtonian effect of the cosmological constant was to shield the gravitational interaction on a distance @xmath28 .
this is actually incorrect .
lematre @xcite was the first to understand that the cosmological constant can be interpreted as a force of `` cosmic repulsion '' . in the newtonian context , the modified poisson equation including the cosmological constant is @xmath29 , leading to the static solution @xmath30 . ] .
he obtained a model of the universe with a positive curvature @xmath31 that is finite ( yet unbounded ) with a `` radius '' @xmath32 , a density @xmath33 , and a mass @xmath34 .
the same year , de sitter @xcite discovered another static solution of the einstein equations ( according to the review of robertson @xcite , this solution was suggested to de sitter by ehrenfest ) .
this universe is empty ( @xmath35 ) and has a radius @xmath36 . in this model ,
the stars and nebulae are just `` test particles '' that do not curve the universe .
de sitter managed to predict that the light emitted by these sources should be redshifted because ( i ) they move away from the observer due to the effect of the cosmological constant , and ( ii ) there is a slowing down of time with increasing distance from the observer ( de sitter effect ) .
he mentioned that some observations ( presumably those made by slipher since 1912 ) seem to confirm that claim .
he also predicted a linear , or quadratic , relationship between the redshift and the distance .
however , his interpretation of the redshifts remained enigmatic for a long time .
therefore , the early cosmological debate revolved around two deficient models : the model of einstein which was unable to explain redshifts , and the model of de sitter which contained no matter and was rather mysterious .
both were built on the assumption that the universe is static . in 1922 and 1924 , friedmann @xcite discovered non - static cosmological solutions of the einstein equations describing the temporal variation of space .
he considered the case of spherical ( positive curvature @xmath31 ) and hyperbolic ( negative curvature @xmath37 ) universes .
he described expanding and contracting solutions , as well as cyclic solutions , and foresaw the possible beginning of the universe in a singularity .
he obtained for the first time an estimate of the age of the universe of the order of @xmath38 billion years . at that time
, einstein did not believe in the significance of these non stationary solutions .
he first argued that the friedmann solutions were not compatible with the field equations of general relativity @xcite , then retracted his statement @xcite but remained convinced that the friedmann solutions are not physically relevant @xcite .
apart from einstein , who criticized them , the papers of friedmann remained unnoticed by the physical community until 1930 . in 1922 ,
lanczos @xcite found a non - stationary form of de sitter s solution in which the scale factor increases exponentially rapidly .
he had the key to an expanding universe but did not recognize its physical significance . in 1923 , weyl @xcite published a paper in which he tried to find his own way to de sitter s redshift .
he noted that the particles in de sitter s universe move away from each other with a velocity proportional to their distance , but did not provide an explicit system of coordinates to make his claim more precise . in 1925
, lematre @xcite found a system of coordinates in which de sitter s model is non - static and leads to a linear relationship between the velocity ( redshift ) and the distance .
this model explains the redshifts of distant objects in an unambiguous way .
this model is euclidean @xmath39 ( which was problematic for lematre ) and the scale factor increases exponentially rapidly with time due to the cosmological constant . in 1927 , lematre @xcite rediscovered independently from friedmann the non - stationary cosmological solutions of the einstein equations .
in addition , having spent two years in harvard and at the mit , he was aware of cosmological observations and related the expansion of space predicted by the theory of general relativity to the recession motion of galaxies observed by slipher , hubble and strmberg .
he interpreted the redshift of the nebulae as a consequence of the expansion of the universe , predicted a linear relationship between the recession velocity and the distance ( the nowdays called `` hubble relationship '' ) and , using available data @xcite , obtained an estimate of the `` hubble constant '' @xmath40 two years before hubble @xcite .
the work of lematre , like the one of his predecessor friedmann , was still not appreciated by einstein who found it `` abominable '' @xcite . in 1928 and 1929 , robertson @xcite reinterpreted de sitter s solution dynamically .
however , like lanczos @xcite and weyl @xcite , he did not `` see '' the expanding universe . in 1930 ,
eddington , who was working with mcvittie on possible dynamical solutions of the cosmological einstein equations ( following a discussion that he had with de sitter during a meeting of the royal astronomical society @xcite ) , was alerted by lematre about his own results .
he publicized them in a paper @xcite in which he showed that the einstein static universe is unstable ( a result that was implicit in the paper of lematre ) .
he also emphasized a non - singular solution of lematre in which the universe starts from the static solution of einstein for @xmath41 , then expands ( this model is often called the eddington - lematre model ) . in parallel , de sitter advertised lematre s discovery of an expanding universe in several publications @xcite .
together with the supporting observations of hubble @xcite , and independent theoretical work by robertson @xcite , this really meant the demise of the static universe . in 1931 ,
einstein was forced to admit the theoretical and observational reality of the expansion of the universe .
he took an extreme position and renounced to the cosmological constant @xcite that he considered being the `` biggest blunder '' of his life ( in contrast , lematre @xcite always defended the cosmological constant as one of the most fundamental ingredients of modern cosmology ) . in 1932 , einstein published with de sitter a one - page paper @xcite in which they consider an expanding universe in a euclidean space ( zero curvature @xmath39 ) without pressure ( @xmath42 ) and without cosmological constant ( @xmath43 ) .
this is the celebrated einstein - de sitter ( eds ) solution that describes the matter era .
the main motivation of these authors was to show that an expansion of the universe is possible even without the cosmological constant . in the meantime
, lematre had gone one step further .
if the universe is expanding , it may have started from a _ singularity _ in the past . in 1931 , he proposed a model @xcite in which the universe begins its expansion from a singular initial state , or more precisely , from a near - singularity that he called the `` primeval atom '' @xcite .
this expansion is followed by a phase of stagnation during which the universe remains close to the static einstein solution . in lematre s model , it is during this phase of stagnation that galaxies form . at later times
, the universe enters in the de sitter phase of exponential expansion .
a new controversy started with einstein who did not believe in the singular model of lematre that implies that the universe has a `` beginning '' .
this controversy remained until the end of their life . on the other hand , in 1960 , fred hoyle , who promoted a `` steady state theory '' @xcite in which the universe has always been identical to itself and in which matter is created spontaneously , and continuously , made fun of lematre during a meeting in pasadena by introducing him as the `` big bang man '' @xcite .
this is how the term `` big bang '' was introduced in cosmology and became popular . up to the mid-1960 s , it was not clear whether the early universe had been hot or cold .
the discovery by penzias and wilson @xcite of the @xmath44 cosmic microwave background radiation ( cmb ) , whose existence had been predicted by the hot universe theory @xcite , immediately led widespread acceptance to the big bang theory . according to that theory , the early universe was filled with an ultrarelativistic gas of photons , electrons , positrons , quarks , antiquarks , etc .
this is described by an equation of state @xmath45 from which it is easily deduced that the scale factor evolves as @xmath46 , the density as @xmath47 , and the temperature as @xmath48 . at @xmath20 ,
the scale factor vanishes while the energy density and the temperature become infinite .
this is why the point @xmath20 is known as the point of the initial cosmological singularity ( big bang ) .
as the universe expands , the density and the temperature decrease .
when the universe is sufficiently cooled , particles and antiparticles annihilate each other , the photon - gas energy density falls off relatively rapidly and the main contribution of the matter density starts to come from the small excess of baryons over antibaryons , as well as from other fields and particles which now comprise the dark matter .
however , the hot universe theory suffers from a series of difficulties .
the first difficulty is the singularity problem .
the energy density goes to @xmath49 at @xmath20 and the solution can not be formally continued for @xmath12 .
a natural question is : `` what was before @xmath20 ? '' .
of course , one can argue that space and time started with the big bang , so that this question is meaningless .
however , this answer is not quite satisfactory . on the other hand , close to the big bang
, the density can be arbitrarily large .
this is in conflict with the laws of physics .
when the density becomes high enough , quantum effects should be taken into account in the problem . as a result , the planck density @xmath50 , the planck length @xmath51 , the planck mass @xmath52 , the planck temperature @xmath53 , and the planck time @xmath54 should play a fundamental role in the very early universe .
the standard big bang theory suffers from other problems that are known as the flatness problem ( or the fine - tuning problem ) , the horizon problem ( or the causality problem ) , and the monopole problem .
these problems can be solved at one stroke in the framework of inflationary cosmology @xcite . in that scenario ,
the universe was in an unstable vacuum - like state described by some hypothetical scalar field @xmath55 with its origin in quantum fluctuations .
this leads to an equation of state @xmath56 , implying a constant energy density @xmath57 , of the order of the planck density @xmath58 , called the vacuum energy . as a result of this vacuum energy ,
the universe expands exponentially rapidly in a tiniest fraction of a second .
this `` inflation '' can solve the above - mentioned problems . at the same time , the inflation theory predicts the scale invariant spectrum of the primordial density fluctuations and explains why the universe is spatially flat ( or behaves _ as if _ it were spatially flat due to the smallness of @xmath59 ) .
this is therefore a very attractive theory . in this paper , we propose a model of the early universe based on a polytropic equation of state with index @xmath10 . before that
, we first explain how we came to such a description . in previous works @xcite
, we explored the possibility that the universe is filled with self - gravitating bose - einstein condensates ( bec ) with short - range interactions ( see a detailed list of references in @xcite and recent additional references in @xcite ) . in the strong coupling limit ( the so - called thomas - fermi approximation ) , the equation of state of the bec is @xmath60 , where @xmath61 is the scattering length @xcite .
this is the equation of state of a polytrope with index @xmath10 @xcite .
initially , this model was introduced in order to describe dark matter halos .
a nice feature of the bec model is that it avoids the presence of cusps at the center of the halos because of the effective repulsion due to the heisenberg principle or because of the self - interaction of the bosons which plays the same role as the pauli exclusion principle for fermions .
since cusps are not observed in real dark matter halos , the bec model ( or the fermion model ) is favored as compared to the standard cold dark matter ( cdm ) model that predicts cuspy density profiles .
of course , at large scales , these models become equivalent .
the bec model was then applied to cosmology in order to describe the evolution of the universe as a whole @xcite . in @xcite
, we investigated the evolution of a `` bec universe '' with an equation of state of the form @xmath62 .
we considered the case of repulsive ( @xmath7 ) and attractive ( @xmath8 ) self - interaction .
more generally , we studied a `` cosmic fluid '' described by an equation of state of the form @xmath63 . when @xmath39 , we recover the standard linear equation of state @xmath64 describing radiation ( @xmath2 ) , pressureless matter ( @xmath3 ) , stiff matter ( @xmath4 ) , and vacuum energy ( @xmath5 ) .
therefore , the generalized equation of state @xmath65 can be viewed as the sum of a linear term describing a `` classical '' universe and a quadratic term due to the bec . at late times , when the density is low , the effect of the bec is negligible and we recover the classical universe . on the contrary ,
the effect of the bec becomes important in the early universe where the density is high .
therefore , it is interesting to study how the bec affects the early evolution of the universe . in our previous paper @xcite , we assumed that becs form after the radiation era , when the temperature has sufficiently decreased .
this prevented us from extrapolating the solution before the radiation era ( see the remark in sec .
8.2 of @xcite ) . in the present paper
, we explore the possibility that the equation of state @xmath62 may hold _ before _ the radiation era .
therefore , we assume that the early universe is described by an equation of state of the form @xmath63 with @xmath2 . actually , the study of this generalized equation of state is interesting even if its origin is not connected to becs .
furthermore , for this class of models , the friedmann equations can be solved analytically , which is an additional source of motivation .
first results were given in @xcite where it was shown that the repulsive and attractive models behave very differently .
for repulsive self - interactions ( @xmath7 ) , the universe starts at @xmath20 from a singularity at which the density is infinite but the radius finite .
for attractive self - interactions ( @xmath8 ) , the universe always existed in the past . for @xmath41
, its density tends to a constant value while its radius tends to zero exponentially rapidly . in the present paper
, we complete the study of the generalized equation of state @xmath63 .
first , we analyze its thermodynamical properties and study the temporal evolution of the temperature . secondly , for @xmath8
, we show that the exponential growth of the scale factor can account for a phase of inflation with constant density that we identify with the planck density @xmath58 . finally ,
taking @xmath2 , we study the transition between a pre - radiation era and the radiation era .
we mention the analogy with a second order phase transition where the control parameter is the time @xmath66 , the order parameter is the scale factor @xmath67 , and the planck constant @xmath24 plays the role of finite size effects ( the standard big bang theory is recovered for @xmath25 ) . for @xmath8 ,
the radius of the universe passes from @xmath68 at @xmath69 to @xmath15 at @xmath70 before entering in the radiation era and increasing algebraically as @xmath71 . during the inflationary phase
, the density remains approximately constant with the planck value ( @xmath72 ) before decreasing as @xmath73 .
the temperature passes from @xmath74 at @xmath69 to @xmath75 at @xmath76 before decreasing algebraically as @xmath77 .
it achieves its maximum value @xmath78 at @xmath79 . for @xmath8 , the initial radius of the universe @xmath80
is already `` large '' so there is no inflation .
the velocity of sound is less than the speed of light only for @xmath81 , corresponding to @xmath82 , @xmath83 , and @xmath84 .
this marks the beginning of the physical universe in this model .
the paper is organized as follows . in sec .
[ sec_basic ] , we recall the basic equations of cosmology and discuss famous cosmological models . in secs . [ sec_ges ] and [ sec_dark ] ,
we study the generalized equation of state @xmath85 for any value of the parameters @xmath86 , @xmath87 and @xmath9 .
we discuss two classes of solutions depending whether @xmath87 is positive or negative . in secs .
[ sec_inf ] and [ sec_ni ] , we consider a specific model of physical interest corresponding to @xmath2 , @xmath10 , and @xmath88 . for @xmath8 ( negative polytropic pressure )
, we get a model of inflationary universe without singularity .
it describes in a unified manner the transition from a pre - radiation era to the radiation era . for @xmath7 ( positive polytropic pressure )
, we get a model of non - inflationary universe with a new type of initial singularity . in sec .
[ sec_analogy ] , we develop an analogy with second order phase transitions where the planck constant plays the role of finite size effects .
in this section , we recapitulate the basic equations of cosmology @xcite and discuss some famous models of universe that were mentioned in the introduction
. this will prepare the ground for the next sections in which we discuss generalized cosmological models .
some readers may directly go to sec . [ sec_ges ] . in a space with uniform curvature ,
the line element is given by the friedmann - lematre - roberston - walker ( flrw ) metric @xmath89 where @xmath67 represents the radius of curvature of the @xmath90-dimensional space , or the scale factor . by an abuse of language
, we shall call it the `` radius of the universe '' . on the other hand
, @xmath87 determines the curvature of space .
the universe can be closed ( @xmath7 ) , flat ( @xmath39 ) , or open ( @xmath8 ) .
if the universe is isotropic and homogeneous at all points in conformity with the line element ( [ e1 ] ) , and contains a uniform perfect fluid of energy density @xmath91 and isotropic pressure @xmath92 , the energy - momentum tensor @xmath93 is @xmath94 the einstein equations @xmath95 relate the geometrical structure of the spacetime ( @xmath96 ) to the material content of the universe ( @xmath97 ) . for the sake of generality
, we have included the cosmological constant @xmath26
. given ( [ e1 ] ) and ( [ e2 ] ) , these equations reduce to @xmath98 @xmath99 where dots denote differentiation with respect to time .
these are the well - known cosmological equations first derived by friedmann @xcite .
the friedmann equations are usually written in the form @xmath100
@xmath101 @xmath102 where we have introduced the hubble parameter @xmath103 . among these three equations ,
only two are independent .
the first equation , which can be viewed as an `` equation of continuity '' , can be directly derived from the conservation of the energy momentum tensor @xmath104 which results from the bianchi identities . for a given barotropic equation of state @xmath105
, it determines the relation between the density and the scale factor .
then , the temporal evolution of the scale factor is given by eq .
( [ f3 ] ) .
equivalent expressions of the `` equation of continuity '' are @xmath106 @xmath107 @xmath108 introducing the volume @xmath109 and the energy @xmath110 , eq .
( [ f5 ] ) becomes @xmath111 .
this can be viewed as the first principle of thermodynamics for an adiabatic evolution of the universe @xmath112 .
this thermodynamical interpretation of the friedmann equation ( [ f1 ] ) was first given by lematre @xcite .
in most parts of this paper ( except in sec . [ sec_fm ] ) , we shall take @xmath43 .
this is because we are interested in the very early universe in which the pressure of radiation dominates over the dark energy ( cosmological constant ) . on the other hand
, we consider a flat universe ( @xmath39 ) in agreement with the observations of the cosmic microwave background ( cmb ) @xcite .
then , the friedmann equations reduce to @xmath113 @xmath114 @xmath115 the friedmann equations can also be derived in the context of newtonian cosmology . in this framework ,
milne @xcite and mccrea & milne @xcite first obtained the friedmann equations without pressure ( _ i.e. _ for @xmath116 ) .
later , mccrea @xcite introduced relativistic effects in newtonian cosmology and obtained the complete set of friedmann equations with pressure .
another derivation was given by harrisson @xcite who obtained a relativistic generalization of the euler - poisson system in the context of newtonian cosmology .
finally , the continuity equation was corrected in a recent paper by lima _
actually , the derivation of the friedmann equations with relativistic effects ( pressure ) from newtonian cosmology was first exposed by lematre @xcite as early as 1931 ( see also @xcite ) . however , for lematre , the newtonian derivation of the cosmological equations is just a pedagogical means and does not reflect the true ( relativistic ) nature of the problem . in modern textbooks of astrophysics @xcite , newtonian cosmology with pressure
is presented as a rigorously valid description of the universe in a region small compared to the hubble length @xmath117 .
_ remark : _ considering eq .
( [ f2 ] ) , we see that the cosmological constant corresponds to a force @xmath118 in the newtonian interpretation .
a positive cosmological constant ( de sitter ) acts as a force of cosmic repulsion and a negative cosmological constant ( anti de sitter ) acts as a force of cosmic attraction whose effect is similar to that of a harmonic oscillator .
on the other hand , a positive pressure has an attractive character while a negative pressure has a repulsive character .
gravity is attractive ( the universe decelerates ) when @xmath119 and repulsive ( the universe accelerates ) when @xmath120 . from the first principle of thermodynamics
@xmath121 one can derive the thermodynamical equation @xcite : @xmath122 for a given barotropic equation of state @xmath105 , this equation can be used to obtain the relation @xmath123 between the temperature and the density .
combining eq .
( [ t2 ] ) with eq .
( [ f6 ] ) , we get @xmath124 where @xmath125 is the entropy of the universe in a volume @xmath126 .
this confirms that the friedmann equations ( [ f1])-([f3 ] ) imply the conservation of the entropy @xcite . in 1917 , einstein @xcite constructed a static model of the universe .
he assumed that the universe if filled with a pressureless gas ( @xmath42 ) and introduced a cosmological constant @xmath26 in the equations of general relativity , @xmath31 and @xmath127 . ] .
setting @xmath128 in eqs .
( [ f1])-([f3 ] ) , it is necessary that @xmath129 and @xmath31 .
then , @xmath130 the static einstein universe is finite ( though unbounded ) with a positive curvature and a density fixed by @xmath26 and @xmath131 .
the volume of this universe is @xmath132 .
its mass @xmath133 is therefore @xmath134 the same year , de sitter @xcite found another cosmological solution of the einstein field equations corresponding to an empty universe ( @xmath135 ) with a cosmological constant @xmath129 . in his original article , he aimed at constructing a static solution .
he obtained a radius @xmath36 .
however , it was understood later by lematre @xcite that , by using a proper change of variables in the metric , the de sitter solution actually describes a dynamical universe whose radius increases exponentially rapidly with time . setting @xmath135 and @xmath39 in eqs .
( [ f1])-([f3 ] ) , we get @xmath136 in this model , the hubble parameter is constant : @xmath137 . it can be written @xmath138 where @xmath36 is the radius of the original ( static ) de sitter model .
the solution ( [ fm3 ] ) describes an empty universe ( @xmath35 ) that expands because of the repulsive nature of the cosmological constant .
any model with @xmath129 goes to a de sitter model for @xmath139 .
the lematre model of 1927 corresponds to @xmath31 , @xmath129 , @xmath42 and @xmath140 @xcite .
the equation of continuity ( [ f1 ] ) implies @xmath141 which can be interpreted as the conservation of mass - energy for a pressureless universe .
we can write this relation in the form @xmath142 , where @xmath143 is einstein s mass and @xmath86 is a dimensionless constant . introducing the dimensionless scale factor @xmath144 and the dimensionless time @xmath145 , we can rewrite the friedmann equation ( [ f3 ] ) in the form @xmath146 for @xmath4 , this equation admits the static einstein solution @xmath147 ( _ i.e. _ @xmath148 ) .
however , as shown by lematre @xcite , and emphasized by eddington @xcite , this solution is unstable and , for sufficiently large times , the radius of the universe expands .
this solution , according to which the radius of the universe is equal to the einstein radius for @xmath149 and increases for large times ( behaving ultimately like the de sitter solution ( [ fm3 ] ) for @xmath150 ) is called the eddington - lematre model ( see fig . [ lemaitre ] ) .
this model does not have an initial singularity . for @xmath4 ,
the differential equation ( [ fm4 ] ) has the analytical solution @xmath151 where @xmath152 is defined up to an additional constant . for @xmath149
, we obtain @xmath153 with @xmath154 . for @xmath150
, we get @xmath155 with @xmath156 . to the best of our knowledge
, these expressions have not been given previously . .
] in 1931 , lematre @xcite considered a model in which @xmath86 is slightly larger than one ( _ i.e. _ it contains more matter than the static einstein model ) .
this model starts from a singularity at @xmath20 .
the radius first increases as @xmath157 ( this corresponds to the solution ( [ dust1 ] ) found later by einstein & de sitter @xcite ) , then slows down and reaches a minimum rate at @xmath158 , after which it re - expands rapidly , ultimately approaching the de sitter solution ( [ fm3 ] ) for @xmath150 .
the main feature of this model is the existence of a `` stagnation period '' during which the radius remains close to @xmath158 . during the stagnation period
, the differential equation ( [ fm4 ] ) takes the approximate form @xmath159 whose solution is @xcite : @xmath160 where @xmath161 is the time at which @xmath162 reaches its minimum .
if @xmath86 is close to unity , then @xmath163 will remain close to the einstein static value @xmath164 for a long time @xmath165 . for @xmath4
, we recover the eddington - lematre model .
however , the historical interest of the lematre model @xcite of 1931 is that it predicts for the first time that the universe may have emerged from a _ singularity _ ( or near - singularity ) at @xmath20 in which the radius is equal to zero and the density is infinite .
this is what lematre called the `` primeval atom '' @xcite , and which became later the big bang theory . in early works of cosmology
, it was assumed that the universe is pressureless ( @xmath42 ) . in later works ,
pressure effects have been taken into account .
many workers have considered a linear equation of state @xmath166 with @xmath167 .
the pressure is positive for @xmath168 and negative for @xmath169 .
for this equation of state , the friedmann equations ( [ f7 ] ) leads to the relation @xmath170 in the following , we assume @xmath171 ( the case @xmath5 is treated in sec . [ sec_vacuum ] ) . substituting eq .
( [ leos5 ] ) in eq .
( [ f9 ] ) and solving the resulting equation for @xmath67 , we find that the radius of the universe increases with time as @xmath172 we have considered expanding solutions and defined the origin of time @xmath20 such that @xmath173 . the hubble parameter and the density decrease with time as @xmath174 in all these models ,
the universe is singular at @xmath20 .
the radius vanishes while the density and the hubble parameter are infinite ( big bang singularity ) . for the equation of state ( [ leos1 ] ) , the thermodynamical equation ( [ t2 ] ) can be integrated into @xmath175 the temperature decreases with time when @xmath168 and increases with time when @xmath169 .
according to eq .
( [ f8 ] ) , the universe is decelerating for @xmath176 and accelerating for @xmath177 . for @xmath178 ,
the scale factor increases linearly with time ( @xmath179 ) .
finally , the square of the velocity of sound is @xmath180 .
the velocity of sound is real for @xmath181 ( it is less than the speed of light for @xmath182 ) and imaginary for @xmath169 .
it vanishes for @xmath3 .
a pressureless universe ( @xmath42 ) made of baryonic matter or dark matter ( dust ) corresponds to @xmath3 .
this is the so - called einstein - de sitter ( eds ) universe . in this model , @xmath183 @xmath184 the equation of state of a radiation ( photon gas ) is @xmath45 , corresponding to @xmath2 .
this yields @xmath185 @xmath186 @xmath187 the relation between the energy density and the temperature corresponds to the stefan - boltzmann law @xmath188 introducing the planck scales , we can write the foregoing equations in the form @xmath189 @xmath190 @xmath191 a stiff equation of state @xmath192 , for which the velocity of sound @xmath193 is equal to the speed of light @xmath194 , corresponds to @xmath4 . in that case , @xmath195 @xmath196 @xmath197 for the equation of state @xmath56 , corresponding to @xmath5 , the density and the hubble parameter are constant @xmath198 and the scale factor increases exponentially rapidly as @xmath199 lematre @xcite was the first to understand that the effect of the cosmological constant @xmath26 is equivalent to that of a fluid with density @xmath200 [ see eq .
( [ f3 ] ) or ( [ fm3 ] ) ] described by an equation of state @xmath56 .
he interpreted @xmath201 as the vacuum energy .
the origin of the vacuum energy was later discussed by sakharov @xcite and zeldovich @xcite in relation to quantum field theory .
however , there is a fundamental difficulty in interpreting @xmath201 as the vacuum energy because quantum field theory predicts that the vacuum energy density should be of the order of @xmath58 which is @xmath202 orders of magnitude larger than @xmath201 .
this is known as the cosmological constant problem @xcite .
therefore , @xmath201 may not be related to the vacuum energy ( hence to quantum fluctuations ) but could be a fundamental property of general relativity ( see discussion in paper ii ) . in the modern terminology , the effect of the cosmological constant , or the effect of an exotic fluid , on the accelerated expansion of the universe is called dark energy @xcite .
the equation of state @xmath56 also appeared in the paper of mccrea @xcite in his discussion of the steady state theory of hoyle @xcite . in this theory ,
in which there is a continuous creation of matter , the density and the hubble parameter are constant , leading to eq .
( [ vacuum2 ] ) . according to hoyle @xcite ,
a continuous creation of matter can be achieved by a proper modification of the einstein equations .
mccrea @xcite noted that the same result could be obtained with an equation of state @xmath56 without invoking creation of matter and without having to modify the einstein equations .
the equation of state @xmath56 also appeared in inflationary models of the early universe @xcite .
it leads to a constant density which is identified with the planck density @xmath58 .
in that case , this equation of state may well be related to the vacuum energy density .
the previous models , based on a linear equation of state , lead to a finite time singularity at @xmath20 ( except the model with @xmath5 ) .
this poses several problems .
we briefly recall what the problems are and show in the next sections how a modification of the equation of state can resolve them .
we call @xmath203 the present value of the hubble constant ( the subscript @xmath204 refers to present - day values ) .
the hubble time is @xmath205 .
it gives an estimate of the age of the universe .
the characteristic `` size '' of the universe ( distance of the cosmological horizon ) is the hubble length @xmath206 . using @xmath207
, we find that @xmath208 and @xmath209 . for a flat space ( @xmath39 )
, the density of the universe @xmath210 can be obtained from the friedmann equation @xmath211 numerically , @xmath212 . on the other hand ,
the temperature measured from the cosmic microwave background ( cmb ) is @xmath213 .
the early universe is dominated by radiation . in the standard model
, the radiation phase leads to a finite time singularity at @xmath20 ( big bang ) .
it is expected that quantum mechanics will come into play at the planck time @xmath18 when the universe is very `` small '' . at that time , the density and the temperature are ( see sec .
[ sec_radiation ] ) : @xmath214 and @xmath215 the particle horizon at the planck time is of the order of the planck length @xmath216 . since @xmath217 is invariant during the evolution of the universe , we have @xmath218 if we take @xmath219 , we find that the present size of the universe is of order @xmath220 which is obviously too small .
the same argument can be turned the other way round . using the present size of the universe @xmath209 , we find that its size at the planck time is @xmath221 which is much larger than the planck length @xmath222 by a factor @xmath223 .
it is instructive to recover this result in a slightly different , but equivalent , manner . since @xmath224 is conserved during the evolution of the universe , we have @xmath225 . writing @xmath226 , and using eq .
( [ nec2 ] ) , we obtain @xmath227 since @xmath228 the foregoing equation can be rewritten @xmath229 since @xmath230 ( meaning that the hubble time @xmath205 is tremendously larger than the planck time @xmath18 ) , we see that @xmath231 the fact that @xmath232 makes problem @xcite
. one must find a mechanism to bring the `` size '' of the universe from @xmath13 to @xmath233 in a very short lapse of time .
this can be achieved with the inflationary scenario .
the usual inflationary scenario is based on the equation of state @xmath56 ( vacuum energy ) . as we have seen in sec .
[ sec_vacuum ] , this implies that @xmath57 and @xmath234 are constant so that the scale factor increases exponentially rapidly with time , as @xmath235 with @xmath236 .
if we assume that @xmath237 , we obtain @xmath238 so that the exponential time scale is of the order of the planck time . this solves the above - mentioned problem @xcite . in the following sections , we show that a polytropic equation of state with @xmath9 produces similar results and leads to simple cosmological models unifying the phases of inflation and radiation .
we consider a generalized equation of state of the form @xmath239 this equation of state was introduced in our previous paper @xcite for the specific index @xmath10 .
this is the sum of a standard linear equation of state @xmath1 and a polytropic equation of state @xmath240 , where @xmath87 is the polytropic constant and @xmath241 is the polytropic index . concerning the linear equation of state , we assume @xmath167 ( the case @xmath5 is treated specifically in appendix [ sec_eosgm ] ) .
this equation of state describes radiation ( @xmath2 ) , pressureless matter ( @xmath3 ) , and vacuum energy ( @xmath5 ) . concerning the polytropic equation of state
, we remain very general .
this polytropic equation of state may correspond to self - gravitating becs with repulsive ( @xmath7 ) or attractive ( @xmath8 ) self - interaction ( the standard bec model corresponds to @xmath10 ) @xcite , but it may have another origin . for an equation of state of the form ( [ ges0 ] ) with positive index @xmath9 , the polytropic component dominates the linear component when the density is high .
these models , studied in the present paper , describe the early universe .
conversely , when @xmath242 , the polytropic component dominates the linear component when the density is low .
these models , studied in paper ii , describe the late universe . as will be discussed in detail in paper ii , these two situations are strikingly similar .
they reveal a form of `` symmetry '' between the early and the late universe . in papers i and ii , we assume that @xmath243 .
this corresponds to the `` normal '' case where the density decreases with the radius ( see below ) .
the opposite case @xmath244 , leading to a `` phantom universe '' where the density increases with the radius , is considered in paper iii .
the purpose of our series of papers is to make an exhaustive study of all the possible cases @xmath245 , even if some of them seem to be unphysical or in conflict with the known properties of our universe . actually , by rejecting the cases leading to past or future singularities
, we will be led naturally to the `` good '' model ( see paper ii ) . for the equation of state ( [ ges0 ] ) ,
the friedmann equation ( [ f7 ] ) becomes @xmath246 assuming @xmath243 , this equation can be integrated into @xmath247 where @xmath248 and @xmath249 is a constant of integration .
the upper sign corresponds to @xmath7 ( repulsive self - interaction ) and the lower sign corresponds to @xmath8 ( attractive self - interaction ) .
we may draw a parallel between these two types of distributions and the statistics of fermions ( @xmath250 ) and bosons ( @xmath251 ) . and @xmath8 .
we have taken @xmath10 and @xmath2 . ] for @xmath7 , the density is defined only for @xmath252 .
when @xmath253 , @xmath254 when @xmath139 , @xmath255 corresponding to the linear equation of state . in the same limits , @xmath256 and @xmath257 , respectively . for @xmath8
, the density is defined for all @xmath163 .
when @xmath258 , the density tends to a finite value @xmath259 when @xmath139 , @xmath260 corresponding to the linear equation of state . in the same limits , @xmath261 and @xmath257 , respectively .
some curves giving the evolution of the density @xmath57 as a function of the scale factor @xmath163 are plotted in fig .
[ rhoa ] for @xmath7 and @xmath8
. for the equation of state ( [ ges0 ] ) , the thermodynamical equation ( [ t2 ] ) becomes @xmath262 this equation can be integrated into @xmath263 where @xmath264 is a constant of integration .
this relation between the temperature and the density can be viewed as a generalized stefan - boltzmann law .
combined with eq .
( [ ges3 ] ) , we obtain @xmath265 let us consider some asymptotic limits . when @xmath266 , eq .
( [ ges6 ] ) reduces to @xmath267 returning eq .
( [ leos10 ] ) for a linear equation of state .
as we have already indicated , in this regime , the temperature decreases to zero when @xmath168 and increases to @xmath49 when @xmath169 .
when @xmath3 , it tends to a constant value @xmath264 .
when @xmath7 and @xmath253 , @xmath268 when @xmath8 and @xmath258 , @xmath269 the extremum of temperature ( when it exists ) is located at @xmath270 @xmath271 @xmath272 some curves giving the evolution of the temperature @xmath273 as a function of the scale factor @xmath163 are plotted in figs .
[ atempkpos ] and [ atempkneg ] for @xmath7 and @xmath8 . .
we have taken @xmath10 , @xmath178 , @xmath3 and @xmath2 . ] .
we have taken @xmath10 , @xmath178 , @xmath3 and @xmath2 . ] finally , the entropy ( [ t4 ] ) is given by @xmath274 and we explicitly check that it is a constant
. we can rewrite the equation of state ( [ ges0 ] ) as @xmath276 with @xmath277 we define @xmath278 @xmath279 corresponding to a possible point where the pressure vanishes ( @xmath280 ) . , @xmath281 and @xmath282 as a function of the scale factor for @xmath7 .
we have taken @xmath10 and @xmath2 . ] , @xmath281 and @xmath282 as a function of the scale factor for @xmath8 .
we have taken @xmath10 and @xmath2 . ] for @xmath7 , @xmath283 when @xmath253 and @xmath284 when @xmath139 . for @xmath168 ,
the pressure is always positive ( @xmath285 ) . for @xmath169 ,
the pressure is positive when @xmath286 and negative when @xmath287 . for @xmath8 , @xmath288 when @xmath258 and @xmath284 when @xmath139 . for @xmath169 ,
the pressure is always negative ( @xmath289 ) . for @xmath168 ,
the pressure is negative when @xmath290 and positive when @xmath287 .
some curves giving the evolution of @xmath291 as a function of the scale factor @xmath163 are plotted in figs .
[ wqcpos ] and [ wqcneg ] for @xmath7 and @xmath8 . for the equation of state ( [ ges0 ] ) ,
the velocity of sound is given by @xmath292 the velocity of sound vanishes at the point defined by eqs .
( [ ges10a])-([ges11 ] ) where the temperature is extremum . at that point
the pressure is extremum with value @xmath293 the case @xmath294 corresponds to an imaginary velocity of sound .
we define @xmath295 @xmath296 @xmath297 corresponding to a possible point where the velocity of sound is equal to the speed of light .
different cases have to be considered .
we first assume @xmath7 .
when @xmath253 , @xmath298 ; when @xmath139 , @xmath299 . for @xmath168
, @xmath300 is always positive . for @xmath169
, @xmath300 is positive when @xmath301 and negative when @xmath302 .
the velocity of sound is larger than the speed of light when @xmath303 and smaller when @xmath304 .
we now assume @xmath8 . when @xmath258 , @xmath305 ; when @xmath139 , @xmath299 . for @xmath169 , @xmath300 is always negative . for @xmath168 ,
@xmath300 is negative when @xmath306 and positive when @xmath302 . for @xmath182 ,
the velocity of sound is always smaller than the speed of light .
some curves giving the evolution of @xmath282 as a function of the scale factor @xmath163 are plotted in figs .
[ wqcpos ] and [ wqcneg ] for @xmath7 and @xmath8 .
the deceleration parameter is defined by @xmath307 the universe is decelerating when @xmath308 and accelerating when @xmath309 . using the friedmann equations ( [ f8 ] ) and ( [ f9 ] ) , we obtain for a flat universe @xmath310 for the equation of state ( [ ges0 ] ) , using eq .
( [ gesw2 ] ) , we get @xmath311 we define a critical density , scale factor , and temperature @xmath312 @xmath313 corresponding to a possible inflexion point ( @xmath314 ) in the curve @xmath67 .
different cases have to be considered .
we first assume @xmath7 .
when @xmath253 , @xmath315 ; when @xmath139 , @xmath316 .
for @xmath176 , the universe is always decelerating ( @xmath308 ) . for @xmath177 ,
the universe is decelerating when @xmath317 and accelerating when @xmath318 .
we now assume @xmath8 . when @xmath258 , @xmath319 ; when @xmath139 , @xmath320 . for @xmath177 ,
the universe is always accelerating ( @xmath309 ) . for @xmath176 ,
the universe is accelerating when @xmath321 and decelerating when @xmath318 .
some curves giving the evolution of @xmath281 as a function of the scale factor @xmath163 are plotted in figs .
[ wqcpos ] and [ wqcneg ] for @xmath7 and @xmath8 . the temporal evolution of the scale factor @xmath67 is determined by the friedmann equation ( [ f9 ] ) . introducing the normalized radius @xmath322 ,
the density ( [ ges3 ] ) can be written @xmath323 substituting this expression in eq .
( [ f9 ] ) , we obtain the differential equation @xmath324 where @xmath325 and @xmath326 . in general
, we shall select the sign @xmath327 corresponding to an expanding universe ( @xmath328 ) . the solution can be written as @xmath329 or , after a change of variables @xmath330 , as @xmath331 the integral can be expressed in terms of hypergeometric functions .
some simple analytical expressions can be obtained for specific values of @xmath332 .
actually , we can have a good idea of the behavior of the solution of eq .
( [ dark2 ] ) by considering asymptotic limits ( see below ) .
the complete solution is represented in the figures by solving eq .
( [ dark2 ] ) numerically .
the universe starts from a primordial singularity at @xmath20 ( see below for a revision of this statement ) with a finite radius @xmath147 and an infinite density , pressure , and temperature .
when @xmath333 , @xmath334 @xmath335 @xmath336 then , the universe expands indefinitely .
when @xmath337 , the density decreases to zero and we recover the solution corresponding to a linear equation of state @xmath338 @xmath339 @xmath340 for @xmath176 , the universe is always decelerating ( @xmath341 ) . for @xmath177 ,
the universe is decelerating for @xmath342 and accelerating for @xmath343 .
the velocity of sound is less than the speed of light when @xmath344 . for @xmath10 , using the identity @xmath345 the general solution of eq .
( [ dark2 ] ) is @xmath346 for @xmath177 , the time @xmath347 at which the universe starts accelerating is given by @xmath348 the time @xmath349 at which the velocity of sound is equal to the speed of light is given by @xmath350 this gives the time at which the physical universe begins
. it may be meaningless to consider smaller times because the velocity of sound is greater than the speed of light .
for @xmath351 , the general solution of eq .
( [ dark2 ] ) is @xmath352 typical evolutions of the scale factor @xmath353 corresponding to @xmath177 , @xmath2 , and @xmath354 are represented in fig .
[ generalisationpos ] . for different values of @xmath86 ( specifically @xmath355 , @xmath178 , and @xmath2 ) .
we have taken @xmath10 . ]
the universe starts from @xmath41 with a vanishing radius @xmath356 , a vanishing temperature @xmath357 , and a finite density @xmath358 .
when @xmath41 , @xmath359 this corresponds to an exponential expansion ( early inflation ) due to the fact that the density is approximately constant .
the universe expands indefinitely .
when @xmath337 , the density decreases to zero and we recover the solution ( [ dark9])-([toto2 ] ) corresponding to a linear equation of state . for @xmath360 ,
the universe is always accelerating ( @xmath361 ) . for @xmath176
, the universe is accelerating when @xmath362 and decelerating when @xmath343 . in this model ,
the velocity of sound is always less than the speed of light . for @xmath10 , using the identity @xmath363 the general solution of eq .
( [ dark2 ] ) is @xmath364 where @xmath365 is a constant of integration . for @xmath176 ,
the time @xmath347 at which the universe starts decelerating is given by @xmath366 when @xmath41 , we obtain from eq .
( [ dark17 ] ) the asymptotic behavior @xmath367 with @xmath368 . for @xmath351 , the general solution of eq .
( [ dark2 ] ) is @xmath369 typical evolutions of @xmath353 corresponding to @xmath177 , @xmath2 , and @xmath354 are represented in fig .
[ generalisationneg ] . for different values of @xmath86 ( specifically @xmath355 , @xmath178 , and @xmath2 ) .
we have taken @xmath10 . ]
the generalized equation of state ( [ ges0 ] ) with @xmath9 and @xmath8 can be used to describe the transition between a phase of early inflation where the density of the universe is constant ( @xmath358 ) and a phase where the universe has a linear equation of state ( @xmath1 ) . therefore
, if we take @xmath2 , it describes the transition between the pre - radiation era and the radiation era .
this provides a model of non - singular inflationary universe . for simplicity
, we shall take @xmath10 ( in the bec model , this corresponds to the standard index @xcite ) . for @xmath2 , @xmath10 , and @xmath8 , eqs .
( [ ges0 ] ) , ( [ ges3 ] ) , ( [ ges5 ] ) , and ( [ ges6 ] ) become @xmath370 @xmath371 @xmath372 @xmath373 where @xmath374 , and we have noted @xmath375 for @xmath249 . when @xmath376 , the density has a constant value @xmath377 .
this gives rise to a phase of early inflation .
the value @xmath259 defines a fundamental upper bound @xmath378 for the density .
it is natural to identify it with the planck density @xmath379 ( see sec .
[ sec_nec ] ) .
this is how quantum mechanics is taken into account in our model .
since the planck density is obtained from a pure dimensional analysis , we should write @xmath380 , where @xmath381 is a constant of order unity ( actually , the planck constant should be defined so as to correspond exactly to @xmath378 ) . for simplicity
, we shall take @xmath382 this fixes the constant @xmath11 .
if the pre - radiation era is made of becs , this relation determines the ratio between the coupling constant @xmath383 of the bosons and their mass @xmath384 ( see appendix [ sec_annbec ] ) .
when @xmath385 , we can make the approximation @xmath386 .
this correspond to the radiation era described by an equation of state @xmath45 .
the conservation of @xmath387 implies that @xmath388 .
writing @xmath226 , where @xmath210 is the present density of the universe , and using eq .
( [ nec2 ] ) , we obtain @xmath389 using eq .
( [ nec8 ] ) , we get @xmath390 the value of the characteristic length @xmath375 is @xmath391 it provides a lengthscale that is intermediate between the planck length @xmath13 and the present size of the universe @xmath209 .
it corresponds to the typical size of the universe at the end of the inflationary phase ( or at the beginning of the radiation era ) . on the other hand , in the radiation era ,
( [ infb2 ] ) reduces to @xmath392 , so we can identify @xmath393 with the stefan - boltzmann constant @xmath394 . using eq .
( [ rad6 ] ) , this yields @xmath395 numerically , @xmath396 .
regrouping these results , we obtain the basic equations of the non - singular model @xmath397
@xmath398 @xmath399 @xmath400 @xmath401 the entropy ( [ ges12 ] ) is given by @xmath402 numerically , @xmath403 . finally , we note that @xmath404 .
as the universe expands from @xmath405 to @xmath406 , the density decreases from @xmath58 to @xmath204 , the equation of state parameter @xmath291 increases from @xmath407 to @xmath408 , the deceleration parameter @xmath281 increases from @xmath407 to @xmath409 , and the ratio @xmath282 goes from @xmath410 to @xmath408 .
the temperature starts at @xmath357 , reaches a maximum and decreases to zero . during the inflationary phase ( @xmath376 ) ,
the density has an approximately constant value @xmath411 .
the corresponding hubble parameter is @xmath412 numerically , @xmath413 and @xmath414 .
integrating eq .
( [ ip1 ] ) , we find that the scale factor increases exponentially rapidly with time as @xmath415 we have determined the constant of integration such that @xmath416 .
this universe exists at any time in the past , so there is no singularity . when @xmath417 , the radius of the universe is smaller than the planck length @xmath13 and it tends to zero exponentially rapidly when @xmath41 .
when @xmath418 , the radius of the universe increases exponentially rapidly on a timescale of the order of the planck time @xmath419 .
this phase of inflation connects the pre - radiation era to the radiation era on a very short interval of time ( of the order of a few planck times ) .
we call @xmath420 the time at which the scale factor @xmath67 reaches the value @xmath15 . using the approximate expression ( [ ip2 ] ) of the scale factor during the inflationary phase ,
we obtain @xmath421 numerically , @xmath422 .
this time corresponds typically to the end of the inflation .
therefore , during the inflation , the radius increases exponentially rapidly from @xmath68 at @xmath69 to @xmath423 at @xmath424 .
on the other hand , in the inflationary phase , the temperature is related to the density and to the scale factor by @xmath425 using the approximate expression ( [ ip2 ] ) , we find that the temperature increases exponentially rapidly with time as @xmath426 when @xmath417 , the temperature is less than @xmath427 , and it tends to zero exponentially rapidly when @xmath41 .
therefore , the pre - radiation era is extremely cold ( this is consistent with the bec model which assumes @xmath357 @xcite ) .
when @xmath418 , the temperature increases exponentially rapidly .
it passes from @xmath428 at @xmath69 to @xmath429 at @xmath424 .
these values are based on the asymptotic results ( [ ip2 ] ) and ( [ ip6 ] ) .
they will be slightly revised in sec .
[ sec_gp ] using the exact solution ( [ gp1 ] ) of the friedmann equations .
after the inflation ( @xmath385 ) , the universe enters in the radiation era . in that case , the quadratic pressure ( @xmath430 ) is negligible as compared to the pressure of radiation ( @xmath431 ) and we recover the standard radiation model of sec .
[ sec_radiation ] .
the density is related to the scale factor by @xmath432 .
the friedmann equation ( [ f9 ] ) becomes @xmath433 yielding @xmath434 using the approximate expression ( [ rp2 ] ) of the scale factor during the radiation era , we find that the time at which @xmath21 is @xmath435 numerically , @xmath436 . this marks the beginning of the radiation era . again , this numerical value will be revised in sec . [ sec_gp ] .
we also have @xmath437 and @xmath438 during the radiation era , the density and the temperature decrease algebraically as the universe expands .
the equation of state ( [ infb8a ] ) interpolates smoothly between the pre - radiation era described by a constant density @xmath439 ( vacuum energy ) and the radiation era described by a density @xmath440 .
it provides therefore a unified description of the early universe . using the results of sec .
[ sec_darkatt ] , the general solution of the friedmann equation ( [ f9 ] ) is @xcite : @xmath441 where @xmath365 is a constant of integration .
it is determined such that @xmath442 at @xmath20 . setting @xmath443 we get @xmath444 numerically , @xmath445 and @xmath446 . for @xmath41
, we have the exact asymptotic result @xmath447 with @xmath448 . due to the smallness of @xmath449 , a very good approximation of @xmath365 and @xmath450 is given by @xmath451 and @xmath452 . with this approximation , eq .
( [ gp3b ] ) returns eq .
( [ ip2 ] ) . with the time @xmath66 in logarithmic scales .
this figure clearly shows the phase of inflation connecting the pre - radiation era to the radiation era . during the inflation , the scale factor increases by @xmath453 orders of magnitude in less than @xmath454 . ] with the time @xmath66 in linear scales .
the dashed line corresponds to the standard radiation model of sec .
[ sec_radiation ] exhibiting a finite time singularity at @xmath20 ( big bang ) . when quantum mechanics is taken into account ( as in our simple model ) ,
the initial singularity is smoothed - out and the scale factor at @xmath20 is equal to the planck length @xmath13 .
this is similar to a second order phase transition where the planck constant plays the role of finite size effects ( see sec .
[ sec_analogy ] for a development of this analogy ) . ] with the time @xmath66 in logarithmic scales . during the inflation ,
the density remains approximately constant with the planck value @xmath455 which represents an upper bound . ] with the time @xmath66 in linear scales .
the dashed line corresponds to the standard radiation model of sec .
[ sec_radiation ] exhibiting a finite time singularity at @xmath20 .
quantum mechanics limits the rise of the density to the planck value @xmath379 . ] with the time @xmath66 in logarithmic scales . for @xmath456 ,
the universe is extremely cold ( @xmath457 ) . during the inflation , the temperature increases by @xmath458 orders of magnitude in less than @xmath454 . during the radiation era
, the temperature decreases algebraically . ] with the time @xmath66 in linear scales .
the dashed line corresponds to the standard radiation model of sec .
[ sec_radiation ] exhibiting a finite time singularity at @xmath20 .
quantum mechanics limits the rise of the temperature to about half the planck value @xmath459 . in the pre - radiation era
, the temperature drops to zero exponentially rapidly . ] at @xmath69 , @xmath460 @xmath461 numerically , @xmath68 , @xmath462 , and @xmath427 . using the results of sec .
[ sec_dec ] , we find that the universe is accelerating for @xmath321 ( _ i.e. _ @xmath463 ) and decelerating for @xmath318 ( _ i.e. _ @xmath464 ) where @xmath465 the corresponding temperature is @xmath466 the time @xmath347 at which the universe starts decelerating is given by @xmath467 this is the time at which the curve @xmath67 presents an inflexion point .
it turns out that this inflexion point coincides with @xmath375 ( this property is true only for @xmath2 ) .
therefore , it coincides with our definition of the end of the inflation : @xmath468 .
numerically , @xmath469 , @xmath470 , @xmath471 , and @xmath472 .
the temperature starts from @xmath357 at @xmath41 , increases exponentially rapidly during the inflation , reaches a maximum value , and decreases algebraically during the radiation era . using the results of sec .
[ sec_gestemperature ] , we find that the point corresponding to the maximum temperature is @xmath473 @xmath474 it is achieved at a time @xmath475 numerically , @xmath476 , @xmath477 , @xmath478 , and @xmath479 . the evolution of the scale factor , density and temperature as a function of time are represented in figs .
[ taloglog]-[ttemplinlin ] in logarithmic and linear scales . in this model ,
the universe starts at @xmath41 with a vanishing radius @xmath405 , a finite density @xmath480 , and a vanishing temperature @xmath357 .
the universe exists at any time in the past and does not present any singularity .
it has been called `` aioniotic '' universe in @xcite since it has no origin nor end . to make contact with the big bang theory
, we define the `` original '' time @xmath69 as the time at which the radius of the universe is equal to the planck length .
thus @xmath68 .
the corresponding density and temperature are @xmath462 and @xmath427 .
we note that quantum mechanics regularizes the finite time singularity that arises in the standard big bang theory .
this is similar to finite size effects in second order phase transitions ( see sec .
[ sec_analogy ] ) .
we also note that the universe is very cold at @xmath69 , unlike what is predicted by the big bang theory ( a naive extrapolation of the law @xmath48 leads to @xmath481 ) .
the universe first undergoes a phase of inflation during which its radius and temperature increase exponentially rapidly [ see eqs .
( [ ip2 ] ) and ( [ ip6 ] ) ] while its density remains approximately constant [ see eq .
( [ infb4 ] ) ] .
the inflation `` starts '' at @xmath482 and ends at @xmath483 . during this very short lapse of time
, the radius of the universe grows from @xmath68 to @xmath15 , and the temperature grows from @xmath427 to @xmath484 .
by contrast , the density does not change significatively : it goes from @xmath462 to @xmath485 . after the inflation
, the universe enters in the radiation era and , from that point , we recover the standard model . the radius increases
algebraically [ see eq .
( [ rp2 ] ) ] while the density and the temperature decrease algebraically [ see eqs .
( [ rp4 ] ) and ( [ rp5 ] ) ] .
the temperature achieves its maximum value @xmath486 at @xmath487 . at that moment
, the density of the universe is @xmath488 and its radius @xmath489 . during the inflation ,
the universe is accelerating and during the radiation era it is decelerating . the transition takes place at a time @xmath472 coinciding with the end of the inflation ( _ i.e. _ @xmath490 ) .
we note that the inflationary process described above is relatively different from the usual inflationary scenario @xcite . in standard inflation ,
the universe is radiation dominated up to @xmath491 , but expands exponentially by a factor @xmath492 in the interval @xmath493 with @xmath494 .
for @xmath495 the evolution is again radiation dominated . at @xmath496
the temperature is about @xmath497 ( this corresponds to the epoch at which most `` grand unified theories '' have a significant influence on the evolution of the universe ) . during the exponential inflation ,
the temperature drops drastically ; however , the matter is expected to be reheated to the initial temperature of @xmath497 by various high energy processes @xcite . in the inflationary process described previously , the evolution of the temperature is different . _
remark : _ of course , our definition of the `` original '' time @xmath20 , corresponding to @xmath498 , is relatively arbitrary .
what essentially matters is the evolution of the different variables on a tiny interval of time of the order of @xmath499 .
the equation of state ( [ ges0 ] ) with @xmath9 and @xmath7 leads to a model of non - inflationary universe with a new type of initial singularity .
we take the same values of @xmath86 , @xmath332 , @xmath259 , @xmath375 and @xmath264 as in the previous section .
we also take @xmath19 .
the basic equations of the singular model are @xmath500 @xmath501 @xmath502 @xmath503 @xmath504 the entropy is still given by eq .
( [ infb12 ] ) .
as the universe expands from @xmath21 to @xmath49 , the density and the temperature decreases from @xmath49 to @xmath204 , the parameter @xmath291 decreases from @xmath49 to @xmath408 , the deceleration parameter @xmath281 decreases from @xmath49 to @xmath409 , and the ratio @xmath282 decreases from @xmath49 to @xmath408 . with the time @xmath66 in logarithmic scales .
there is no inflation since the universe already starts with a `` large '' radius @xmath15 .
however , the initial condition is singular since the density and the temperature are infinite . ] with the time @xmath66 in linear scales .
the dashed line corresponds to the standard radiation model of sec .
[ sec_radiation ] ( big bang ) . ] with the time @xmath66 in logarithmic scales .
the slope goes from @xmath505 in the pre - radiation era to @xmath506 in the radiation era . ] with the time @xmath66 in linear scales .
the dashed line corresponds to the standard radiation model of sec .
[ sec_radiation ] . ] with the time @xmath66 in logarithmic scales .
the slope goes from @xmath507 in the pre - radiation era to @xmath508 in the radiation era . ] with the time @xmath66 in linear scales .
the dashed line corresponds to the standard radiation model of sec .
[ sec_radiation ] . ] for the equation of state ( [ ni1a ] ) , using the results of sec .
[ sec_darkrep ] , the general solution of the friedmann equation ( [ f9 ] ) is @xcite : @xmath509 the universe starts from a singularity at @xmath20 .
the initial radius of the universe is finite ( with the value @xmath80 ) but the initial density and the initial temperature are infinite .
the universe is always decelerating . the density and the temperature decrease indefinitely as the universe expands .
there is no inflation since the universe starts with a `` large '' radius @xmath510 .
when @xmath333 , @xmath511 @xmath512 @xmath513 when @xmath514 , the universe enters in the radiation era .
in that case , we recover the results of sec .
[ sec_rp ] .
the condition that the velocity of sound must be less than the speed of light imposes a maximum bound on the density . using the results of secs .
[ sec_gessound ] and [ sec_darkrep ] , we find that the physical universe starts at a time @xmath515 at this time , its radius , density and temperature are @xmath516 @xmath517 for @xmath518 , the solution ( [ ni4 ] ) may be unphysical because the velocity of sound is larger than the speed of light .
numerically , @xmath519 , @xmath520 , @xmath521 and @xmath522 . the evolution of the scale factor , density and temperature as a function of time are represented in figs .
[ taloglogpos]-[ttemplinlinpos ] in logarithmic and linear scales .
the standard big bang theory is a classical theory in which quantum effects are neglected . in that case
, it exhibits a finite time singularity : the radius is equal to zero at @xmath20 while the density and the temperature are infinite . for @xmath12 ,
the solution is not defined and we may take @xmath405 . for @xmath523
the radius increases as @xmath46 .
this is similar to a second order phase transition if we view the time @xmath66 as the control parameter ( _ e.g. _ the temperature @xmath273 ) and the scale factor @xmath163 as the order parameter ( _ e.g. _ the magnetization @xmath524 ) .
it is amusing to note that the exponent @xmath525 is the same as in mean field theories of second order phase transitions ( _ i.e. _ @xmath526 ) but this is essentially a coincidence . when quantum mechanics effects are taken into account , as in the simple model of sec .
[ sec_inf ] , the singularity at @xmath20 disappears and the curves @xmath67 , @xmath527 and @xmath528 are regularized .
in particular , we find that @xmath529 at @xmath20 , instead of @xmath405 , due to the finite value of @xmath24 .
this is similar to the regularization due to finite size effects ( _ e.g. _ the system size @xmath530 or the number of particles @xmath531 ) in ordinary phase transitions . in this sense , the classical regime @xmath532 is similar to the thermodynamic limit ( @xmath533 or @xmath534 ) in ordinary phase transitions . , dashed line ) in the model of sec .
[ sec_inf ] .
the singularity at @xmath20 is replaced by an inflationary expansion from the pre - radiation era to the radiation era .
we can draw an analogy with second order phase transitions where the planck constant plays the role of finite size effects ( see the text for details ) . ] in the model of sec .
[ sec_ni ] . the standard big bang solution ( @xmath25 )
is represented by a dashed line . for @xmath535 ,
the initial radius is strictly positive and it tends to zero as @xmath532 . ] to study the convergence toward the classical big bang solution when @xmath532 , it is convenient to use a proper normalization of the parameters .
we call @xmath536 the true value of the planck constant ( in this section , the subscript @xmath204 refers to true values of the parameters ) .
then , we express time in terms of @xmath537 and lengths in terms of @xmath538 .
we introduce @xmath539 and @xmath540 . using @xmath541 and eq .
( [ infb5 ] ) , we obtain @xmath542 where @xmath543 . therefore @xmath544 where @xmath545 . using these relations and eq .
( [ gp1 ] ) , we obtain @xmath546 where @xmath547 is defined by eq .
( [ gp3 ] ) .
this equation describes a phase transition between the pre - radiation era and the radiation era .
the normalized planck constant @xmath548 plays the role of finite size effects .
finite size scalings are explicit in eq .
( [ ana3 ] ) . for @xmath549
, we recover the non - singular model of sec .
[ sec_inf ] . for @xmath550
, we recover the singular radiation model of sec .
[ sec_radiation ] ( big bang ) . the convergence towards this singular solution as @xmath551 is shown in fig . [ phasetransition ] .
we can make the same study for the model of sec .
[ sec_ni ] although the analogy with phase transitions with finite size effects is less clear since this model remains singular for finite values of the planck constant .
nevertheless , we have represented the effect of quantum mechanics on the form of the curve @xmath67 in fig .
[ phasetransitionpos ] . for finite values of @xmath24 , using eq .
( [ ni4 ] ) , the evolution of the scale factor is given by @xmath552 at @xmath20 , the scale factor is @xmath553 .
in this paper , we have carried out an exhaustive study of the generalized equation of state ( [ ges0 ] ) for positive indices @xmath9 . for @xmath2 , @xmath10 and @xmath11 ,
this equation of state describes in a unified manner the transition between the pre - radiation era ( @xmath439 ) and the radiation era ( @xmath440 ) in the early universe .
it provides a model of early inflation .
the case of negative indices @xmath242 is treated in paper ii . for @xmath3 , @xmath554 and @xmath555 , where @xmath201 is the cosmological density , this equation of state describes in a unified manner the transition between the matter era ( @xmath556 ) and the dark energy era ( @xmath557 ) in the late universe .
it provides a model of late inflation . combining these two approaches ,
we obtain a simple non - singular model for the whole evolution of the universe based on two symmetric polytropic equations of state ( see paper ii ) .
when we started our study , we were mainly interested in the equation of state @xmath558 introduced in our previous paper @xcite .
this equation of state was obtained in the context of bose - einstein condensates , assuming that they are the constituents of dark matter .
although becs should form after the radiation era when the universe has cooled sufficiently , we wondered whether this equation of state could have some sense before the radiation era . in that picture , the pre - radiation era would correspond to primordial becs with attractive self - interaction @xmath559 ( see appendix [ sec_annbec ] ) .
since the pre - radiation era is extremely cold ( see sec .
[ sec_inf ] ) , the bec model ( which assumes @xmath357 ) may be an interesting suggestion to develop further . however , we progressively realized that the study of a generalized equation of state of the form @xmath0 was of interest beyond the context of becs . therefore , our present point of view is more general .
we believe that the equation of state ( [ ges0 ] ) may have several origins that should be discussed in future works . by studying this equation of state in full generality ,
we realized that the positive indices @xmath9 describe the early universe while the negative indices @xmath242 describe the late universe .
furthermore , a positive polytropic pressure ( @xmath7 ) leads to past or future singularities ( or peculiarities ) while a negative polytropic pressure ( @xmath8 ) leads to non - singular models .
they exhibit phases of early and late inflation associated with a maximum density @xmath560 ( planck density ) corresponding to the vacuum energy in the past and a minimum density @xmath561 ( cosmological density ) corresponding to the dark energy in the future .
therefore , in the polytropic model , the description of the early and late universe appears to be very `` symmetric '' .
this result is obtained in a purely theoretical manner , without reference to observations .
strikingly , this symmetry is consistent with what we know about the real universe .
this point will be further developed in paper ii .
another motivation of our study was to discuss some aspects of the history of science and publicize ( in the same vein as @xcite ) the important contributions of lematre in cosmology , which may not be sufficiently recognized .
they include : 1 .
the first correct understanding of the mysterious de sitter model @xcite ; 2 .
the re - discovery of the friedmann equations and their thermodynamical interpretation @xcite ; 3 .
the instability of the einstein static universe @xcite ; 4 .
the interpretation of the redshifts as a results of the expansion of the universe in accord with the einstein theory of relativity @xcite ; 5 .
the discovery of the hubble law and the calculation of the hubble constant two years before hubble @xcite ; 6 .
the basis of newtonian cosmology @xcite ; 7 .
the correct place of the cosmological constant in the poisson equation @xcite ; 8 .
the importance of the cosmological constant and its interpretation in terms of a vacuum energy density @xcite ; 9 .
the notion of a primordial singularity ( primeval atom ) that finally led to the big bang theory @xcite ; 10 .
the importance of quantum mechanics in the early universe @xcite .
some other important contributions of lematre have been collected recently in @xcite .
in this appendix , we specifically study the equation of state ( [ ges0 ] ) with @xmath5 , @xmath9 , and @xmath7 , namely @xmath563 it generalizes the equation of state @xmath56 of the vacuum .
for the equation of state ( [ eosgm0 ] ) , the continuity equation ( [ f7 ] ) can be integrated into @xmath564 where @xmath565 and @xmath249 is a constant of integration .
the density is defined for @xmath566 . when @xmath253 , @xmath567 and @xmath256 ; when @xmath139 , @xmath568 and @xmath257
. the thermodynamical equation ( [ t2 ] ) can be integrated into @xmath569 where @xmath264 is a constant of integration . combined with eq .
( [ eosgm1 ] ) , we obtain @xmath570 when @xmath253 and @xmath139 , @xmath571 .
the temperature has a minimum at @xmath572 @xmath573 the evolution of the density and temperature as a function of the scale factor is plotted in fig .
[ appendixarhotemp ] . .
we have taken @xmath10 . ] , @xmath281 , and @xmath282 as a function of the scale factor @xmath163 .
we have taken @xmath10 . ]
the equation of state can be written as @xmath574 with @xmath575 the pressure vanishes ( @xmath280 ) at @xmath576 when @xmath253 , @xmath283 ; when @xmath139 , @xmath288 .
the pressure is positive for @xmath286 and negative for @xmath287 .
the deceleration parameter is given by eqs .
( [ dec1 ] ) and ( [ dec2 ] ) . together with eq .
( [ eosgm5 ] ) , we obtain @xmath577 the curve @xmath67 presents an inflexion point ( @xmath578 ) at @xmath579 when @xmath253 , @xmath315 ; when @xmath139 , @xmath319 .
the universe is decelerating when @xmath317 and accelerating when @xmath318 .
finally , the velocity of sound is given by @xmath580 the velocity of sound vanishes at the point defined by eqs .
( [ eosgm4a])-([eosgm4b ] ) at which the temperature reaches its minimum . at that point ,
the pressure is minimum with value @xmath581 when @xmath253 , @xmath298 ; when @xmath139 , @xmath582 .
the velocity of sound is real when @xmath301 and imaginary when @xmath302 . on the other hand ,
the velocity of sound is equal to the speed of light at @xmath583 @xmath584 for @xmath10 , this corresponds to the point at which the pressure vanishes ( @xmath280 ) .
the velocity of sound is larger than the speed of light when @xmath303 and smaller when @xmath304 .
the evolution of @xmath291 , @xmath281 , and @xmath282 as a function of the scale factor @xmath163 is represented in fig .
[ appendixwqc ] .
setting @xmath322 , the friedmann equation ( [ f9 ] ) can be written @xmath585 where @xmath586 .
its solution is @xmath587 where we have determined the constant of integration such that @xmath147 at @xmath20 .
the density and the temperature evolve as @xmath588 @xmath589 the universe starts at @xmath20 with a radius @xmath147 and an infinite density and infinite pressure .
then , the universe expands indefinitely while the density decreases to zero .
the physical solution , for which the velocity of sound is smaller than the speed of light , starts at @xmath496 where @xmath590 . for @xmath10 ,
this corresponds to the time at which the pressure becomes negative .
the expansion of the universe is decelerating when @xmath591 and accelerating when @xmath592 , where @xmath593 .
the velocity of sound becomes imaginary when @xmath594 , where @xmath595 .
this corresponds to the time at which the temperature reaches its minimum .
the evolution of the scale factor with time is plotted in fig . [ appendixta ] . . for this index , @xmath596 , @xmath597 , and @xmath598 . ]
at @xmath357 , and in the strong coupling limit , the equation of state of bose - einstein condensates ( becs ) with short - range interactions is @xmath599 where @xmath384 is the mass of the bosons and @xmath61 is the s - scattering length @xcite .
therefore , a self - interacting bec has a polytropic equation of state @xmath600 with index @xmath601 ( _ i.e. _ @xmath10 ) and polytropic constant @xmath602 .
furthermore , the scattering length @xmath61 can be positive or negative ( positive scattering corresponds to repulsive self - interaction , and negative scattering to attractive self - interaction ) . in terrestrial bec experiments ,
some atoms like @xmath603 have a negative scattering length @xcite . as a result
, the pressure of a bec can be _
negative_. therefore , there exist physical systems in which it is possible to obtain polytropic equations of state with negative pressure .
furthermore , the origin of these equations of state is due to quantum mechanics since the planck constant explicitly appears in eq .
( [ annbec1 ] ) .
this may be a justification ( but not the only one ) for the generalized equation of state ( [ ges0 ] ) .
it is convenient to define a coupling constant @xmath604 by @xcite : @xmath605 we can now use @xmath606 as independent variables instead of @xmath607 .
the equation of state ( [ annbec1 ] ) can be rewritten as @xmath608 in the inflationary model of sec .
[ sec_inf ] , in which the pre - radiation era is described by a `` fluid '' with a polytropic equation of state of the form ( [ annbec3]-a ) with @xmath8 , the maximum density is @xmath374 . if we assume that this `` fluid '' is made of becs , then , using eq .
( [ annbec3]-b ) , we obtain @xmath609 assuming that @xmath610 is of the order of the planck density @xmath58 , _
i.e. _ @xmath611 , we find that the coupling constant of the bec scales as @xmath612 . of course , the nature of these `` primordial '' becs remains highly speculative .
it is possible that this bec description is just an effective one . | we construct models of universe with a generalized equation of state @xmath0 having a linear component and a polytropic component .
the linear equation of state @xmath1 describes radiation ( @xmath2 ) , pressureless matter ( @xmath3 ) , stiff matter ( @xmath4 ) , and vacuum energy ( @xmath5 ) .
the polytropic equation of state @xmath6 may be due to bose - einstein condensates with repulsive ( @xmath7 ) or attractive ( @xmath8 ) self - interaction , or have another origin . in this paper
, we consider positive indices @xmath9 . in that case , the polytropic component dominates in the early universe where the density is high . for @xmath2 , @xmath10 and @xmath11 , we obtain a model of early universe describing the transition from a pre - radiation era to the radiation era .
the universe exists at any time in the past and there is no singularity .
however , for @xmath12 , its size is less than the planck length @xmath13 . in this model
, the universe undergoes an inflationary expansion with the planck density @xmath14 that brings it to a size @xmath15 at @xmath16 ( about @xmath17 planck times @xmath18 ) . for @xmath2 , @xmath10 and @xmath19
, we obtain a model of early universe with a new form of primordial singularity : the universe starts at @xmath20 with an infinite density and a finite radius @xmath21 .
actually , this universe becomes physical at a time @xmath22 from which the velocity of sound is less than the speed of light .
when @xmath23 , the universe evolves like in the standard model .
we describe the transition from the pre - radiation era to the radiation era by analogy with a second order phase transition where the planck constant @xmath24 plays the role of finite size effects ( the standard big bang theory is recovered for @xmath25 ) . |
it remains a theoretical challenge to explain the observed pattern of quark and lepton masses and mixings , in particular the striking differences between the quark sector and the neutrino sector .
promising elements of a theory of flavour are grand unification ( gut ) based on the groups @xmath5 , @xmath6 or @xmath7 , supersymmetry , the seesaw mechanism and additional flavour symmetries @xcite .
a successful example is the froggatt - nielsen mechanism @xcite based on spontaneously broken abelian symmetries , which parametrizes quark and lepton mass ratios and mixings by powers of a small ` hierarchy parameter ' @xmath8 .
the resulting structure of mass matrices also arises in compactifications of higher - dimensional field and string theories , where the parameter @xmath8 is related to the location of matter fields in the compact dimensions or to vacuum expectation values of moduli fields ( cf .
@xcite ) . in this article
we consider a froggatt - nielsen symmetry which commutes with the gut group @xmath5 , and which naturally explains the large @xmath9 mixing @xcite .
this symmetry implies a particular hierarchy pattern in the majorana mass matrix for the light neutrinos , @xmath10 which can be regarded as a key element for our analysis .
the predicted dirac and majorana neutrino mass matrices are also consistent with leptogenesis @xcite . despite these successes ,
the predictive power of the froggatt - nielsen mechanism is rather limited due to unknown @xmath0 coefficients in all entries of the mass matrices .
for example , the considered model @xcite can accommodate both a small as well as a large ` solar ' mixing angle @xmath11 @xcite .
to get an idea of the range of possible predictions for a given flavour structure , it is instructive to treat the @xmath0 parameters as random variables @xcite . in the following we shall employ monte - carlo techniques to study quantitatively the dependence of yet undetermined , but soon testable parameters of the neutrino sector on the unknown @xmath12 factors of the mass matrices . using the already measured neutrino masses and mixings as input ,
we find surprisingly sharp predictions which indicate a large value for the smallest mixing angle @xmath4 in accordance with recent results from t2k @xcite , minos @xcite and double chooz @xcite , a value for the lightest neutrino mass of @xmath13 ev and one majorana phase in the mixing matrix peaked at @xmath14 .
as far as orders of magnitude are concerned , the masses of quarks and charged leptons approximately satisfy the relations @xmath15 with @xmath16 for masses defined at the gut scale .
this mass hierarchy can be reproduced by a simple @xmath17 flavour symmetry .
grouping the standard model leptons and quarks into the @xmath5 multiplets @xmath18 and @xmath19 , the yukawa interactions take the form @xmath20 where @xmath21 denote the charge conjugates of right - handed neutrinos and @xmath22 are flavour indices .
note that the yukawa matrix @xmath23 for the right - handed neutrinos can always be chosen to be real and diagonal .
@xmath24 , @xmath25 and @xmath26 are the higgs fields for electroweak and @xmath27 symmetry breaking , i.e. , their vacuum expectation values generate the dirac masses of quarks and leptons and the majorana masses for the right - handed neutrinos , respectively . in this setup ,
the yukawa couplings are determined up to complex @xmath12 factors by assigning @xmath17 charges to the fermion and higgs fields in eq . ,
@xmath28 with the charge assignment given in tab .
[ tab_fn - charges ] the mass relations in eq . are reproduced .
additionally , perturbativity of the yukawa couplings and constraints on @xmath29 require @xmath30 .
.froggatt - nielsen charge assignments . from ref .
[ cols="^,^,^,^,^,^,^,^,^,^,^,^,^",options="header " , ] _ masses _ + from eq . and tab .
[ tab_fn - charges ] one obtains for the dirac neutrino mass matrix @xmath31 and the majorana mass matrix of the right - handed neutrinos @xmath32 , @xmath33 with the electroweak and @xmath27 symmetry breaking vacuum expectation values @xmath34 and @xmath35 , respectively . in the seesaw formula
@xmath36 the dependence on the right - handed neutrino charges drops out , and one finds for the light neutrino mass matrix , @xmath37 the charged lepton mass matrix is given by @xmath38 note that the second and third row of the matrix @xmath39 have the same hierarchy pattern .
this is a consequence of the same flavour charge for the second and third generation of leptons , which is the origin of the large neutrino mixing .
hence , diagonalizing @xmath39 can a priori give a sizable contribution to the mixing in the lepton sector .
_ mixing _ + the lepton mass matrices are diagonalized by bi - unitary and unitary transformations , respectively , @xmath40 with @xmath41 . from @xmath42 and @xmath43 one obtains the leptonic mixing matrix @xmath44 , which is parametrized as @xcite @xmath45 with @xmath46 and @xmath47 . since the light neutrinos are majorana fermions , all three phases are physical . in the following
we study the impact of the unspecified @xmath12 factors in the lepton mass matrices on the various parameters of the neutrino sector by using a monte carlo method , taking present knowledge on neutrino masses and mixings into account .
naively , one might expect large uncertainties in the predictions for the observables of the neutrino sector obtained in this setup .
for instance , the neutrino mass matrix is calculated by multiplying three matrices , in which each entry comes with an unspecified @xmath12 factor , cf .
. however , carrying out the analysis described below and calculating the @xmath48 confidence intervals , we find that in many cases our results are sharply peaked , yielding a higher precision than only an order - of - magnitude estimate .
_ monte - carlo study _
the unknown @xmath12 coefficients of the yukawa matrices @xmath49 , @xmath50 and @xmath23 are constrained by the experimental data on neutrino masses and mixings , with the @xmath51 confidence ranges given by @xcite : @xmath52 in the following we explicitly do not use the current bound on the smallest mixing angle ( @xmath53 at @xmath51 @xcite ) .
this allows us to demonstrate that nearly all values we obtain for @xmath4 automatically obey the experimental bound , cf .
[ fig_mixing ] . in a numerical monte - carlo study
we generate random numbers to model the 39 real parameters of the three mass matrices .
factors in each @xmath50 and @xmath49 , as well as three real @xmath12 factors in @xmath23 .
note that here we are treating the low energy yukawa couplings as random variables , which are related to the couplings at higher energy scales via renormalization group equations .
however , we expect that the effect of this renormalization group running can essentially be absorbed into a redefinition of the effective scale @xmath54 , hence leaving the results presented in the following unchanged . ]
the absolute values are taken to be uniformly distributed in @xmath55 $ ] on a logarithmic scale .
the phases in @xmath49 and @xmath50 are chosen to be uniformly distributed in @xmath56 . in the following
, we shall refer to those sets of coefficients which are consistent with the experimental constraints in eq . as hits . in a preliminary run
, we consider the neutrino mixing matrix @xmath43 , with the effective scale @xmath57 treated as random variable in the interval @xmath58\times10^{15}$ ] gev .
we find that the percentage of hits strongly peaks at @xmath59 gev .
this is interesting for two reasons .
firstly , it implies that given @xmath30 , the high seesaw scale lies in the range @xmath60 .
note that the upper part of this mass range is close the gut scale , which is important for recent work on the connection of leptogenesis , gravitino dark matter and hybrid inflation @xcite .
secondly , this result allows us to fix the parameter @xmath54 in the following computations without introducing a significant bias . in the main run , for fixed @xmath54 , we include the mixing matrix @xmath42 of the charged leptons to compute the full pmns matrix .
we require the mass ratios of the charged leptons to fulfill the experimental constraints up to an accuracy of @xmath61 and allow for @xmath62 to achieve the correct normalization of the charged lepton mass spectrum . finally , imposing the @xmath63 constraints on the two large mixing angles of the full pmns matrix
, we find parameter sets of @xmath12 factors which yield mass matrices fulfilling the constraints in eq . .
our final results are based on roughly 20000 such hits .
for each hit we calculate the observables in the neutrino sector as well as parameters relevant for leptogenesis .
the resulting distributions are discussed below .
_ statistical analysis _ in our theoretical setup the relative frequency with which we encounter a certain value for an observable might indicate the probability that this value is actually realized within the large class of flavour models under study . in the following
we shall therefore treat the distributions for the various observables as probability densities for continuous random variables .
that is , our predictions for the respective observables represent best - guess estimates according to a probabilistic interpretation of the relative frequencies . for each observable we would like to deduce measures for its central tendency and statistical dispersion from the respective probability distribution .
unfortunately , it is infeasible to fit all obtained distributions with one common template distribution .
such a procedure would lack a clear statistical justification , and it also appears impractical as the distributions that we obtain differ substantially in their shapes .
we therefore choose a different approach .
we consider the median of a distribution as its centre and we use the @xmath64 ` confidence ' interval around it as a measure for its spread .
of course , this range of the confidence interval is reminiscent of the @xmath65 range of a normal distribution .
more precisely , for an observable @xmath66 with probability density @xmath67 we will summarize its central tendency and variability in the following form @xcite , @xmath68 here , @xmath69 and @xmath70 denote the @xmath71- and @xmath72-quantiles with respect to the density function @xmath67 .
the central value @xmath73 is the median of @xmath67 and thus corresponds to its @xmath74-quantile .
all three values of @xmath66 can be calculated from the quantile function @xmath75 , @xmath76 : p \leq f(x)\right\ } \,,\quad f(x ) = \int_{x_\textrm{min}}^x dt \ :
f(t ) \,,\end{aligned}\ ] ] where @xmath77 stands for the cumulative distribution function of @xmath66 .
we then have : @xmath78 intuitively , the intervals from @xmath79 to @xmath69 , @xmath73 , and @xmath70 respectively correspond to the @xmath66 ranges into which @xmath71 , @xmath74 or @xmath72 of all hits fall .
this is also illustrated in the histogram for @xmath80 in fig .
moreover , we have included vertical lines into each plot to indicate the respective positions of @xmath69 , @xmath73 , and @xmath70 . in our case
the median is a particularly useful measure of location .
first of all , it is resistant against outliers and hence an appropriate statistic for such skewed distributions as we observe them .
but more importantly , the average absolute deviation from the median is minimal in comparison to any other reference point .
the median is thus the best guess for the outcome of a measurement if one is interested in being as close as possible to the actual result , irrespective of the sign of the error . on the technical side the definition of
the median fits nicely together with our method of assessing statistical dispersion .
the @xmath81 confidence interval as introduced above is just constructed in such a way that equal numbers of hits lie in the intervals from @xmath69 to @xmath73 and from @xmath73 to @xmath70 , respectively . in this sense ,
our confidence interval represents a symmetric error with respect to the median . as a test of the robustness of our results
, we checked the dependence of our distributions on the precise choice of the experimental error intervals .
the results presented here proved insensitive to these variations .
for definiteness , we therefore stick to the @xmath51 intervals .
we also checked the effect of taking the random @xmath82 factors to be distributed uniformly on a linear instead of a logarithmic scale .
again , the results proved to be robust .
_ mass hierarchy _ an important open question which could help unravel the flavour structure of the neutrino sector is the mass hierarchy .
since the sign of @xmath83 is not yet known , we can not differentiate with current experimental data between a normal hierarchy with one heavy and two light neutrino mass eigenstates and an inverted hierarchy , which has two heavy and one light neutrino mass eigenstate .
measuring the mikheyev - smirnov - wolfenstein ( msw ) effect of the earth could resolve this ambiguity . with the procedure described above
, all hits match the structure of the normal hierarchy and there are no examples with inverted hierarchy .
it is however notable that imposing the structure of the neutrino mass matrix given by eq . alone
does not exclude the inverted mass hierarchy . only additionally imposing the measured bounds on the mixing angles
rejects this possibility .
_ mixing angles _ the mixing in the lepton sector is described by the matrix @xmath84 given in eq . .
of the three angles , two are only bounded from one side by experiment : for the largest mixing angle @xmath85 there exists a lower bound , whereas the smallest mixing angle @xmath4 is so far only bounded from above .
recent results from t2k @xcite , minos @xcite and the preliminary result of double chooz @xcite point to a value of @xmath4 just below the current experimental bound .
the respective best fit points , assuming a normal hierarchy , are @xmath86 ( t2k ) , @xmath87 ( minos ) and @xmath88 ( double chooz ) .
the 90@xmath89 and 68@xmath89 confidence regions respectively read @xmath90 with the procedure described above , we find sharp predictions for the smallest and the largest mixing angle within the current experimental bounds , @xmath91 the corresponding distributions are shown in fig .
[ fig_mixing ] .
these results are quite remarkable : the atmospheric mixing angle points to maximal mixing , while the rather large value for @xmath4 is consistent with the recent t2k , minos and double chooz results .
in our monte - carlo study we observe that the dominant contribution to the strong mixing in the lepton sector is primarily due to the neutrino mass matrix @xmath92 .
the numerical results are not much affected by including the charged lepton mixing matrix @xmath42 .
the pmns matrix is thus approximately given by the matrix @xmath43 which diagonalizes the light neutrino mass matrix @xmath92 .
_ absolute mass scale _
the absolute neutrino mass scale is a crucial ingredient for the study of neutrinoless double - beta decay and leptogenesis .
although inaccessible in neutrino oscillation experiments , different experimental setups have succeeded in constraining this mass scale .
cosmological observations of the fluctuations in the cosmic microwave background , of the density fluctuations in the galaxy distribution and of the lyman-@xmath93 forest yield a constraint for the sum of the light neutrino masses , weighted by the number of spin degrees of freedom per majorana neutrino , @xmath94 , @xcite @xmath95 the planck satellite is expected to be sensitive to values of @xmath96 as low as roughly @xmath97 ev @xcite .
a further constraint arises from measuring the @xmath98-spectrum in tritium decay experiments .
the current bound @xcite is @xmath99 by comparison , the katrin experiment , which will start taking data soon , aims at reaching a sensitivity of @xmath100 @xcite .
finally , the neutrino mass scale can also be probed by neutrinoless double - beta decay .
the relevant effective mass is @xmath101 here , ref .
@xcite claims a value of @xmath102 ev .
dedicated experiments , such as gerda @xcite with a design sensitivity of @xmath103 ev , are on the way .
note that @xmath104 does not only depend on the absolute neutrino mass scale and the mixing angles , but also on the phases @xmath105 and @xmath106 in the pmns matrix .
we find sharp predictions for the neutrino mass parameters discussed above .
the lightest neutrino , @xmath107 , is found to be quite light , cf .
[ fig_m1mtritium ] , @xmath108 hence favouring a relatively low neutrino mass scale beyond the reach of current and upcoming experiments .
more precisely , we find for the neutrino mass parameters discussed above : @xmath109 _ cp - violating phases _ the small value of the mass parameter measured in neutrinoless double - beta decay , @xmath104 , is due to the relative minus sign between the @xmath110 and @xmath111 terms in eq . , caused by a strong peak of the value for the majorana phase @xmath106 at @xmath112 , @xmath113 this is depicted in fig .
[ fig_masses ] .
an analytic analysis of how this phenomena arises from the structure of the neutrino mass matrix , cf .
eq . , is presented in appendix [ app_alpha1 ] . for the other majorana phase @xmath114 and the dirac phase @xmath115
we find no such distinct behaviour but approximately flat distributions .
_ leptogenesis parameters _ finally , leptogenesis @xcite links the low energy neutrino physics to the high energy physics of the early universe .
the parameters that capture this connection are the effective neutrino mass of the first generation @xmath116 and the @xmath117 violation parameter @xmath118 @xcite , @xmath119 ^ 2 } { 8 \pi ( h^{(\nu ) \ , \dagger } h^{(\nu)})_{11 } } f\left(\frac{m^2_j}{m^2_{1 } } \right)\,,\ ] ] with @xmath120 and @xmath121 denoting the masses of the heavy neutrinos . here
, @xmath116 determines the coupling strength of the lightest of the heavy neutrinos to the thermal bath and thus controls the significance of wash - out effects .
it is bounded from below by the lightest neutrino mass @xmath110 .
the absolute value of the @xmath117 violation parameter @xmath118 is bounded from above by @xcite @xmath122 with the procedure described above , we find @xmath123 and hence a clear preference for the strong wash - out regime @xcite .
notice that there typically is a hierarchy between @xmath116 and @xmath110 of about one order of magnitude .
the relative frequency of the @xmath117 violation parameter @xmath118 peaks close to the upper bound @xmath124 , with the majority of the hits lying within one order of magnitude or less below @xmath124 , cf .
[ fig_cp ] .
this justifies the use of @xmath124 when estimating the produced lepton asymmetry in leptogenesis . here ,
in the discussion of @xmath118 , we assumed hierarchical heavy neutrinos , @xmath125 .
_ theoretical versus experimental input _
+ the results of this section are obtained by combining two conceptually different inputs , on the one hand the hierarchy structure of the neutrino mass matrix @xmath92 given by eq . and on the other hand the experimentally measured constraints listed in eq . . in general
, the distributions presented above really arise from the interplay between both of these ingredients .
for example , the hierarchy structure alone does not favour a large solar mixing angle @xmath11 and the ratio @xmath126 tends to be too large ( cf .
this discrepancy is eased by generating the random coefficients in eq . via the seesaw mechanism .
imposing the experimental constraints finally singles out the subset of parameter sets used for the distributions presented above . as another example , consider the smallest mixing angle @xmath4 and the lightest neutrino mass eigenstate @xmath110 . in these cases ,
the hierarchy structure of the neutrino mass matrix automatically implies small values , similar to those shown in the distributions above .
however , the exact distributions including the precise position of the peaks only arise after implementing the experimental constraints . a notable exception to
this scheme is the majorana phase @xmath106 . here
the peak at @xmath14 is a result of the hierarchy structure of the neutrino matrix @xmath92 alone , as demonstrated in appendix [ app_alpha1 ] .
in summary , we find that starting from a flavour symmetry which accounts for the measured quark and lepton mass hierarchies and large neutrino mixing , the present knowledge of neutrino parameters strongly constrains the yet unknown observables , in particular the smallest mixing angle @xmath4 , the smallest neutrino mass @xmath110 , and the majorana phase @xmath106 . this statement is based on a monte - carlo study : treating unspecified @xmath0 parameters of the considered froggatt - nielsen model as random variables , the observables of interest are sharply peaked around certain central values .
we expect that these results hold beyond froggatt - nielsen flavour models .
an obvious example are extradimensional models which lead to the same type of light neutrino mass matrix ( cf .
@xcite ) . on the other hand , quark - lepton mass hierarchies and the presently known neutrino observables can not determine the remaining observables in a model - independent way .
this is illustrated by the fact that our present knowledge about quark and lepton masses and mixings is still consistent with @xmath127 as well as with an inverted neutrino mass hierarchy ( cf .
@xcite ) . as a consequence ,
further measurements of neutrino parameters will be able to falsify certain patterns of flavour mixing and thereby provide valuable guidance for the theoretical origin of quark and lepton mass matrices .
* acknowledgements * the authors thank g. altarelli , f. brmmer , g. ross , d. wark , w. winter and t. yanagida for helpful discussions and comments .
this work has been supported by the german science foundation ( dfg ) within the collaborative research center 676 `` particles , strings and the early universe '' .
the complex phases of the @xmath0 coefficients in the neutrino mass matrix @xmath128 and the lepton mass matrix @xmath39 are randomly distributed .
one would thus naively expect that also the majorana phases @xmath106 and @xmath114 in the pmns matrix can take arbitrary values . by contrast , the distribution of values for @xmath106 that we obtain from our numerical monte - carlo study , cf . fig . [ fig_masses ] , clearly features a prominent peak at @xmath14 . in this appendix
we shall demonstrate by means of a simplified example how the structure of the neutrino mass matrix @xmath128 may partly fix the phases of the corresponding mixing matrix @xmath43 .
consider the following simplified majorana mass matrix @xmath128 for the light neutrinos , @xmath129 where @xmath130 is an arbitrary complex phase between @xmath131 and @xmath132 .
for simplicity , let us neglect any effects on the mixing matrix @xmath43 from the diagonalization of @xmath39 .
that is , we define @xmath43 such that @xmath133 , with @xmath134 denoting the eigenvalues of @xmath135 , @xmath136 + \mathcal{o}\left(\eta^4\right)\,,\\ \frac{m_3 ^ 2}{v^2}= & \
: 4\left(1 + \eta^2\left[1-\sin^2\left(\varphi/2\right)\right]\right ) + \mathcal{o}\left(\eta^4\right)\nonumber \,.\end{aligned}\ ] ] notice that the first two mass eigenvalues are nearly degenerate .
this is a consequence of the particular hierarchy pattern of the matrix @xmath128 which originally stems from the equal flavour charges of the @xmath137 and @xmath138 multiplets .
the relative sign of the @xmath139 contributions to @xmath140 and @xmath141 eventually shows up again in entries of @xmath43 , for instance , @xmath142\right ) + \mathcal{o}\left(\eta\right ) \,.\end{aligned}\ ] ] with @xmath143 the phase @xmath144 \:\textrm{mod}\:\pi\right)$ ] in the matrix @xmath43 represents the analog of the majorana phase @xmath106 in the pmns matrix , cf .
according to our explicit results for @xmath145 and @xmath146 it is independent of the arbitrary phase @xmath130 to leading order in @xmath8 , @xmath147 + \frac{i}{2 } \textrm{arg } \left[-z\right ] \right)\right ] \:\textrm{mod}\:\pi\right ) = \pi \,.\end{aligned}\ ] ] in a similar way we may determine the phase analogous to the majorana phase @xmath114 .
however , due to the hierarchy between the mass eigenvalues @xmath110 and @xmath148 , the first and third column of the matrix @xmath43 differ significantly from each other , thus leading to a phase that depends on @xmath130 at all orders of @xmath8 . including corrections to all orders in @xmath8 and scanning over the phase @xmath130
numerically shows that the maximal possible deviation of @xmath106 from @xmath112 is , in fact , of order @xmath149 . adding more complex phases to the matrix @xmath128 in eq .
gradually smears out the peak in the distribution of @xmath106 values .
the distribution that is reached in the case of six different phases is already very similar to the one in fig .
[ fig_masses ] .
we conclude that despite the need for corrections the rough picture sketched in this appendix remains valid : the hierarchy pattern of the neutrino mass matrix directly implies that @xmath106 tends to be close to @xmath14 . for a review and references
see , for example , + s. raby , eur .
j. c * 59 * , 223 ( 2009 ) c. d. froggatt and h. b. nielsen , nucl .
b * 147 * , 277 ( 1979 ) y. grossman and m. neubert , phys .
b * 474 * , 361 ( 2000 ) ; t. gherghetta and a. pomarol , nucl .
b * 586 * , 141 ( 2000 ) ; w. buchmuller , k. hamaguchi , o. lebedev and m. ratz , nucl .
b * 785 * , 149 ( 2007 ) j. sato and t. yanagida , phys .
b * 430 * , 127 ( 1998 ) ; n. irges , s. lavignac and p. ramond , phys .
d * 58 * , 035003 ( 1998 ) w. buchmuller and t. yanagida , phys .
b * 445 * , 399 ( 1999 ) f. vissani , jhep * 9811 * , 025 ( 1998 ) l. j. hall , h. murayama and n. weiner , phys .
lett . *
84 * , 2572 ( 2000 ) ; j. sato and t. yanagida , phys .
b * 493 * , 356 ( 2000 ) ; f. vissani , phys .
b * 508 * , 79 ( 2001 ) k. abe _ et al .
_ [ t2k collaboration ] , phys .
lett . * 107 * , 041801 ( 2011 ) p. adamson _ et al . _
[ minos collaboration ] , phys .
lett . * 107 * , 181802 ( 2011 ) .
talk given by h. de kerret at the sixth international workshop on low energy neutrino physics ( lownu11 ) at seoul , korea during november 9 - 12 , 2011 k. nakamura _ et al .
_ [ particle data group ] , j. phys .
g * 37 * , 075021 ( 2010 ) w. buchmuller , k. schmitz and g. vertongen , phys .
b * 693 * , 421 ( 2010 ) ; w. buchmuller , k. schmitz and g. vertongen , nucl . phys .
b * 851 * , 481 ( 2011 ) ; w. buchmuller , v. domcke and k. schmitz , arxiv:1202.6679 [ hep - ph ] .
[ planck collaboration ] , esa - sci(2005)1 ( 2006 ) , arxiv : astro - ph/0604069 m. beck [ katrin collaboration ] , j. phys .
. ser .
* 203 * ( 2010 ) 012097 h. v. klapdor - kleingrothaus , a. dietz , h. l. harney , i. v. krivosheina , mod . phys .
lett . * a16 * , 2409 - 2420 ( 2001 ) g. meierhofer [ gerda collaboration ] , j. phys .
* 312 * ( 2011 ) 072011 .
m. fukugita , t. yanagida , phys .
* b174 * , 45 ( 1986 ) . for reviews containing the relevant formulae
see , for example , + w. buchmuller , r. d. peccei and t. yanagida , ann .
nucl . part .
* 55 * , 311 ( 2005 ) ; s. davidson , e. nardi and y. nir , phys .
rept . *
466 * , 105 ( 2008 ) s. davidson and a. ibarra , phys .
b * 535 * , 25 ( 2002 ) ; see also k. hamaguchi , h. murayama and t. yanagida , phys .
d * 65 * , 043512 ( 2002 ) g. altarelli , f. feruglio and i. masina , jhep * 0301 * , 035 ( 2003 ) ; i. masina and c. a. savoy , phys .
d * 71 * , 093003 ( 2005 ) [ hep - ph/0501166 ] .
f. plentinger , g. seidl and w. winter , phys .
d * 76 * , 113003 ( 2007 ) ; f. plentinger , g. seidl and w. winter , nucl .
b * 791 * , 60 ( 2008 ) t. asaka , w. buchmuller , l. covi , phys .
lett . *
b563 * , 209 - 216 ( 2003 ) for a review and references see , for example , + g. altarelli , f. feruglio , rev .
phys . *
82 * , 2701 - 2729 ( 2010 ) ; h. ishimori , t. kobayashi , h. ohki , y. shimizu , h. okada , m. tanimoto , prog .
suppl . * 183 * , 1 - 163 ( 2010 ) | flavour symmetries of froggatt - nielsen type can naturally reconcile the large quark and charged lepton mass hierarchies and the small quark mixing angles with the observed small neutrino mass hierarchies and their large mixing angles .
we point out that such a flavour structure , together with the measured neutrino mass squared differences and mixing angles , strongly constrains yet undetermined parameters of the neutrino sector . treating unknown @xmath0 parameters as random variables ,
we obtain surprisingly accurate predictions for the smallest mixing angle , @xmath1 , the smallest neutrino mass , @xmath2 , and one majorana phase , @xmath3 .
desy 11 - 173 + december 2011 1.5 cm * predicting @xmath4 and the neutrino mass scale from quark lepton mass hierarchies * w. buchmller , v. domcke , and k. schmitz + _ deutsches elektronen - synchrotron desy , 22607 hamburg , germany _ 0.75 cm |
spectroscopic properties of galaxies have provided valuable information on different aspects of galaxy evolution , such as star formation histories along the hubble sequence , chemical evolution of gas and stars , and the physical conditions associated with agn and starburst activity in the central regions of galaxies ( e.g. , kennicutt 1998 , 1999 ) . in particular , spectroscopic studies of galaxies at optical wavelengths have played vital roles not only in determining the current and the past star formation rate but also in deducing the physical mechanisms relevant to the time evolution of star formation ( see kennicutt 1998 for a recent review ) .
the change in the galactic global star formation rate with redshift has recently been investigated for a variety of different distant star forming objects by many authors ( e.g. , madau 1996 ; steidel et al .
1999 ; ellis 1997 ) .
these recent works have suggested that optical data alone can not provide a comprehensive view of the star formation properties of distant galaxies owing to heavy dust extinction .
a major and important challenge is to properly assess the extent to which the stellar light is affected by dust extinction .
future near - infrared spectroscopic studies of the redshifted h@xmath9 emission from distant , dust - obscured galaxies is crucial to revealing the degree of extinction and reddening and thereby clarifying the real star formation histories of these galaxies ( kennicutt 1999 ) . detailed spectroscopic studies of distant rich clusters ( dressler & gunn 1983 , 1992 ; couch & sharples 1987 ; poggianti et al .
1999 ) have provided important clues as to the nature and evolution of galaxies in these dense environments . in particular ,
dressler & gunn ( 1983 , 1992 ) discovered galaxies with no detectable emission lines but with very strong balmer line absorption ( the so - called `` e+a '' galaxies ) and concluded that such galaxies were `` poststarburst '' objects seen within @xmath10gyr of a severely curtailed burst of star formation .
furthermore , couch & sharples ( 1987 ) made the first attempts to establish an evolutionary link between star - forming , poststarburst , and passively evolving galaxies observed in distant clusters .
these attempts have been progressed significantly further by dressler et al .
( 1999 ) and poggianti et al .
( 1999 ) through their detailed modeling of an extensive data base of spectroscopic observations they obtained in 10 rich clusters in the range @xmath11 .
the wealth of data allowed them to delineate the different spectral classes pertinent to distant cluster populations much more clearly , with three distinct classes of emission line object [ e(a ) , e(b ) , e(c ) ] being identified on the basis of the strength of [ oii]@xmath13727 emission and h@xmath2 absorption , two poststarburst classes ( a+k and k+a ) identified amongst galaxies with no detectable [ oii ] emission but enhanced h@xmath2 absorption , and a passive `` k '' class where neither of these features had any prominence .
attempts were made to identify the evolutionary pathways between these different classes of object in order to identify the most likely progenitor to the sizeable population of poststarburst galaxies .
galaxies with e(a ) spectra characterized by strong balmer absorption [ ew(h@xmath12 ] and relatively modest [ oii ] emission , were previously interpreted ( in the context of dust - free models ) as poststarburst galaxies with some residual star formation ( couch et al .
however , poggianti et al s detailed spectral analysis and comparison with liu & kennicutt s ( 1995 ) sample of spectra for present - day merging galaxies suggested that the e(a ) types were more likely dusty starburst galaxies .
this suggestion has subsequently been reinforced by the observation that a significant number of e(a ) galaxies are either luminous infrared galaxies ( lirgs ) or ultra - luminous infrared galaxies ( ulirgs ) , objects which are often dominated by dusty starbursts ( poggianti & wu 2000 ) .
independent evidence for the presence and importance of dust in determining the observed properties of distant cluster galaxies has been presented by other authors .
smail et al .
( 1999 ) found that 5 out of 10 galaxies classified as either a+k or k+a types by dressler et al .
( 1999 ) in abell 851 ( at @xmath13 ) were detectable in deep 1.4ghz continuum maps obtained with the vla . with this radio emission most likely attributable to starburst activity
, the lack of any such signature in the optical spectrum could only imply that it must be totally obscured by dust .
further to this , owen et al .
( 1999 ) investigated the radio properties of galaxies in a rich cluster at z @xmath14 ( a2125 ) which had large blue galaxy fractions ( 0.19 ) and found that the optical line luminosities ( e.g. , h@xmath9+[nii ] ) were often weaker than one would expect for the star formation rates implied by the radio emission .
they accordingly suggested that dust obscuration , which is larger than is usually found locally , hides most of optical light from the star - forming regions .
although these observational results have identified the possible importance of dust and dusty starbursts in distant cluster populations , little theoretical work has been done to address the formation and evolution of such galaxies .
the purpose of this paper is to report the results of a _ numerical _ investigation of the spectroscopic evolution of dusty starburst galaxies and possible evolutionary links between galaxies of different spectral class . using a one - zone chemical and spectrophotometric evolution model
, we consider in particular , the degree to which dust extinction ( represented by @xmath0 ) depends on the age of the starburst populations and thereby try to clarify the time evolution in the equivalent width of the h@xmath2 absorption and [ oii ] emission line features .
the main questions that we address are : ( 1)when and how does a disk galaxy hosting a dusty starburst develop an e(a)-type spectrum ?
( 2)what are the requisite physical conditions for the formation of e(a ) galaxies ?
( 3)can a dusty galaxy show an a+k or k+a spectrum even in its starburst phase ?
( 4)are there any evolutionary links between the e(a ) , e(b ) , e(c ) , a+k , k+a , and k spectral types ?
a key feature of our approach is to adopt the _ selective dust extinction _ framework originally proposed by poggianti & wu ( 2000 ) whereby the extinction suffered by a stellar population is inversely proportional to its age with the youngest populations having the greatest extinction .
we confirm that such age - dependent extinction is essential to reproducing the spectral properties of the putative dusty starburst population in distant clusters .
the layout of this paper is as follows : in 2 , we summarize the numerical models used in the present study and describe briefly the methods for deriving spectral energy distributions ( seds ) affected by internal dust extinction . in 3 , we present our numerical results on the time evolution of dusty starburst galaxies in the ew([oii])-ew(h@xmath2 ) plane . in 4 , we discuss the possible evolutionary connections between the different spectral types ( e.g. , e(a ) , e(b ) , and k+a ) in the context of our dusty starburst models .
the conclusions of the present study are given in 5 .
in the present paper , each spectral class is defined as follows : e(a ) for galaxies with 5 @xmath15 ew([oii ] ) @xmath7 40 and 4 @xmath15 ew(h@xmath2 ) , e(b ) for 40 @xmath15 ew([oii ] ) and ew(h@xmath2 ) @xmath7 4 , e(c ) for 5 @xmath15 ew([oii ] ) @xmath7 40 and ew(h@xmath2 ) @xmath7 4 , a+k for ew([oii ] ) @xmath7 5 and 7 @xmath15 ew(h@xmath2 ) , k+a for ew([oii ] ) @xmath7 5 and 3 @xmath15 ew(h@xmath2 ) @xmath7 7 , k for ew([oii ] ) @xmath7 5 and ew(h@xmath2 ) @xmath7 3 .
we adopt a one - zone chemical and spectrophotometric evolution model of a disk galaxy with a starburst , and thereby investigate the time evolution of the line strength of h@xmath2 absorption and that of [ oii ] emission , both of which are important properties for classifying the spectral types of galaxies in distant clusters ( dressler et al .
the present model is essentially the same as that adopted in our previous studies ( shioya & bekki 1998 , 2000 ) , with the details of our treatment of chemical evolution given in shioya & bekki ( 1998 ) .
accordingly , we describe only briefly the present spectrophotometric model . one of the main differences between our model and previous ones ( e.g. , couch & sharples 1987 ; newberry et al . 1990
; charlot & silk 1994 ; jablonka & alloin 1995 ; poggianti & barbaro 1996 ; poggianti et al . 1999 ) is that we explicitly include the effects of dust on both the photometric and spectroscopic properties of stellar populations as also the gaseous emissions . with this new development to our code
, we can clearly demonstrate how the dust extinction affects the time evolution of the line strengths of h@xmath2 and [ oii ] ( ew(h@xmath2 ) and ew([oii ] ) ) in dusty starburst galaxies .
we follow the chemical evolution of galaxies by using the model described in matteucci & tornamb ( 1987 ) which includes metal - enrichment processes of type ia and ii supernovae ( snia and snii ) .
we adopt the salpeter initial mass function ( imf ) , @xmath16 , with upper mass limit @xmath17 and lower mass limit @xmath18 .
we calculate the photometric properties of galaxies as follows : the monochromatic flux of a galaxy with age @xmath19 , @xmath20 , is described as : @xmath21 where @xmath22 is the monochromatic flux of a single stellar population with age @xmath23 and metallicity @xmath24 , and @xmath25 is the time - dependent star formation rate described below . in the present study ,
we use the spectral library gissel96 which is the most recent version of that published by bruzual & charlot ( 1993 ) .
the star formation history of a disk galaxy is characterized by three epochs .
the first is the epoch of galaxy formation ( @xmath26 ) during which the interstellar gas is consumed by star formation at a moderate rate . in the present study
, @xmath26 is 0 for all models .
the second is @xmath27 at which the starburst begins in the disk . here
@xmath27 is considered to be a free parameter .
the third is @xmath28 at which @xmath29 star formation ceases is defined as the epoch at which the stellar mass fraction reaches a value @xmath30 in our models ; @xmath30 is also assumed to be a free parameter which controls the duration of starburst . throughout the evolution of disk galaxies ,
the star formation rate is assumed to be proportional to the gas mass fraction ( @xmath31 ) : @xmath32 where @xmath33 is a parameter which controls the star formation rate .
this parameter @xmath33 is given as follows : @xmath34 in the following , we refer to these three different epochs as the `` prestarburst '' , `` starburst '' , and `` poststarburst '' phases , respectively .
the value of the @xmath35 parameter , which controls the star formation histories of disk galaxies , is suggested to vary for different hubble types ( e.g. , arimoto , yoshii , & takahara 1992 ) .
for example , @xmath35 = 0.225 @xmath36 corresponds to a plausible star formation rate for sb disks ( e.g. , arimoto et al .
1992 ) . in comparison , observations of starburst galaxies ( e.g. , planesas et al
. 1997 ) would suggest that @xmath37 is an order of magnitude larger than @xmath35 ( i.e. , @xmath37 = @xmath38 ) . by changing the value of @xmath37
, we can investigate how the strength of a starburst influences the evolution in spectral type within a dusty galaxy .
we adopt the selective dust extinction model investigated in detail by shioya & bekki ( 2000 ) . in this model , the amount of dust extinction associated with the starburst is assumed to depend on the age of stellar population : @xmath39 where @xmath40 and @xmath41 are free parameters which control the the extinction at any given time during starburst . based on the monochromatic flux derived in equation ( 1 ) and the value of @xmath0 given by equation ( 4 ) , we calculated a ` dust - absorbed ' sed for galaxies of different age @xmath19 using the extinction law inferred by calzetti et al .
( 1995 ) and adopting the so - called screen model . to derive fluxes for various gaseous emission lines ( h@xmath2 and [ oii ] )
, we first calculate the number of lyman continuum photons , @xmath42 , by using a non - absorbed sed .
if all the lyman continuum photons are consumed in ionizing the surrounding gas , the luminosity of h@xmath43 can be calculated according to the relation : @xmath44 ( leitherer & heckman 1995 ) . to calculate luminosities of other emission lines , e.g. , [ oii ] and h@xmath2
, we use the luminosity ratios relative to h@xmath43 tabulated in pegase ( fioc & rocca - volmerange 1997 ) , which are calculated for an electron temperature of 10,000k and an electron density of 1@xmath45 . the strength of absorption lines are derived using gissel96 ( bruzual & charlot 1993 ) . thus the seds derived here comprise the following components : stellar continuum , gaseous emission , and stellar absorption . shioya & bekki ( 2000 ) have already demonstrated the importance of selective dust extinction in reproducing e(a ) type galaxy spectra by comparing the model results with and without dust extinction . furthermore , they pointed out that a dusty starburst galaxy with selective dust extinction evolves from an e(b ) , to an e(a ) , to an a+k , to a k+a , and finally to a k type galaxy . since they have already indicated the fundamental role of the selective dust extinction in spectroscopic evolution of dusty starburst galaxies , we detail here the parameter dependence of the spectroscopic evolution in the ew([oii])-ew(h@xmath2 ) plane . to demonstrate the importance of each of the six free parameters , @xmath40 , @xmath41 , @xmath27 , @xmath46 , @xmath37 , and @xmath30 , we firstly compute a fiducial model ( model 1 ) with @xmath40 = 5.0 mag , @xmath41 = @xmath47 yr , @xmath27 = 7.64 gyr , @xmath35 = 0.225 gyr@xmath48 , @xmath37 = 2.25 gyr@xmath48 , and @xmath30 = 0.95 .
model 1 can then be compared with a model without dust extinction ( model 2 ) , to show the typical behavior of selective dust extinction in the evolution of starburst galaxies on the ew([oii])-ew(h@xmath2 ) plane .
the dependence of the evolution on the six parameters , @xmath40 , @xmath41 , @xmath27 , @xmath35 , @xmath37 , and @xmath30 is then investigated via the computation of 17 models with a range of parameter values : model 1 , 3 , 4,and 5 with different @xmath40 , model 1 , 6 , and 7 with different @xmath41 , model 1 , 8 , 9 , 10 , and 11 with different @xmath27 , model 1 , 12 , and 13 with different @xmath35 , model 1 , 14 , and 15 with different @xmath37 , and model 1 , 16 , and 17 with different @xmath49 .
the parameter values are given in table 1 .
we assume that the initial gas mass of the model galaxy is @xmath50 .
figure 1 shows the star formation history and the color evolution ( in @xmath51 ) for the fiducial model ( model 1 ) with @xmath35 = 0.225 @xmath36 ( corresponding to sb type star formation history ) . during the starburst ,
the @xmath51 color becomes rather blue ( 0.2 mag at @xmath19 = 8.56 gyr ) due to the formation of massive stars .
however , the color is not so blue compared with the model without selective dust extinction ( model 2 ) because of the heavy dust obscuration suffered by the young stellar population in the fiducial model .
this result implies that selective dust extinction affects the global colors as well as the spectral characteristics of starburst galaxies .
figure 2 shows the evolution of the emission and absorption lines by plotting their strength at @xmath19 = 1 , 7.64 , 8.56 , 8.58 , and 13.0 gyr for the fiducial model . before the starburst ( @xmath19 @xmath15 7.64 gyr )
, the model clearly shows a number of strong emission lines , whereas the absorption lines ( e.g. , the h@xmath2 absorption line ) are rather weak as a result of the continuous and moderate star formation . during the starburst ( 7.64 @xmath7 @xmath19 @xmath7 8.58 gyr ) ,
the emission lines do not become so remarkably strong ; nevertheless , the star formation rate becomes more than a factor of five larger than that of the prestarburst phase .
this is because the ionizing photons from very young ob stars are very heavily obscured by the dust .
after the starburst ( @xmath19 @xmath5 8.58 gyr ) , the emission lines become even less remarkable due to the death of the young , massive stars and the intermediate aged ones as well . on the other hand , the stellar absorption lines ( e.g. , h@xmath2 )
become prominent after the starburst ( @xmath19 @xmath5 8.58 gyr ) owing to the presence of an older stellar population .
figure 3 describes the time evolution of ew([oii ] ) and ew(h@xmath2 ) , which clearly demonstrates the importance of the selective dust extinction in the spectroscopic evolution of dusty starburst galaxies .
the equivalent width of [ oii ] does _ not _ increase and is well below 40 during the starburst .
the equivalent width of h@xmath2 becomes rapidly smaller just after the starburst begins , although the h@xmath2 line can still be observed as an absorption line .
a few @xmath52 years after the starburst begins , the equivalent width of h@xmath2 becomes larger than 4 .
consequently , this model reproduces the properties of the e(a ) class with ew(h@xmath2 ) @xmath53 and relatively modest [ oii ] emission , ew([oii])@xmath54 ; the location in the ew([oii])-ew(h@xmath2 ) plane is shown in figure 4 .
the physical reason for the successful reproduction of the model is clearly described in our previous paper ( shioya & bekki 2000 ) .
the time scale during which the model is located within the e(a ) region is estimated to be @xmath55gyr .
these results thus confirm the earlier suggestion of poggianti & wu ( 2000 ) of the vital role of selective dust extinction in reproducing e(a ) spectra .
figure 5 shows the time evolution of spectroscopic properties in the ew([oii])-ew(h@xmath2 ) plane for models with different values of @xmath40 ( 1 @xmath7 @xmath40 @xmath7 10.0 mag for model 2 , 3 , 4 , and 5 ) .
since parameters other than @xmath40 are fixed at the same values adopted for the fiducial model , this figure demonstrates how the evolution of dusty starburst galaxies in the ew([oii])-ew(h@xmath2 ) plane differs according to the amount of dust extinction . as can be seen in the figure
, the locus on the plane is appreciably different between the models , which implies that the degree of dust extinction ( represented by @xmath40 ) is an important influence on the observed spectroscopic evolution . during the starburst , the model with the larger @xmath40 shows weaker [ oii ] emission , weaker h@xmath2 emission , and stronger h@xmath2 absorption .
thus the model with the largest @xmath40 ( e.g. , model 5 with @xmath40 = 10 mag ) is located towards the bottom
right of the e(a ) region during the dusty starburst .
figure 6 describes the evolution in the ew([oii])-ew(h@xmath2 ) plane for the models with different values of @xmath41 ( @xmath56yr for model 1 , 6 , and 7 ) .
parameters other than @xmath41 are fixed at the same values of the fiducial model .
this figure clearly demonstrates that the time - scale during which young massive stars are heavily obscured by dust also governs the spectroscopic evolution . for the parameter range adopted here ,
a larger number of young massive stars with different ages are obscured by dust for the model with larger @xmath41 , i.e. , the age - dependent dust extinction is weaker and thus young stars with different ages ( @xmath57yr ) are more uniformly obscured by dust .
therefore , the model with larger @xmath41 shows lower ew([oii ] ) and higher ew(h@xmath2 ) and , as a result , is located in the bottom - right hand region of the ew([oii])-ew(h@xmath2 ) plane during the starburst .
we stress here that for the model with larger @xmath41 (= @xmath58yr and @xmath59yr ; model 6 and 7 , respectively ) , the time - scale during which the galaxies show an e(a)-type spectrum is appreciably shorter compared to the fiducial model .
furthermore , the galaxies show a k+a / a+k spectral signature in the late phase of the starburst .
this result suggests that not only the galaxies with e(a ) spectra but also those with k+a
/ a+k spectra are possible dusty starburst candidates in distant clusters . for the models with @xmath60yr ,
not only young massive stars ( e.g. , ob stars ) which emit most of the ionizing photons but also late b and a stars are strongly affected by dust extinction .
hence a galaxy with a @xmath41 value of this size does not show strong h@xmath2 absorption during the starburst and poststarburst phases .
therefore the galaxy can be located in k+a region even in the dusty starburst phase .
this can be contrasted with models with @xmath61yr , where the emission lines from ionized gas regions around massive stars are not obscured so heavily by dust . as a result of this , even if we adopt a rather large value for @xmath40 ( i.e. , more than 10 mag ) , neither the [ oii ] nor h@xmath2 lines are weakened . in this case , a dusty starburst galaxy shows an e(b ) spectrum during its starburst phase .
these results on the @xmath41dependences within our models confirms that age - dependent dust extinction is a crucially important factor for determining the spectroscopic properties of dusty starburst galaxies .
the present one - zone model , however , can not determine with any great certainty what controls the age - dependence ( i.e. , the parameter value of @xmath41 ) owing to the limitations of the model .
we further discuss this problem in a later section .
figure 7 describes how the star formation history and the evolution of spectroscopic properties on the ew([oii])-ew(h@xmath2 ) plane are controlled by the parameter @xmath27 in dusty starburst galaxies . for models with a smaller @xmath27 ,
the starburst occurs at earlier epochs and the magnitude of the starburst is larger .
this is because there is more gas to fuel starbursts at earlier times .
the evolution of galaxies on the ew([oii])-ew(h@xmath2 ) plane is not so remarkably different between the models with different @xmath27 ( = 4.46 , 6.93 , 7.64 , 8.47 , and 9.45 gyr : corresponding to model 8 , 9 , 1 , 10 , and 11 , respectively ) .
the model with smaller @xmath27 ( e.g. , model 9 ) , which represents a higher redshift starburst galaxy , shows discernibly lower ew([oii ] ) and higher ew(h@xmath2 ) and thus is located at the bottom - right of the e(a ) region during the starburst .
this is primarily because a larger number of young massive stars are created and then obscured preferentially in the model with smaller @xmath27 .
these results firstly imply that irrespective of the epoch at which a dusty starburst occurs , disk galaxies can exhibit an e(a ) spectrum during this phase if the extinction parameters are within certain reasonable ranges .
secondly , they imply that the higher redshift dusty starburst galaxies are more likely to be located to the right and to the bottom of the e(a ) region in the ew([oii])-ew(h@xmath2 ) plane
. it would be very interesting if future spectroscopic observations could confirm this predicted redshift - dependent behavior in dusty starburst galaxies .
figure 8 describes how the evolution of starburst galaxies in the ew([oii])-ew(h@xmath2 ) plane depends on their prior star formation histories .
the three models with sa ( model 13 ) , sb ( model 1 ) , and sc ( model 12 ) type star formation histories all pass through the e(a ) region during the starburst , which implies that a dusty starburst galaxy can show an e(a ) spectrum irrespective of its longer - term activity in its quiescent state .
figure 9 shows the dependence of the spectroscopic evolution on the parameter @xmath37 controlling the strengths of starbursts in sb type disk galaxies . since parameters other than @xmath37 are the same between these models ( with @xmath37 = 1.0 , 2.25 , and 5.0 gyr@xmath48 , corresponding to models 14 , 1 , and 15 , respectively ) , this figure describes how the strength of a starburst in a galaxy determines the location of the dusty galaxy in the ew([oii])-ew(h@xmath2 ) plane . for the adopted parameter range ( @xmath62gyr@xmath48 ) ,
the models show e(a ) spectra during starbursts , although the location of each model in the e(a ) region is different between these models : the models with larger @xmath37 ( i.e. , those with a stronger starburst ) show stronger ew(h@xmath2 ) and are thus located to the right of the e(a ) region .
the two models with @xmath63 and 5gyr@xmath48 evolve from e(a ) to a+k galaxies whereas the model with @xmath64 , corresponding to a galaxy with a rather weak and longer starburst , evolve from e(a ) to k+a without passing through the a+k region .
this result implies that dusty starburst galaxies that have undergone a rather weak starburst , are more likely to exhibit a k+a spectrum straight after the starburst .
clearly the _ strength _ of the starburst is important to the evolution of a dusty starburst in the ew([oii])-ew(h@xmath2 ) plane .
figure 10 demonstrates how the evolution is dependent on the duration of the starburst .
there is about 1.8gyr difference in the starburst duration between the models , with @xmath30 = 0.90 ( model 16 ) , 0.95 ( model 1 , fiducial model ) , and 0.99 ( model 17 ) .
all three of the models pass through the e(a ) region during the starburst , albeit at different locations .
the models with smaller @xmath30 values ( model 1 and 14 ) corresponding to those with the shorter starburst durations
evolve from e(a ) into a+k and then into k+a types during the starburst and poststarburst phases .
in contrast , the model with the largest @xmath30 value and thus the longest starburst duration , evolves directly from an e(a ) to a k+a type without passing through the a+k spectral phase .
( this dependence on @xmath30 is strikingly similar to that on @xmath37 . )
this result implies that a dusty starburst galaxy with the longer starburst duration is more likely to show the k+a spectra in the poststarburst phase . to summarize the above , weak [ o ii ] equivalent width during the starburst is favoured by large dust obscuration , obscuration which is slow to clear relative to the characteristic age of the starburst .
furthermore , the location and movement within the ew([oii])-ew(h@xmath2 ) plane of a dusty starburst galaxy during both the starburst and poststarburst phases has been demonstrated to be dependent on a number of parameters : @xmath40 , @xmath41 , @xmath27 , @xmath37 , @xmath35 , and @xmath30 . in turn , these parameters could depend on the relative spatial distribution of gas and stars , the radial variation in the metallicity and density of the gas , the dynamical evolution of gas and stars , and the effectiveness of supernovae feedback in the evolution of the interstellar medium for a dusty starburst galaxy .
the one - zone model adopted here is , by its very nature , unable to determine the extent to which these important factors ( e.g. , the relative spatial distribution of gas and stars ) are closely associated with the origin of the six parameters .
this means that although the present study can indicate which of those parameters are important to the spectroscopic evolution of dusty starburst galaxies , it can not so clearly reveal the origin of the spectroscopic properties of the galaxies .
thus future studies , in which the above five parameters are not treated as free parameters but as ones that can be predicted from a certain theoretical calculation are required to understand more clearly the origin of each spectral type ( e.g. , e(a ) ) in the context of distant cluster galaxy evolution .
calzetti et al . ( 1994 ) and mayya & prabhu ( 1996 ) .
have already proposed the idea that the older populations can be located at the outer part of molecular clouds and thus do not suffer so severely from dust extinction .
poggianti & wu ( 2000 ) furthermore pointed out that stellar populations with the ages between a few @xmath8 and 1.5 @xmath65yrs , which are responsible for the strong balmer line absorption , have had sufficient time to drift away from the dusty molecular clouds where they were born .
this selective obscuration of young stellar populations probably works effectively wherever a starburst occurs within molecular clouds .
both shioya & bekki ( 2000 ) and the present study have demonstrated that if younger starburst populations are more heavily obscured by dusty gas than old ones in a galaxy , the galaxy shows an e(a)-type spectrum .
this was first suggested by poggianti & wu ( 2000 ) , who also proposed that if the location and thickness of dust depended on the age of the embedded stellar populations in a starburst galaxy , then the effects it would have on its spectroscopic properties would be quite remarkable .
why is it then that only some fraction of dusty starburst galaxies show e(a ) spectra ?
moreover , when observing them in distant clusters , it should not be forgotten that their spectrum is an integrated one , comprising a luminosity - weighted ` mean ' of the light emitted over a large region of the galaxy .
thus while the selective obscuration suggested by poggianti & wu ( 2000 ) is clearly important , it is the _ global _ ( 1001000pc ) rather than the local ( @xmath66pc ) processes that need to be investigated in in detail to explain the origin of the selective extinction .
we point out here that a difference in the degree of dust extinction between the central ( @xmath15 a few 100 pc ) and outer ( @xmath67kpc ) regions of a galaxy ( @xmath3 1 kpc ) could produce the above age - dependent extinction effect . to be more specific , if a young stellar population is formed in a secondary starburst in the central region where there are greater quantities of dust and hence more dust extinction , its contrast in properties to the older and less obscured populations in the outer regions of the galaxy will make the extinction appear as though it is selective .
this idea that the _ age _ of the stellar population varies globally with radius is not unreasonable and unrealistic , given that secondary starbursts occur preferentially in the central parts of galaxies due to the efficient inward transfer of gas driven by non - axisymmetric structure ( such as stellar bars ) and galaxy interaction and merging ( barnes & hernquist 1992 ; mihos & hernquist 1995 ) .
furthermore , numerical simulations of the gaseous and stellar distribution in merging disk galaxies ( which trigger a dusty starburst ) have demonstrated that the central young stellar component formed by a nuclear starburst , is more heavily obscured by dust than the outer , older component initially located in the merger progenitor disks ( bekki , shioya , & tanaka 1999 ) .
bekki & shioya ( 2000 ) also demonstrated that since younger stellar populations are more centrally concentrated and the gas density is higher in the inner region , the mean gas density around stars ( which is critically important for @xmath0 ) depends strongly on stellar age in such a way that the density around younger populations is higher than that around older populations .
therefore we strongly suggest that galaxies which undergo central dusty starbursts are more likely to show selective dust extinction .
if the suggested radial dependence of stellar age and dust extinction is plausible and realistic , are there physical processes operating in distant rich clusters which have an equally realistic chance of causing it ? physical processes which are possibilities include : tidal galaxy interaction ( icke 1985 ; noguchi 1987 ) , tidal effects of cluster cores ( byrd & valtonen 1990 ) , galaxy ` harassment ' ( moore et al . 1996 ) , minor merging ( mihos & hernquist 1994 ) , unequal - mass merging ( barnes 1996 ; bekki 1998 ) , and major merging ( barnes & hernquist 1991 ) . given the high probability of high speed galaxy encounters in the rich cluster environment , galaxy harassment seems to be the most promising candidate .
however it has the following three disadvantages in explaining the nature of e(a ) galaxies .
firstly , the blue cluster galaxies of modest luminosity , i.e. in the luminosity regime where galaxy harassment is effective , show predominantly e(b) rather than e(a)type spectra ( poggianti et al .
secondly , there is no clear tendency that e(a ) galaxies are preferentially located within the cores of clusters ( dressler et al .
1999 ) , where galaxy harassment is most likely to trigger a starburst .
thirdly , combined spectroscopic and morphological studies of the galaxies in distant clusters ( e.g. couch et al .
1998 ) have not revealed any significant population of starburst objects which has the grossly deformed spiral arm pattern predicted by moore et al .
of course these observations do not necessarily rule out galaxy harassment as a mechanism for forming e(a ) galaxies ; rather , if galaxy harassment is important , then it will produce only low luminosity galaxies located near the cluster core .
the second observation above would also appear to rule out tidal effects within cluster cores as being important to e(a ) formation as well . a rather high fraction ( @xmath3 50 % ) of low@xmath68lirgs and ulirgs considered to be ongoing major mergers
are observed to have e(a ) spectra ( liu & kennicutt 1995 , poggianti & wu 2000 ) .
the rate of galaxy merging is found to increase very steeply with redshift ( @xmath69 ; van dokkum et al .
1999 ) and , furthermore , there is some evidence that this trend is seen amongst e(a ) galaxies ( dressler et al . 1999
; poggianti et al .
on the other hand , the e(a ) galaxies found in the distant cluster samples are predominantly _ disk _ galaxies , which makes it hard to reconcile them with being the products of major mergers , which tend to destroy any disk structure .
we suggest , therefore , that minor and unequal - mass galaxy merging , which is different from major merging in that it can leave disk systems intact , is a more likely mechanism for reproducing the fundamental properties of e(a ) galaxies .
mihos & hernquist ( 1994 ) demonstrated that minor merging can excite non - axisymmetric structure in a late - type disk , trigger nuclear starbursts , and transform the disk into an early - type system such as an sa or s0 galaxy .
furthermore , unequal - mass merging between two disks has been demonstrated to create a disk with structure and kinematics strikingly similar to that of an s0 ( barnes 1996 ) . on that note ,
bekki ( 1998 ) also found that such merging can trigger nuclear starburst , form a very gas - poor disk , and create bulge with characteristics remarkably similar to that of typical s0s .
hence as well as contributing to the formation of dusty nuclear starbursts and thus to the radial dependence of dust extinction , this more minor type of merging may also provide an evolutionary link between late - type disks , e(a ) galaxies , and early - type disks ( s0s ) and thus account for the observed decrease in the cluster s0 fraction with redshift ( e.g. , dressler et al .
1996 ) . an early - type disk galaxy formed by a minor / unequal - mass merger , will experience a poststarburst phase ( with k+a / a+k spectra ) , and should contain a bulge remarkable for its young stellar population and thus bluer colors ( bekki 1998 ) . accordingly , a key observational test for assessing the relative importance of minor / unequal - mass merging in the formation of e(a ) galaxies with selective dust extinction , is to investigate the number fraction of disk galaxies with k+a / a+k spectra and prominent @xmath70 bulges in distant clusters .
we now focus on the origin of the difference between the e(a ) and e(b ) spectral classes : why do some starburst galaxies show e(b ) spectra while others show e(a ) spectra ? we have demonstrated that the spectral type of a dusty starburst galaxy has a basic dependence @xmath40 in that it will exhibit an e(a ) or e(b ) spectrum if @xmath40 is large or small , respectively .
poggianti et al .
( 1999 ) modeled the time evolution of the global colors and luminosities of distant cluster galaxies by using a stellar population synthesis code and found that the e(b ) galaxies were likely to evolve into sub
luminous systems once the starburst had ceased .
they also found that the hii region metallicities of the e(b ) galaxies are significantly lower than those of galaxies in other spectral classes , and that these low metallicities were consistent with them being low - luminosity galaxies .
these observational results suggest that one of the physical parameters that can govern the difference in spectral class is galactic luminosity .
given our result that the spectral class of a dusty starburst galaxy depends on @xmath40 , this would imply that the luminosity and @xmath40 are somehow related .
how can this be ?
one possibility is that for the more luminous galaxies , the thermal feedback of supernovae explosions that occur during starbursts is simply not strong enough to blow off some fraction of the interstellar gas surrounding the young stars and thus to modify the spatial distribution of the gas .
the key factor here is the deeper gravitational potential well of the galaxy which prevents the gas from escaping the galaxy . as a result of this , the mean gaseous density ( which is an important factor in determining @xmath40 ) continues to remain high , as also its dust extinction
consequently , the galaxy can show an e(a ) spectrum . for a less luminous galaxy ,
most of the dusty interstellar gas can be blown away by the thermal feedback effect of the supernovae , as a result of the shallower gravitational potential .
consequently , the density of dusty gas around the starburst populations is reduced as also the degree of dust extinction .
the galaxy thus shows an e(b ) spectrum .
hence we suggest that owing to the difference in the effectiveness of supernovae feedback in galaxies with different luminosities , galactic luminosity can determine the degree of dust extinction and optical spectral type .
another possibility for the origin of e(b ) galaxies is that they are normal star - forming systems but with a significantly lower metallicity . although poggianti et al .
( 1999 ) argued that e(b ) galaxies must be starburst objects because a spiral - like star formation history was not able to produce strong enough [ oii ] emission , viz .
however , this was based on models with solar metallicity and indeed poggianti et al .
acknowledged that the relation between sfr and [ oii ] strength can change with metallicity . in estimating the strength of the [ oii ] flux for a galaxy with a given star formation rate , the h@xmath43 flux is first derived and then used to estimate the [ oii ] strength by adopting a certain [ oii]/h@xmath43 ratio .
accordingly , if less luminous and less metal - rich galaxies have a larger [ oii]/h@xmath43 ratio , then they will have larger [ oii ] fluxes for a given star formation rate .
the large [ oii ] fluxes observed in e(b ) galaxies could therefore be the result of such a metallicity effect rather than due to a high star formation rate .
indeed , kobulnicky et al . ( 1999 ) have found that the [ oii]/h@xmath43 ratio correlates with oxygen abundance ( 12+log(o / h ) ) in such a way that a galaxy with smaller gaseous metallicity is more likely to show a larger [ oii]/h@xmath43 ratio .
this observational result seems to support the above possibility that e(b ) are just normal star - forming galaxies with less metal - rich gas .
if this is the case , we do not need to consider the reason why there are two different dusty starburst populations .
smail et al .
( 1999 ) found that a significant fraction of the k+a population in abell 851 ( @xmath13 ) appeared as sources in deep vla 1.4ghz continuum maps of this cluster , and suggested that even k+a galaxies classified _ optically _ by their lack of [ oii ] emission but moderately strong h@xmath2 absorption can be systems hosting and ongoing dusty starburst .
balogh & morris ( 2000 ) also found examples of k+a galaxies with h@xmath9 emission and , accordingly , concluded that some fraction of k+a galaxies are likely to be undergoing significant star formation .
these observations have raised the following two questions : ( i)are such k+a galaxies really dusty starburst ones ? , and ( ii)what determines the spectral types of dusty starburst galaxies ( or , why are there dusty starburst populations with spectral characteristics , viz .
e(a ) and k+a ) ? concerning the first question , smail et al .
( 1999 ) argued that the radio emission from the k+a galaxies most likely had a starburst origin ( being the synchrotron radiation emitted from type ii supernovae associated with the newly formed population of massive stars ) with star formation rates of the order of 1030@xmath72@xmath73 .
their ` poststarburst ' spectral signature at optical wavelengths would imply that this star formation was completely obscured . although this interpretation is plausible and consistent with other observational results such as color images , submillimeter fluxes , and the [ oii ] emission line strengths for these galaxies ( smail et al . 1999 ; poggianti et al . 1999 ) , we point out that an alternative interpretation is that there are significant spatial variations in star - forming activity across these galaxies , with the central regions in a ` poststarburst ' phase and the outer regions still in a ` starburst ' phase .
the galaxy shows a k+a spectrum at optical wavelengths because the earlier @xmath74 star formation ( which forms the poststarburst population ) occurred more intensively than the @xmath75 ongoing star formation : the integrated spectrum is simply dominated by the light from the central poststarburst region . at radio wavelengths ,
the flux comes exclusively from the outer star - forming region .
the validity of this interpretation can be assessed observationally , if the detailed two - dimensional spatial distribution of say h@xmath9 emission is mapped .
furthermore , we can confirm whether smail et al.s interpretation of these objects is robust by investigating their infrared fluxes and thereby obtaining more accurate star formation rates .
thus it is reasonable to say that future h@xmath9 and infrared observations will resolve this first question . concerning the second problem
, our present numerical results can provide some clues .
our models for starbursts with @xmath41 = @xmath4 and @xmath8 show e(a ) and k+a spectra , respectively , which implies that the difference in @xmath41 yields the spectral type difference
. the larger value of @xmath41 means that the age - dependence of dust extinction is weaker and thus interstellar dust uniformly obscures the starburst populations of different ages .
accordingly , dusty starburst galaxies with k+a spectral types are likely to be formed , if younger stellar populations are not so preferentially obscured by dust ( i.e. , the selective dust extinction does not effectively work ) .
if the difference in @xmath41 is really an important factor in explaining the origin of the e(a)/k+a dichotomy , the next question is what physical processes drive the @xmath41 difference .
answering this question is obviously beyond the scope of this paper , because @xmath41 is a free parameter of the present one - zone model .
however , we suggest that the difference in spatial distribution of stars , gas , and dust between dusty starburst galaxies is responsible for the @xmath41 difference . in our future papers , we will investigate how the radial distribution of stars and gas affects the spectral types of dusty starburst galaxies .
although the present one - zone model of dusty starburst galaxies has clarified some essential ingredients of the spectroscopic evolution of distant cluster galaxies , it has the following three limitations in analyzing the spectroscopic properties of galaxies .
firstly , the degree of dust extinction ( represented by @xmath0 ) , which is one of the most important parameters in determining the seds of galaxies , is a free parameter in the one - zone model .
@xmath0 is generally considered to depend on gaseous density , gaseous metallicity , dust composition , and the relative distribution of gas and stars ( e.g. , mazzei , xu , & de zotti 1992 ) and so is not calculated in principle by the one - zone model .
secondly , the dependence of @xmath0 on the ages of the starburst populations is assumed in an arbitrary way and fixed for the one - zone model .
the age dependence , which is another important factor in this context , depends on the relative spatial distribution of dusty gas and young stars and thus can not be predicted from the one - zone model .
thirdly , the star formation history of a galaxy is freely controlled for the one - zone model .
the time evolution of the star formation rate is suggested to be controlled by dynamical processes such as local gravitational instability in disk galaxies ( e.g. , kennicutt 1989 , 1998 ) ; this is not addressed by the one - zone model since it is unable to follow the dynamical evolution of galaxies . in future work , we suggest that the above three properties ( e.g. , @xmath76 ) in the one - zone model should not be freely parameterized but predicted from theoretical or numerical studies so that spectroscopic properties can be investigated in a more self - consistent manner . the most effective way to remove the above three limitations in the classical one - zone model of spectrophotometric evolution is to derive both the detailed distributions of gas , dust , and stars and the time - evolution of the star formation rate , by performing numerical simulations of dusty starburst galaxies . we have already developed a new code by which we can derive photometric evolution of galaxies based on the detailed spatial distribution of gas and stars and on the time evolution of star formation rate ( bekki & shioya 2000 ) . however , this code can not yet enable us to calculate spectroscopic properties of dusty starburst galaxies due to the time evolution of gaseous emission not yet being included . accordingly , an important future task is to derive emission line properties in numerical simulations . from now on , we refer to these future numerical simulations as ` spectrodynamical simulations ' , by which we can investigate not only spectroscopic properties but also morphological , structural , and kinematical ones .
these will address the following five questions : ( 1)can younger starburst populations actually be obscured more heavily by dust than older ones , and what physical processes are closely associated with the suggested selective dust extinction ? ( 2)what determines the strength of the age - dependence of @xmath0 in dusty starburst galaxies and is the strength time - dependent ?
( 3)how do physical processes specific to cluster galaxy evolution , such as galaxy harassment ( moore et al . 1994 ) and ram - pressure stripping ( farouki & shapiro 1984 ) , determine the relative spatial distribution of dusty gas and young stars ?
( 4)are supernovae feedback effects important for explaining the spectroscopic properties of less luminous galaxies with e(b ) spectra ?
( 5)what is the most important dynamical process for the origin of each of the spectral classes observed in clusters ?
we have investigated the evolution in spectral properties of dusty starburst galaxies , adopting the fundamental assumption that the amount of dust extinction ( quantified by @xmath0 ) depends on the age of the stellar population according to : @xmath77 , where @xmath40 and @xmath41 are parameters controlling the degree of extinction and @xmath19 and @xmath78 represent the age of the galaxy and that of its stellar population , respectively .
the main conclusions from this investigation can be summarized as follows : 1 . if the young stellar populations ( with an age of @xmath79yr ) are more heavily obscured by dust than the older ones ( age@xmath80yr ) in a dusty starburst galaxy , the integrated spectrum of the galaxy is an e(a ) type characterized by strong h@xmath2 absorption and relatively modest [ oii ] emission .
the time interval over which the galaxy exhibits an e(a ) spectrum is fairly long , lasting for @xmath55gyr of the @xmath81gyr duration of the starburst .
2 . there are requisite physical conditions for a dusty starburst galaxy to show an e(a ) spectrum .
for example , the mean @xmath40 should be within a certain range ( @xmath82 ) . the age dependence of dust extinction
also should be fairly strong in a sense that young stars with the ages of @xmath3 @xmath4 yr are preferentially obscured by dust .
( i.e. , @xmath41 @xmath3 @xmath4 yr ; more selective ) .
3 . a dusty starburst galaxy with an e(a )
spectrum will evolve into a poststarburst galaxy with an a+k ( or k+a ) spectrum , 0.2gyr after the starburst has ended , and then into one with a passive k - type spectrum , 1gyr after the starburst .
this result clearly demonstrates an evolutionary link between the following spectral types : e(a ) , a+k , k+a , and k. 4 .
the evolutionary path of a dusty starburst galaxy on the ew([oii])ew(h@xmath2 ) plane depends on the history of star formation within the disk before the starburst , the epoch , duration , and strength of the starburst , plus the adopted age - dependence of @xmath0 .
for example , galaxies of later hubble type ( i.e. with smaller @xmath35 ) are more likely to pass through the bottom - right part of the e(a ) region during a starburst .
5 . for a plausible star formation history ,
a starburst galaxy with the earlier burst epoch is more likely to show slightly lower ew([oii ] ) and slightly higher ew(h@xmath2 ) strengths .
this result implies that a dusty starburst at higher redshift is likely to be located in the bottom - right part of the e(a ) region on the ew([oii])ew(h@xmath2 ) plane .
a dusty starburst galaxy can show an a+k or k+a spectrum even in the starburst phase , if the age dependence of @xmath0 is weak ( @xmath83 ) for a fixed @xmath40 .
this result implies that if young massive stars with different ages ( @xmath57 ) are uniformly obscured by dust , the galaxy shows an a+k or k+a spectrum .
the duration of a starburst can control the spectral type evolution such that a dusty galaxy with a longer starburst time - scale shows a k+a spectrum rather than an e(a ) spectrum during the starburst phase .
this result , combined with the previous point implies that the origin of possible starburst k+a galaxies detected by vla radio observations ( smail et al .
1999 ) can result either from the weak age - dependent dust extinction or from the relatively long starburst time scales in these dusty starburst galaxies .
these results clearly demonstrate that the dependence of @xmath0 on ages of young starburst populations is a crucially important factor for determining spectral types of galaxies .
we furthermore stress that the age dependence can be closely associated with the relative distribution of dusty gas and young stars for dusty starburst galaxies in distant clusters .
we will investigate the importance of the relative spatial distribution of gas and stars in determining spectroscopic properties of galaxies by using spectrodynamical simulations in our future works .
we are grateful to the referee michael dopita for valuable comments , which contribute to improve the present paper . y.s .
thanks the japan society for promotion of science ( jsps ) research fellowships for young scientist .
k.b . acknowledges the large australian research council ( arc ) .
we used the [ oii]/h@xmath43 emission line ratio ( adopting a value of 3.01 throughout this paper ; listed in the table of the pegase paper published by fioc & rocca - volmerange 1997 ) to calculate the [ oii ] emission line strength of a dusty starburst galaxy from the derived h@xmath43 line strength in this work .
it is important for us to confirm whether or not our results , described in the main body of the paper , change if we adopt a [ oii]/h@xmath43 ratio different to that of the pegase model .
we therefore demonstrate here that our results do not depend strongly on the adopted [ oii]/h@xmath43 value , by showing results calculated by the photoionization code cloudy ( ferland et al .
1998 ) . before presenting the results
, we confirm whether our very simple model with the cloudy code can reproduce reasonably well the [ oii]/h@xmath43 values observed for nearby galaxies .
figure 11 shows dereddened [ oii]/h@xmath43 ratios as a function of oxygen abundance ( @xmath84 ) from the observations of kobulnicky et al .
( 1999 ) . in this figure
, we also give our theoretical results based on the cloudy code , with an sed of continuous star formation of 10 myr , an electron density of @xmath85@xmath45 , and an ionizing parameter @xmath86 of @xmath87 , for different metallicity models . as is shown in this figure , the cloudy
based model can successfully reproduce the observational results for galaxies with @xmath84 @xmath15 8.5 .
it should be noted here that the [ oii]/h@xmath43 ratio is less than 3 ( corresponding to the value in the pegase with solar metallicity ) for @xmath88 and has a value of @xmath89 around solar metallicity ( @xmath90 ) .
this result suggests that the [ oii ] emission line strength calculated by pegase is appreciably larger ( by a factor of @xmath91 ) than that of cloudy .
considering the above difference in the [ oii]/h@xmath43 ratio between the two models , we investigate the evolution of a dusty starburst galaxy on the ew([oii])-ew(h@xmath2 ) plane for models 1 and 2 by using the cloudy code with solar - metallicity ( the corresponding [ oii]/h@xmath43 @xmath3 value is 2.08 ) . as is shown in figure 12 ,
even if we use the cloudy code , our key results on e(a ) formation and spectral evolution do not change ( compare figure 4 with figure 12 ) .
however , the two models based on the cloudy code show lower ew([oii ] ) than the pegase - based models during their evolution .
this is a result of the smaller [ oii]/h@xmath43 ratio ( 2.08 ) , which in turn makes ew([oii ] ) smaller . as a natural results of this ,
the model with dust shows an e(c ) rather than an e(b ) spectrum during the actively star - forming phase just prior to the starburst . for all of the present models
, we assumed that the [ oii]/h@xmath43 ratio is fixed at a certain value for all time steps in a model .
this is actually an over - simplification , because the gaseous metallicity ( thus [ oii]/h@xmath43 ratio ) in a galaxy evolves with time as the chemical evolution of the galaxy proceeds .
owing to this simplification , the present models show larger values of ew([oii ] ) in the early evolution phase . in future work
, we will consider the dependence of the [ oii]/h@xmath43 ratio on the gaseous metallicity and thereby construct a more realistic and sophisticated model of the spectral evolution of dusty galaxies .
lastly , we point out one disadvantage of the present model in deriving the [ oii ] line of dusty starburst galaxies .
since we used the pegase code for the calculation of [ oii]/h@xmath43 , the temperature of the hii region in the present model with the pegase code is too low , in particular , for high metallicity , owing to the model s not allowing for any dust depletion onto grains in the hii regions . as a natural result of this , the [ oii ] line strength in the present model
is severely decreased relative to what it should be .
one of the best way to avoid this underestimation of [ oii ] line strength is to use the [ oii]/h@xmath43 derived by dopita et al .
( 2000 ) in which dust physics and the depletion of various elements out of the gas phase are more properly treated .
we have a plan to investigate the [ oii]/h@xmath43 of dusty starburst galaxies in our future papers by using the model by dopita et al .
( 2000 ) .
cccccccc 1 & 0.225 & 7.64 & 2.25 & 0.95 & 8.558 & @xmath92 & 5 + 2 & 0.225 & 7.64 & 2.25 & 0.95 & 8.558 & @xmath92 & 0 + 3 & 0.225 & 7.64 & 2.25 & 0.95 & 8.558 & @xmath92 & 1 + 4 & 0.225 & 7.64 & 2.25 & 0.95 & 8.558 & @xmath92 & 3 + 5 & 0.225 & 7.64 & 2.25 & 0.95 & 8.558 & @xmath92 & 10 + 6 & 0.225 & 7.64 & 2.25 & 0.95 & 8.558 & @xmath93 & 5 + 7 & 0.225 & 7.64 & 2.25 & 0.95 & 8.558 & @xmath94 & 5 + 8 & 0.225 & 4.46 & 2.25 & 0.95 & 5.722 & @xmath92 & 5 + 9 & 0.225 & 6.93 & 2.25 & 0.95 & 7.920 & @xmath92 & 5 + 10 & 0.225 & 8.47 & 2.25 & 0.95 & 9.307 & @xmath92 & 5 + 11 & 0.225 & 9.45 & 2.25 & 0.95&10.193 & @xmath92 & 5 + 12 & 0.056 & 7.64 & 2.25 & 0.95 & 9.189 & @xmath92 & 5 + 13 & 0.325 & 7.64 & 2.25 & 0.95 & 8.242 & @xmath92 & 5 + 14 & 0.225 & 7.64 & 1.00 & 0.95 & 9.870 & @xmath92 & 5 + 15 & 0.225 & 7.64 & 5.00 & 0.95 & 8.030 & @xmath92 & 5 + 16 & 0.225 & 7.64 & 2.25 & 0.90 & 8.148 & @xmath92 & 5 + 17 & 0.225 & 7.64 & 2.25 & 0.99 & 9.921 & @xmath92 & 5 + | by using a one - zone chemical and spectrophotometric evolution model of a disk galaxy undergoing a dusty starburst , we investigate , numerically , the optical spectroscopic properties in order to explore galaxy evolution in distant clusters .
we adopt an assumption that the degree of dust extinction ( represented by @xmath0 ) depends on the ages of starburst populations in such a way that younger stars have larger @xmath0 ( originally referred to as selective dust extinction by poggianti & wu 2000 ) .
in particular , we investigate how the time evolution of the equivalent widths of [ oii]@xmath13727 and h@xmath2 is controlled by the adopted age dependence .
this leads to three main results : ( 1)if a young stellar population ( with the age of @xmath3 @xmath4 yr ) is more heavily obscured by dust than an old one ( @xmath5 @xmath6 yr ) , the galaxy can show an `` e(a ) '' spectrum characterized by strong h@xmath2 absorption and relatively modest [ oii ] emission .
( 2)a dusty starburst galaxy with an e(a ) spectrum can evolve into a poststarburst galaxy with an a+k ( or k+a ) spectrum 0.2gyr after the starburst and then into a passive one with a k - type spectrum 1gyr after the starburst .
this result clearly demonstrates an evolutionary link between galaxies with different spectral classes ( i.e. , e(b ) , e(a ) , a+k , k+a , and k ) .
( 3)a dusty starburst galaxy can show an a+k or k+a spectrum even in the dusty starburst phase if the age - dependence of dust extinction is rather weak , i.e. , if young starburst populations with different ages ( @xmath7 @xmath8 yr ) are uniformly obscured by dust . |
determining the angles of the ckm unitarity triangle , denoted by @xmath15 , @xmath16 and @xmath0 , is one of the important aims of the @xmath1-factories .
methods have been suggested to determine @xmath0 ( @xmath17 arg ( @xmath18 where @xmath19 is the ckm matrix ) using decays of @xmath20 and @xmath21 ( and their cp - conjugates ) to two pseudoscalars belonging to the @xmath13 octet including the effects of the electroweak penguin ( ewp ) diagrams .
these methods rely on the flavor @xmath13 symmetry .
many of these methods neglect the effects of ( soft ) final state rescattering .
in particular , the decay amplitude for @xmath22 is assumed to contain only the weak phase @xmath23 from the penguin diagram with the top quark in the loop .
tree level operators have the weak phase @xmath24 , but since they have the transition @xmath25 , they contribute , in the absence of rescattering , only through annihilation to this decay .
annihilation contributions are argued to be small since they are suppressed by @xmath26 ( in the absence of significant rescattering effects ) .
assuming that the @xmath22 amplitude has no @xmath24 weak phase , references @xcite have suggested methods to determine @xmath0 .
however , rescattering from an intermediate state , for example @xmath27 , created by a ( color - allowed ) spectator decay of a @xmath28 due to tree level operators , can generate a significant amplitude with the weak phase @xmath24 in the decay @xmath29 as well .
so it would be better to have a method to determine @xmath0 which does not use the assumption of no rescattering effects .
rescattering might also enhance annihilation contributions @xcite and thus necessitates their inclusion .
buras and fleischer @xcite gave a method to determine @xmath0 _ without _ neglecting rescattering using @xmath30 , @xmath31 decays and time _ dependent _ measurements of the @xmath32 decay . for this method
, they also require time dependent analysis of , for example , @xmath33 to measure @xmath16 .
gronau and pirjol @xcite suggested a method using time independent measurements of _ all _ the @xmath34 and @xmath35 modes . in their method also rescattering effects are included .
however , it might be diffcult to measure the neutral modes of @xmath21 decays since that will involve tagging at hadron machines . in this paper , we suggest a technique to determine @xmath0 _ including _ rescattering effects ( and the ewp operators ) using @xmath1 meson decays to @xmath2 s and @xmath3 s .
we will illustrate this technique using four cases ; see table [ cases ] . &
+ & + & & + case & @xmath7 & @xmath36 + & & + & & + @xmath6 & @xmath37 & @xmath38 + & & + & @xmath39 & @xmath40 + & & + & & + & & + @xmath41 & @xmath42 & @xmath43 + & & + & @xmath39 & ( cp - averaged ) + & & + & & + & & + @xmath44 & @xmath45 & @xmath46 + & & + & @xmath47 & @xmath48 + & & + & & + & & + @xmath8 & @xmath49 & @xmath50 + & & + & @xmath51 ( cp - averaged ) & @xmath48 + & & + & & + we do _ not _ require any time dependent studies .
the strategy is as follows . in cases
[ intro ] and [ bpipi ] , using @xmath4 @xmath5 decay modes , we determine the strong phases and magnitudes of the tree level and penguin contributions as functions of @xmath0 ( assuming flavor @xmath52 symmetry ) .
then , using flavor @xmath13 symmetry , we _ predict _ the rate for _ one _ @xmath53 mode in case [ bpipi ] . in case
[ intro ] , two @xmath54 modes have to be measured to make a prediction for a third @xmath54 mode .
the measurement of the decay for which we have a prediction ( as a function of @xmath0 ) then determines @xmath0 .
a similar idea can be applied to predict a @xmath5 decay mode as a function of @xmath0 using measurements of @xmath10 ( and some @xmath7 ) modes ( cases [ bpik ] and [ conclude ] ) .
the @xmath49 , @xmath55 modes should be relatively accessible at the @xmath14 and hadron machines . for
the @xmath56 decays to a cp eigenstate , we require ( external ) tagging ( _ i.e. , _ the cp - averaged decay rate is not sufficient ) .
the @xmath21 decay modes were accessible at lep1 and will be accessible at hadron machines .
the @xmath57 decay mode ( case [ intro ] ) might be hard to measure since it requies tagging whereas in cases [ bpipi ] and [ conclude ] , the @xmath21 modes are either `` self - tagging '' ( @xmath58 ) or a cp - averaged measurement is sufficient ( for @xmath59 , @xmath60 ) . in case [ intro ] , we show that if we measure additional @xmath56 modes , a cp - averaged measurement of the decay rate @xmath57 is sufficient .
although flavor @xmath13 symmetry is used in all the four cases ( as in all the other methods mentioned above ) , in the last section , we discuss how to take into account @xmath13 breaking .
we will write the decay amplitudes for the decays @xmath61 , where @xmath62 is a pseudoscalar belonging to the flavor @xmath13 octet , in terms of the 5 linearly independent @xmath13 invariant amplitudes denoted by @xmath63 , @xmath64 , @xmath65 , @xmath66 and @xmath67 ( where @xmath68 and @xmath69 stand for the parts of these amplitudes generated by tree level and penguin operators , respectively ) .
these @xmath13 invariant amplitudes include rescattering effects .
the annihilation amplitudes , @xmath70 , are the ones in which the quark index @xmath71 of @xmath72 is contracted directly with the hamiltonian .
neglecting rescattering effects is equivalent to assuming @xmath73 and @xmath74 .
for example , the tree level part of the decay amplitude @xmath75 contains this combination of the @xmath76 amplitudes and @xmath77 . in this notation @xcite
, the amplitudes for @xmath78 decays can be written as @xmath79 @xmath80 @xmath81 here , @xmath82 ( @xmath83 and @xmath84 ) and @xmath85 , @xmath86 denote the penguin amplitudes due to @xmath87 running in the loop
. can also contribute through rescattering from charm intermediate states .
this rescattering generates a charm - quark penguin topology with the amplitude being proportional to @xmath88 and so can be included as part of @xmath89 . ]
@xmath90 and @xmath91 are the amplitudes for @xmath92 and @xmath93 respectively or in other words the @xmath94 and @xmath95 amplitudes . using the unitarity of the ckm matrix , _
@xmath96 , we get @xmath97 where @xmath98 and @xmath99 .
a similar notation is used for @xmath100 and @xmath101 . in the @xmath49 decay , which contains only the @xmath94 amplitude
, there is no contribution to @xmath102 from the strong penguin diagrams since these diagrams are @xmath95 .
neubert and rosner @xcite showed that @xmath103 , where @xmath104 is the ratio of wilson coefficients ( wc s ) of the ewp operators ( with quark @xmath87 running in the loop ) and the tree level operators in the effective hamiltonian .
we expect @xmath105 since the top quark ewp diagram with @xmath106 exchange is enhanced by @xmath107 and so henceforth we neglect @xmath108 and denote @xmath109 by @xmath110 .
so , we get @xmath111 @xmath112 using eqns.([notation ] ) , ( [ nrresult0 ] ) and ( [ nrresult ] ) , we can rewrite @xmath113 + @xmath114 as @xmath115 since @xmath116 and @xmath117 , we neglect the second term ( _ i.e. _ , the ewp contribution ) in the right hand side of eqn.([a2 ] ) for now and we assume @xmath118 and @xmath119 ( and similarly for @xmath120 replaced by @xmath121 ) , we get @xmath122 so that @xmath123 can be obtained directly from the @xmath49 decay rate . we have chosen a phase convention such that the strong phase of @xmath124 is zero . from eqns.([b0 ] ) and ( [ b+- ] ) , we get @xmath125 the five quantities : @xmath123 , @xmath126 , @xmath69 , @xmath127 and @xmath128 ( where the @xmath129 s are the cp conserving strong phases ) can thus be determined as functions of @xmath0 from the measurements of the five rates : @xmath49 , @xmath130 , @xmath131 and the cp - conjugates of the @xmath56 decays . explicitly , rotating the cp - conjugate amplitudes by @xmath132 ( and denoting them by `` bars '' ) , we get the triangle formed by @xmath49 , @xmath130 and @xmath131 ( eqns.([b+ ] ) , ( [ b0 ] ) and ( [ b+- ] ) ) : @xmath133 and the one formed by the cp - conjugate decays .
these are shown in fig.[figbpipi ] ( from eqn.([approxa2 ] ) , @xmath134 ) . from eqn.([a0 ] ) , we get @xmath135 thus , the length of @xmath136 ( obtained from fig.[figbpipi ] ) gives @xmath69 as a function of @xmath0 .
the angle between @xmath136 and @xmath90 is @xmath137 ( see eqns.([approxa2 ] ) and ( [ idiff1 ] ) ) so that the orientation of @xmath136 ( obtained from fig.[figbpipi ] ) gives @xmath138 independent of @xmath0 .
similarly , @xmath139 gives @xmath127 and @xmath126 ( or knowing @xmath140 and @xmath128 gives @xmath127 and @xmath126 using eqn.([a0 ] ) ) .
= 0.8 [ figbpipi ] there is a discrete ambiguity in this procedure since in fig.[figbpipi ] , the vertices @xmath141 and @xmath142 could be on opposite sides of @xmath143 .
this has been discussed in the literature @xcite @xcite .
all the analysis up to now actually relies only on flavor @xmath52 symmetry .
the @xmath144 and @xmath57 amplitudes are given by @xcite @xmath145 @xmath146 @xmath147 from eqns.([bk+pi- ] ) and ( [ bspipi ] ) and using the notation of eqn.([notation ] ) and rearranging we get @xmath148 using eqns.([nrresult ] ) and ( [ a0 ] ) and assuming @xmath149 , we get @xmath150 _ i.e. , _ from eqns.([b+- ] ) and ( [ bkpibspipi ] ) , we see that the combination of the amplitudes @xmath151 can be obtained from the amplitude for @xmath152 by scaling the tree level contribution in the latter by @xmath153 and the penguin contribution by @xmath154 .
in particular the ewp contribution ( @xmath155 ) is important in the last line of eqn.([bkpibspipi1 ] ) since , due to the ckm factors , it is comparable to the tree level contribution ( unlike in the @xmath152 decay ; see eqn.([a2 ] ) ) . similarly , from eqns.([b0 ] ) , ( [ bk0pi0 ] ) and ( [ bspipi ] ) , we see that the combination of the amplitudes @xmath156 can be obtained from the amplitude @xmath157 by scaling the latter by ckm factors .
this gives @xmath158 we emphasize again that in obtaining the last lines of eqns.([bkpibspipi1 ] ) and ( [ bkpibspipi2 ] ) , it is crucial that we use the neubert - rosner result ( eqn.([nrresult ] ) ) , _
i.e. , _ that @xmath110 is calculable .
if @xmath159 , we can determine @xmath160 from the decay rate @xmath161 .
then , measuring the decay rate @xmath162 gives @xmath163 as a function of @xmath0 ( using eqn.([bkpibspipi1 ] ) ) ( @xmath138 , @xmath164 , @xmath69 and @xmath126 are already known as functions of @xmath0 ) .
knowing @xmath160 and @xmath163 , we have a prediction for the decay rate @xmath165 ( eqn.([bkpibspipi2 ] ) ) and then @xmath0 can be determined by measuring this decay . and another in the final determination of @xmath0 from the @xmath165 rate . ]
thus , we can determine @xmath0 , including rescattering effects , by measuring the @xmath166 decay modes : @xmath49 , @xmath56 and @xmath167 , @xmath162 , @xmath165 and @xmath57 ( any one ) ( or cp - conjugates of the last three modes ) . from eqn.([bspipi ] ) , we see that the decay mode @xmath168 has only an annihilation contribution .
if ( either from experimental measurement of the @xmath57 rate ( or a limit on the rate ) or a theoretical estimate including rescattering ) the annihilation amplitude @xmath169 does turn out to be small ( smaller than , say , the experimental error in the measurement of the ( magnitude ) of @xmath170 ) then , a decay rate @xmath171 can be predicted ( as a function of @xmath0 ) by simply scaling the corresponding @xmath172 amplitude by ckm factors .
thus , in this case , @xmath173 decay modes : @xmath49 , @xmath56 ( and @xmath174 ) @xmath175 and any _ one _ @xmath176 are sufficient to determine @xmath0 . as mentioned in the introduction , we require tagging to measure the @xmath168 decay mode ( _ i.e. _ , the cp - averaged rate is not sufficient in the above method )
. if this tagged decay rate is hard to measure whereas the cp - averaged rate can be measured , we can proceed as follows .
writing @xmath177 , the cp - averaged rate is @xmath178 .
measuring the @xmath176 decay rates , from eqns.([bkpibspipi1 ] ) and ( [ bkpibspipi2 ] ) , we can determine @xmath160 ( and @xmath179 ) as a function of @xmath0 .
similarly , the @xmath180 decay rates will give @xmath181 .
thus , the cp - averaged @xmath168 rate can be predicted as a function of @xmath0 and its measurement gives @xmath0 .
the expression for the @xmath182 decay amplitude in terms of the @xmath13 invariant amplitudes is identical to that for @xmath183 , _ including _ annihilation contributions ( unlike the decay mode @xmath162 ) , modulo ckm factors , _
i.e. , _ up to @xmath184 @xcite .
thus , we also have a prediction ( as a function of @xmath0 ) for this decay rate and also the rate for its cp - conjugate process including _ all _ rescattering effects .
so , the measurement of this cp - averaged decay rate can be used to determine @xmath0 .
using the same technique as in section [ bpipi ] , we can _ predict _ the rate for a @xmath172 ( or a @xmath35 ) decay as a function of @xmath0 _ given _ the decay rates for @xmath185 .
the difference is that in the triangle construction to determine the tree level and penguin contributions to the @xmath176 amplitudes , we have to take into account the ewp contributions to the @xmath185 amplitudes ( the ewp amplitudes were neglected in the @xmath78 triangles in section [ bpipi ] ; see eqns.([a2 ] ) and ( [ approxa2 ] ) ) . , to make a _ prediction _ for @xmath54 decay modes , we did have to include the ewp contribution to the @xmath54 decay amplitudes .
] so , we discuss the application of the technique again .
the decay amplitudes for @xmath186 ( eqns.([bk+pi- ] ) and ( [ bk0pi0 ] ) ) can be written as @xmath187 @xmath188 where @xmath189 and @xmath190 are the amplitudes for @xmath56 decay to @xmath191 and @xmath192 respectively .
then , @xmath193 using eqn.([nrresult ] ) and latexmath:[$| \tilde{c } _ { 15}^t @xmath49 decay ) .
@xmath195 is given by @xmath196 , _
i.e. , _ as mentioned earlier , the ewp contribution _ is _ important for @xmath176 decays .
@xmath123 can be obtained from the @xmath49 decay rate as before .
@xmath189 is given by ( in analogy to @xmath91 of section [ bpipi ] ) @xmath197 as in section [ bpipi ] , the four quantities : @xmath198 , @xmath199 , @xmath200 and @xmath201 can thus be determined as functions of @xmath0 from the measurements of the four decay rates : @xmath162 , @xmath202 and their cp - conjugates . decay amplitudes which can be written in terms of @xmath190 ( the same as for @xmath176 decays ) and @xmath203 which is a different combination of the @xmath13 invariant amplitudes than @xmath189 .
thus , the @xmath204 decay rates are not useful as far as predicting the @xmath205 ( case [ bpik ] ) ( or @xmath35 in case [ conclude ] ) decay rates is concerned . ]
due to the ewp contribution ( see eqn.([a3/2 ] ) ) , the triangle construction is a bit different in this case as shown below . as before ,
we multiply the cp - conjugate amplitudes by @xmath132 to get the `` barred '' amplitudes . in this case
( unlike the case for @xmath90 in section [ bpipi ] ) there is an angle between @xmath190 and @xmath206 denoted by @xmath207 and their magnitudes are functions of @xmath0 ( see eqn.([a3/2 ] ) ) : @xmath208 @xmath209 given @xmath0 , we can thus construct the triangles of eqn.([a3/2 ] ) and its cp - conjugate ( see fig.[figbpik ] ) .
, there is a discrete ambiguity in the orientation of the triangles . ] as in section [ bpipi ] , knowing the magnitudes and orientations of @xmath189 and @xmath210 from fig.[figbpik ] , we can determine @xmath198 , @xmath199 , @xmath211 and @xmath201 as functions of @xmath0 , using equations similar to eqns.([idiff1 ] ) and ( [ idiff2 ] ) .
this construction also shows how to include the ewp contributions to the @xmath49 decay in section [ bpipi ] as follows .
once ewp s are included , as for the case of @xmath190 and @xmath206 , there is an angle @xmath212 between @xmath143 and @xmath213 and the magnitude of @xmath214 will depend on @xmath0 ( see eqn.([a2 ] ) ) : @xmath215 @xmath216 where @xmath217 is given by @xmath218 .
= 0.8 [ figbpik ] the @xmath47 amplitude is given by @xcite : @xmath219 from eqns.([b0 ] ) , ( [ b+- ] ) , ( [ bk+pi- ] ) , ( [ bk0pi0 ] ) and ( [ bkk ] ) , we can see that @xmath220 can be obtained from @xmath221 and + @xmath222 can be obtained from @xmath223 by scaling the @xmath54 amplitudes by appropriate ckm factors , _
i.e. , _ in this case , the decay mode @xmath224 plays the role of the decay mode @xmath168 of case [ intro ] ( compare eqns.([bspipi ] ) and ( [ bkk ] ) ) .
thus , as in case [ intro ] , we can determine @xmath0 , including _ all _ rescattering effects , by measuring the @xmath166 decay modes : @xmath49 , @xmath56 and @xmath180 ( all ) , @xmath47 , @xmath131 and @xmath183 ( or cp - conjugates of the last three modes ) .
if the annihilation amplitudes are small , as in case [ intro ] , we can determine @xmath0 by measuring any _ one _ @xmath172 decay mode , in addition to the @xmath49 , @xmath56 ( and @xmath174 ) @xmath225 decay modes .
as in case [ intro ] , if we measure the cp - conjugate @xmath172 rates as well , then a cp - averaged measurement of the decay rate @xmath47 suffices .
the expressions for the decay amplitudes for @xmath226 and @xmath227 in terms of the @xmath13 invariant amplitudes are identical to those for @xmath162 and @xmath202 , respectively , _ including _ annihilation contributions ( unlike the case of @xmath172 and @xmath176 decays ) , modulo the ckm factors @xcite .
thus , the same method predicts the rates for the @xmath228 decays and the cp - conjugate processes .
it suffices to use the @xmath229 decay ( or its cp - conjugate ) which is a `` self tagging '' mode or the cp - averaged decay rate for @xmath227 .
thus , no external tagging is required .
the analysis of the above three cases is strictly valid only in the flavor @xmath13 limit . in the tree level amplitudes , _
i.e. , _ @xmath124 and the @xmath126 s , the corrections due to @xmath13 breaking can be taken into account more reliably in the factorization approximation and are expected to be given by @xmath230 ( see , for example , gronau _ et al . _ in @xcite @xcite ) .
for example , in fig.[figbpik ] the lengths of @xmath231 , @xmath232 and @xmath233 will have to be multiplied by @xmath230 .
however , since the strong penguin amplitudes include @xmath234 type operators , the ( factorizable ) corrections due to @xmath13 breaking there are less certain , but the corrections are still less than @xmath235 .
this is especially relevant for the cases [ intro ] and [ bpipi ] where the penguin contribution dominates in the @xmath176 , @xmath236 decays and we are _ predicting _ this contribution from the @xmath172 decays using @xmath13 symmetry . in the cases [ bpik ] and [ conclude ] , we use @xmath13 symmetry to predict the penguin contribution in a @xmath237 ( or a @xmath35 ) decay from the @xmath176 decays , but now the _ tree level _
contributions dominate the @xmath172 decay rate and so the uncertainty due to the @xmath13 breaking in the penguin amplitudes is less important .
we have a prediction for more than one rate in some of the cases .
for example , in case [ conclude ] , we can predict both @xmath226 and its cp - conjugate decay rate or in case [ intro ] , neglecting annihilation , we can predict the decay rates @xmath162 and @xmath238 .
so , we can treat the @xmath13 breaking in the penguin amplitudes as an unknown and determine it ( in addition to @xmath0 ) from the measurement of _ seven _ decay rates .
we have also assumed that the @xmath13 breaking in the strong phases is small .
a possible justification is that at the energies of the final state particles @xmath239 , the phase shifts are not expected to be sensitive to the @xmath13 breaking given by , say , @xmath240 ( which is much smaller than the final state momenta ) .
however , it is hard to quantify this effect .
if we measure _ all _ the @xmath78 and @xmath176 decay rates , then we can compute the tree level parts of the amplitudes , both @xmath126 and @xmath198 ( see eqns.([a0 ] ) and ( [ a1/2 ] ) ) , as functions of @xmath0 as discussed in sections [ bpipi ] and [ bpik ] . if the annihilation amplitudes are small , then we have @xmath241 ( in the @xmath13 limit ) since the decay amplitudes @xmath172 and @xmath176 are the same up to ckm factors . to include the @xmath13 breaking in this analysis
, we use the modified relations @xmath242 and @xmath243 .
this can be used to determine @xmath0 , including @xmath13 breaking ( without having to to deal with @xmath13 breaking in the penguin amplitudes and in the strong phases ) .
m. gronau , o.f .
hernndez , d. london , j.l .
rosner , phys .
rev . * d52 * , 6374 ( 1995 ) .
deshpande , x - g .
he , phys .
lett . * 75 * , 3064 ( 1995 ) .
m. gronau , j.l .
rosner , phys .
rev . * d53 * , 2516 ( 1996 ) .
m. neubert , j.l .
rosner , phys .
lett . * 81 * , 5076 ( 1998 ) .
see , for example , m. neubert , phys .
lett . * b424 * , 152 ( 1998 ) ; m. gronau , j.l .
rosner , phys .
rev . * d58 * , 113005 ( 1998 ) .
buras , r. fleischer , hep - ph/9810260 .
m. gronau , d. pirjol , hep - ph/9811335 .
see , for example , a.s .
dighe , phys .
* d54 * , 2067 ( 1996 ) .
m. neubert , j.l .
rosner , phys .
b441 * , 403 ( 1998 ) .
see , for example , h.j .
lipkin , y. nir , h.r .
quinn , a.e .
snyder , phys . rev .
* d44 * , 1454 ( 1991 ) ; m. gronau , o.f .
hernndez , d. london , j.l .
rosner , phys .
rev . * d50 * , 4529 ( 1994 ) .
m. gronau , d. london , phys .
65 * , 3381 ( 1990 ) . | we suggest a new technique to determine the ckm phase @xmath0 _ without _ neglecting the ( soft ) final state rescattering effects .
we use ( time integrated ) @xmath1 meson decay rates to @xmath2 s and @xmath3 s .
a set of @xmath4 @xmath5 ( or @xmath6 @xmath7 and @xmath8 @xmath9 ) decay rates is used to compute the strong phases and magnitudes of the tree level and penguin contributions as functions of @xmath0 .
these are used to _
predict _ a @xmath10 @xmath11 @xmath12 decay rate as a function of @xmath0 ( using @xmath13 symmetry ) .
the measurement of this decay rate then gives @xmath0 .
we illustrate this technique using different cases .
most of the decay modes we use are expected to be accessible at the @xmath1-factories ( @xmath14 or hadron machines ) .
epsf oits-667 + .05 in * determining the weak phase @xmath0 in the presence of rescattering * .15 in k. agashe and n.g .
deshpande .1 in _ institute of theoretical science + 5203 university of oregon + eugene or 97403 - 5203 _ .05 in |
a microlensing event is said to occur whenever a source star and lens star pass each other at an angular separation that places the source within an einstein ring radius of the lens .
the gravitational focusing of the source light by the lens leads to the observed amplification of the source brightness .
the time scale of the event is defined as @xmath3 , where @xmath4 is the einstein ring radius with @xmath5 and @xmath6 being observer - lens and observer - source distances respectively , @xmath7 the mass of the lens , @xmath8 the gravitational constant , c the velocity of light , and @xmath9 is the relative transverse velocity between lens and source projected onto the lens plane .
empirically , the time scale is approximately half of the duration of an event . to illustrate the effect of a population of brown dwarfs on the microlensing time scale frequency distribution
, we adopt the model of zhao ( 1996 ) for the mass distribution in a bar - like galactic bulge , @xmath10 , \label{eq : bulge}\ ] ] where @xmath11 ^ 2+\left(\frac{z^{{\scriptscriptstyle}\prime}}{\sigma_{{\scriptscriptstyle}z}}\right)^4,\cr s_{{\scriptscriptstyle}a}^2&=&\frac{q_{{\scriptscriptstyle}a}^2(x^2+y^2)+z^{{{\scriptscriptstyle}\prime}2}}{\sigma_{{\scriptscriptstyle}z}^2 } \label{eq : blgexp } \end{aligned}\ ] ] with @xmath12 , @xmath13 kpc , @xmath14 kpc and @xmath15 kpc . here
@xmath16 are orthogonal coordinates along the principal axes of the triaxial bar .
the coefficient @xmath17 is chosen such that the mass of the bulge is @xmath18 ( zhao 1996 ) . to this
we add the exponential disk model of bahcall & soneira ( 1980 ) @xmath19 where @xmath20 is the coordinate perpendicular to the plane of the galaxy and r is the radial coordinate in the plane of the galaxy .
we choose @xmath21 kpc for the scale length in the radial direction ( zhao , spergel & rich 1995 ; kent , dame & fasio 1995 ) instead of 3.5 kpc chosen by bahcall & soneira .
the coefficient @xmath22 is chosen such that @xmath23 at @xmath24 and @xmath25 kpc , the sun s distance from the galactic center .
the observations are consistent with the long axis of the bulge inclined about @xmath26 to @xmath27 relative to the line of sight to the galactic center with the near side of the bar lying in the first quadrant ( zhao 1998 ) .
an inclination of @xmath26 is chosen in the following .
we normalize all distances by @xmath28 kpc , all velocity components by @xmath29 km / sec , the circular velocity of the local standard of rest ( lsr ) , the masses by the solar mass @xmath30 and the timescale @xmath31 by @xmath32 days , which is the timescale for an event with a solar mass lens located at a distance of 4 kpc with a source at rest at the galactic center ( 8kpc ) . for @xmath33 , the lens velocity and observer velocity are both assumed to be the circular velocity @xmath34 km / sec in the azimuthal direction consistent with a flat galaxy rotation curve . with these normalizations ,
the number of events @xmath35 at timescale @xmath36 is ( peale 1998 ) @xmath37 where @xmath31 appears only in the expression for @xmath7 in @xmath38 with @xmath39 in eqs .
( [ eq : mcarlo ] ) and ( [ eq : m ] ) @xmath40 are the components of the relative transverse velocity @xmath9 in the directions of increasing galactic latitude and longitude respectively , @xmath41 is the lens mass density , where the numerical coefficients in eq .
( [ eq : mcarlo ] ) mean it is expressed in @xmath42 , @xmath43 , @xmath44 .
the distributions of random velocities about the circular velocity of the galactic model are @xmath45 , where @xmath46 is the rms value of the random velocity @xmath47 and the subscript @xmath48 is either @xmath49 or @xmath50 for lens or source and the subscript @xmath51 is @xmath52 or @xmath53 for the latitude and longitude directions .
the spatial density of sources is @xmath54 , and the exponent @xmath55 follows from the increase in area of the cross section of the angular field with distance and from the assumption that the fraction of stars with luminosities greater than some @xmath56 varies as @xmath57 ( kiraga & paczyski 1994 ) .
the details of the derivation of eq .
( [ eq : mcarlo ] ) are given in peale ( 1998 ) .
the expression for the rate of events for a single source at @xmath6 that has random velocity components within @xmath58 of ( @xmath59 ) for lenses within @xmath60 of @xmath5 with random velocities within @xmath61 of ( @xmath62 ) , and with lens mass within @xmath63 of @xmath7 is multiplied by the fraction of sources within @xmath64 of @xmath6 and by @xmath65 to select a particular time scale .
independent variable @xmath7 is replaced by @xmath31 and the expression integrated over the ranges of all the variables as a function of @xmath66 to yield the time scale frequency distribution of eq .
( [ eq : mcarlo ] ) .
the integral limits correspond to the variables in the same order as the differentials in the integrand , where the integration over @xmath67 starts at 0.1 instead of zero to avoid a singularity .
this is justified as there are not likely to be any sources involved in microlensing events closer to us than 0.8 kpc .
the distributions of lenses and sources along a particular line of sight are identical and follow from the above galactic model .
the assumed circular velocities for the general galactic rotation are @xmath68 km / sec for @xmath69 ( @xmath70 kpc ) consistent with a flat rotation curve except the circular velocities decrease linearly to 0 at the galactic center for @xmath71 ( @xmath72 kpc ) .
the velocity dispersions in the disk are @xmath73 km / sec ( @xmath69 ) , whereas the bulge velocity dispersions are @xmath74 km / sec ( @xmath71 ) .
the velocity dispersions @xmath75 are assigned to disk or bulge dispersions in eq .
( [ eq : mcarlo ] ) according to the locations of the lens and source . evaluation of the integral with a monte carlo technique is described in peale ( 1998 ) .
estimating the mass function for brown dwarfs is a complicated process , since , unlike main sequence stars , there is no unique correspondence between mass and spectral type .
the spectral type l dwarfs , observations of which are the basis of the reid _ et al . _
( 1999 ) discussion , have effective temperatures between 1400 and 2000 k with no @xmath76 absorption , but every brown dwarf with @xmath77 passes through the l dwarf region as it cools , with those of higher mass passing through later ( burrows , _ et al .
the number of l - type brown dwarfs at any epoch depends on the birth rate , the initial mass function , and the evolution of these failed stars ( reid , _ et al .
if the birth rate is assumed known and a theory of evolution adopted , the dependence of the number of visible l dwarfs on the mass function can be used to constrain the latter .
if a power law distribution of mass is assumed , the steeper distributions lead to more visible l dwarfs , and the index is constrained by the observed volume density of l dwarfs in the region reasonably close ( @xmath78 pc ) to the sun .
this procedure leads to @xmath79 for @xmath80 based on information from the denis and 2mass surveys ( reid _ et al .
although model dependent , this result is the first determination of the brown dwarf mass function based on observational data .
what are the implications for the microlensing time scale frequency distribution ?
several star counts yield @xmath81 for @xmath82 with steeper slopes for more massive stars ( e.g. basu & rana 1992 ; holtzman _ et al .
_ 1998 ; reid _ et al .
we shall use the holtzman _
mass functions and modifications thereof in the examples .
figure 1 shows the time scale frequency distributions for the above galactic model for a line of sight toward @xmath83 , which is representative of an average ( see peale 1998 ) over the directions to 24 fields observed by the macho ( massive compact halo objects ) group during the 1993 bulge season . during this season
12.6 million stars were monitored for 189 days ( alcock _ et al .
_ 1996 ) leading to the choice of ordinate in figure 1 .
the distributions are shown for the holtzman _
( 1998 ) mass function ( solid line ) and three modifications , two of which include brown dwarfs down to @xmath84 to accommodate the new discoveries . for continuation of the main sequence mass function into the brown dwarf region with slope -1 ,
the mass of the brown dwarfs is only 5.5% of that of the main sequence stars , and it is about 7% of the main sequence star mass with the reid _ et al . _
slope of -1.3 , where continuity of the mass function is assumed .
these brown dwarf masses are 8% and 10% , respectively , of those main sequence stars with @xmath85a range more appropriate to the solar neighborhood .
the third modification contains no brown dwarfs but has many more low mass , hydrogen burning stars , and it yields a time scale frequency distribution closely matching the empirical data .
( the peaks of these curves are raised by about 10% if the value of @xmath86 km / sec found by mendez _
( 1999 ) is appropriate for the flat part of the galaxy rotation curve . )
all the mass functions are continuous and are normalized to yield a bulge mass of @xmath87 and a mass density at the sun s position in the disk of @xmath88 .
the empirical time scale frequency distribution from the macho 1993 bulge season ( alcock _ et al .
_ 1996 ) , adjusted for observing efficiencies and binned into 5 day intervals , is shown for comparison . as found by several authors ( e.g. peale 1998 ) , mass functions with @xmath81 in the m star region fail to yield a sufficient number of short time scale events near the peak of the empirical distribution _ for the simplest galactic models consistent with observations_. adding the reid _
_ brown dwarfs to the holtzman _ et al
. _ distribution with either the -1.3 or -1 slope increases the number of short time scale events as expected , but both extensions of the mass function fail to yield a sufficient number of short time scale events near the peak of the alcock _
results , and , perhaps more importantly , the peaks of the predicted distributions are shifted toward smaller time scales ( 3 - 5 days ) than those of the observed peak ( 5 - 10 days ) .
( see peale ( 1998 ) for similar consequences for other extensions of the mass function into the brown dwarf region . )
a fairly good match to the empirical data is obtained in figure 1 by increasing the fraction of low mass , but still hydrogen burning stars with a continuation of the -2.2 slope of the holtzman _
_ mass function down to @xmath89 .
the chosen model of the galaxy used in calculating the predicted time scale frequency distributions is consistent with observations ( zhao 1996 , 1998 ) , although many parameters are poorly constrained ( zhao 1998 ) .
if the galaxy model is assumed to be close to the real galaxy , and the local population of brown dwarfs extends throughout the galaxy , then the macho collaboration should have seen a larger fraction of events with 3 - 5 day time scales .
one could conclude that the absence of these events means that there are fewer brown dwarfs in the bulge , where most of the lenses reside , than observed locally .
however , the same caveats and reservations about microlensing constraints on the mass function emphasized by peale ( 1998 ) apply here .
the meagerness of the published microlensing data for bulge events precludes definitive microlensing constraints on the distribution of stars in the galaxy or the mass function , although there are more than 200 events that have not yet been analyzed that would improve this situation .
we also point out that the 10 events in the bin corresponding to time scales from 5 to 10 days in figure 1 involve two clump giant sources with detection efficiencies of about 0.6 and eight much less luminous turnoff and possibly more blended stars as sources with detection efficiencies of 0.15 to 0.2 ( alcock _ et al .
the average of the reciprocals of the efficiencies is nearly 5 meaning this element of the histogram is almost 5 times higher than the actual number of detected events .
the sampling efficiencies for a given time scale event for a distribution of source brightnesses are determined empirically and should be reasonably sound , but the reduction of these efficiencies due to blending of the source stars with other nearby stars by a factor of 0.75 is just an estimate ( alcock et al 1996 ) . until the detection efficiencies for bulge events are determined more rigorously , even the empirical distribution of macho time scales that we are trying to match is only a first approximation .
could the detection efficiency for the 3 - 5 day events be even smaller than the alcock _
_ estimate ?
if so , this element of the histogram would be raised , although it seems unlikely that this region could become the peak of the distribution and thereby accommodate the brown dwarfs .
if the empirical distribution approximates reality , the brown dwarfs might be accommodated by reducing the velocity dispersion ( see examples in peale 1998 ) , but unrealistically low dispersions would be necessary to shift the peak to the macho value , and increasing the mass of the galaxy to yield the observed peak frequency of events would result in too many events with @xmath90 days . finally , one notices the several very long time scale events in the macho distribution that are not consistent with any of the models .
these events may be due to there being more close lenses , such as a spiral arm concentration , where the relative angular velocity of source and lens is small , than are accounted for in the axisymmetric disk model .
other than the triaxial bulge , deviations from axial symmetry in the galaxy model are ignored .
in addition , streaming velocities superposed on the galactic rotation are not accounted for .
could some combination of such additional degrees of freedom hide a galaxy - wide distribution of brown dwarfs in the microlensing data ?
one can not be sure without better constraints on galactic structure and velocity parameters .
for the present , the addition of the brown dwarfs to the mass function appears to widen the discrepancies between predicted and observed microlensing time scale frequency distributions . | the 2-micron all sky survey ( 2mass ) ( skrutskie _ et al . _
1997 ) and the deep near infrared survey of the southern sky ( denis ) ( epchtein _ et al .
_ 1994 ) have revealed a heretofore unknown population of free brown dwarfs that has extended the local mass function down to as small as @xmath0 ( reid _ et al .
_ 1999 ) .
if this local proportion of brown dwarfs extends throughout the galaxy in particular in the galactic bulge one expects an increase in the predicted fraction of short time scale microlensing events in directions toward the galactic bulge .
zhao _
et al . _
( 1996 ) have indicated that a mass function with 30 - 60% of the lens mass in brown dwarfs is not consistent with empirical microlensing data . here
we show that even the much lower mass fraction ( @xmath1 ) of brown dwarfs inferred from the new discoveries appears inconsistent with the data .
the added brown dwarfs do indeed increase the expected number of short time scale events , but they appear to drive the peak in the time scale frequency distribution to time scales smaller than that observed , and do not otherwise match the observed distribution . a reasonably good match to the empirical data ( alcock _ et al .
_ 1996 ) is obtained by increasing the fraction of stars in the range @xmath2 considerably above that deduced from several star counts .
however , all inferences from microlensing about the appropriate stellar mass function must be qualified by the meagerness of the microlensing data and the uncertainties in the galactic model . |
in the last century the fundamental discoveries of nuclear physics have led to the understanding of the thermonuclear energy source of main sequence stars and explained the basic physical processes underlying the solar luminosity ( see e.g. m. schwarzschild @xcite ) . the discovery of pulsars in 1968 ( see hewish et al .
@xcite ) led to the first evidence for the existence of neutron stars , first described in terms of theoretical physics by george gamow as far back as 1936 @xcite .
it became clear that the pulsed luminosity of pulsars , at times @xmath4
@xmath5 larger than solar luminosity , was not related to nuclear burning and could be simply explained in term of the loss of rotational energy of a neutron star ( gold @xcite ) . for the first time it became so clear the possible relevance of strong gravitational fields in the energetics of an astrophysical system .
the birth of x - ray astronomy thanks to riccardo giacconi and his group ( see e.g. giacconi and ruffini @xcite ) led to a still different energy source , originating from the accretion of matter onto a star which has undergone a complete gravitational collapse process : a black hole ( see e.g. ruffini & wheeler @xcite ) . in this case , the energetics is dominated by the radiation emitted in the accretion process of matter around an already formed black hole .
luminosities up to @xmath2 times the solar luminosity , much larger then the ones of pulsars , could be explained by the release of energy in matter accreting in the deep potential well of a black hole ( leach and ruffini @xcite ) .
this allowed to probe for the first time the structure of circular orbits around a black hole computed by ruffini and wheeler ( see e.g. landau and lifshitz @xcite ) .
this result was well illustrated by the theoretical interpretation of the observations of cygnus - x1 , obtained by the uhuru satellite and by the optical and radio telescopes on the ground ( see fig . [ cygx1 ] ) .
of @xmath2 solar luminosities points to the accretion process into a neutron star or a black hole as the energy source .
the absence of pulsation is naturally explained either by a non - magnetized neutron star or a kerr - newmann black hole , which has necessarily to be axially symmetric .
what identifies the black hole unambiguously is that the mass of cygnus x-1 , larger than @xmath6 , exceeds the absolute upper limit of the neutron star mass , estimated at @xmath7 by rhoades and ruffini @xcite . ]
the discovery of gamma - ray bursts ( grbs ) sign a further decisive progress .
the grbs give the first opportunity to probe and observe a yet different form of energy : the extractable energy of the black hole introduced in 1971 ( christodoulou and ruffini @xcite ) , which we shall refer in the following as the blackholic energy .
the blackholic energy , expected to be emitted during the dynamical process of gravitational collapse leading to the formation of the black hole , generates x - ray luminosities @xmath8 times larger than the solar luminosity , although lasting for a very short time .
the extreme regimes of grbs evidence new and unexplored regimes of theoretical physics .
it is the aim of this talk to outline the progress achieved in understanding these astrophysical systems and the theoretically predicted regimes for the first time submitted to direct observational verification .
it is a pleasure to present these results in brazil .
while sitting at the casino de urca , george gamow and mario schoenberg in 1939 identified the basic process leading to the formation and cooling of a newly born neutron star ( see [ gamow ] ) .
they called this process essentially related to the emission of neutrinos and antineutrinos the urca process .
it is a welcomed coincidence that , during the preparation of my talk at the tenth marcel grossmann meeting @xcite , examining the data of the recently observed grb 030329 , we have received a confirmation of a scenario we have recently outlined in three papers giving the theoretical paradigms for the understanding of grbs ( ruffini et al .
@xcite ) .
we have clear evidence , first advanced in the system grb980425/sn1998bw ( ruffini et al .
@xcite , fraschetti et al .
@xcite ) and now confirmed also in the system grb030329/sn1003dh , that there are in these systems three different components : 1 ) the grb source , generated by the collapse to a black hole , 2 ) the supernova , generated by the collapse of an evolved star , 3 ) an additional x - ray source which is not related , unlike what is at times stated in the literature , to the grb afterglow . in honor of the work done in the town of urca by george gamow and mario schoenberg , identifying in the neutrino emission of the urca process the basic mechanism leading to the process of gravitational collapse and the formation of a relativistic compact star , we named these two x - ray sources urca-1 , the one formed in the system grb980425/sn1998bw , and urca-2 , the one formed in the system grb030329/sn2003dh . we shall now recall some of the main steps in reaching this understanding out of the grb phenomenon and explore possible explanation of the origin of these two sources .
we then turn to the analysis of the short grbs .
we first review some progress in the study of the general relativistic collapse of a shell of matter endowed with electromagnetic fields , which has been motivated by the study of the short grbs .
we then deduce from these theoretical developments some consequences for the interpretation of the mass - energy formula of the black hole , as well as some conceptual consequences for the relation between general relativity and thermodynamics .
we also apply some recent progress on the solution of the einstein - maxwell - blasov equations to the thermalization process occurring in the electron - positron plasma generated by the vacuum polarization process .
we finally give a very specific theoretical prediction on the burst structure to be expected in short grbs .
we then proceed to the conclusions and some general considerations about this novel astrophysical scenario .
it is well known how grbs were detected and studied for the first time using the _ vela _ satellites , developed for military research to monitor the non - violation of the limited test ban treaty signed in 1963 ( see e.g. strong @xcite ) .
it was clear from the early data of these satellites , which were put at @xmath9 miles from the surface of earth , that the grbs did not originate either on the earth nor in the solar system .
the mystery of these sources became more profound as the observations of the batse instrument on board of the compton gamma - ray observatory ( cgro ) satellite over @xmath10 years proved the isotropy of these sources in the sky ( see fig .
[ batse2k ] ) .
in addition to these data , the cgro satellite gave an unprecedented number of details on the grb structure , on their spectral properties and time variabilities which became encoded in the fourth batse catalog @xcite ( see e.g. fig .
[ grb_profiles_eng ] ) . out of the analysis of these batse sources
it soon became clear ( see e.g. kouveliotou et al .
@xcite , tavani @xcite ) the existence of two distinct families of sources : the long bursts , lasting more then one second and softer in spectra , and the short bursts ( see fig .
[ slb ] ) , harder in spectra ( see fig .
[ tavani ] ) .
we shall return shortly on this topic .
s ) and long ( @xmath11 s ) grbs are represented . reproduced , by his kind permission , from tavani @xcite where the details are given . ] . ]
the situation drastically changed with the discovery of the afterglow by the italian - dutch satellite bepposax ( costa et al .
@xcite ) and the possibility which led to the optical identification of the grbs by the largest telescopes in the world , including the hubble space telescope , the keck telescope in hawaii and the vlt in chile , and allowed as well the identification in the radio band of these sources .
the outcome of this collaboration between complementary observational technique has led to the possibility of identifying in 1997 the distance of these sources from the earth and their tremendous energy of the order up to @xmath12 erg / second during the burst .
it is interesting , as we will show in the following , that an energetics of this magnitude for the grbs had previously been predicted out of first principles already in 1974 by damour and ruffini @xcite .
the resonance between the x- and gamma ray astronomy from the satellites and the optical and radio astronomy from the ground , had already marked in the seventies the great success and development of the astrophysics of binary x - ray sources ( see e.g. giacconi & ruffini @xcite ) .
this resonance is re - proposed here for grbs on a much larger scale .
the use of much larger satellites , like chandra and xmm - newton , and dedicated space missions , like hete-2 and , in the near future , swift , and the very fortunate circumstance of the coming of age of the development of unprecedented optical technologies for the telescopes offers opportunities without precedence in the history of mankind . in parallel ,
the enormous scientific interest on the nature of grb sources and the exploration , not only of new regimes , but also of totally novel conceptual physical process of the blackholic energy , make the knowledge of grbs an authentic new frontier in the scientific knowledge .
the study of grbs is very likely `` the '' most extensive computational and theoretical investigation ever done in physics and astrophysics .
there are at least three different fields of research which underlie the foundation of the theoretical understanding of grbs .
all three , for different reasons , are very difficult . the first field of research is the field of special relativity . as i always mention to my students in the course of theoretical physics , this field is paradoxically very difficult since it is extremely simple . in approaching special relativistic phenomena
the extremely simple and clear procedures expressed by einstein in his 1905 classic paper @xcite are often ignored .
einstein makes use in his work of very few physical assumptions , an almost elementary mathematical framework and gives constant attention to a proper operational definition of all observable quantities .
those who work on grbs use at times very intricate , complex and often wrong theoretical approaches lacking the necessary self - consistency .
this is well demonstrated in the current literature on grbs . the second field of research essential for understanding the energetics of grbs deals with quantum electrodynamics and the relativistic process of pair creation in overcritical electromagnetic fields .
this topic is also very difficult but for a quite different conceptual reason : the process of pair creation , expressed in the classic works of heisenberg - euler - schwinger @xcite later developed by many others , is based on a very powerful theoretical framework but has never been verified by experimental data .
the quest for creating electron - positron pairs by vacuum polarization processes in heavy ion collisions or in lasers has not yet been successfully achieved in earth - bound experiments ( see e.g. ruffini , vitagliano , xue @xcite ) . as we will show here , there is the tantalizing possibility of observing this phenomenon , for the first time , in the astrophysical setting of grbs on a more grandiose scale . there is a third field which is essential for the understanding of the grb phenomenon : general relativity . in this case , contrary to the case of special relativity , the field is indeed very difficult , since it is very difficult both from a conceptual , technical and mathematical point of view .
the physical assumptions are indeed complex .
the entire concept of geometrization of physics needs a new conceptual approach to the field .
the mathematical complexity of the pseudo - riemannian geometry contrasts now with the simple structure of the pseudo - euclidean minkowski space .
the operational definition of the observable quantities has to take into account the intrinsic geometrical properties and also the cosmological settings of the source . with grbs we have the possibility to follow , from a safe position in an asymptotically flat space at large distance , the formation of the horizon of a black hole with all the associated relativistic phenomena of light bending and time dilatation .
most important , as we will show in details in this presentation , general relativity in connection with quantum phenomena offers , with the blackholic energy , the explanation of the tremendous grb energy sources . for these reasons grbs offer an authentic new frontier in the field of physics and astrophysics .
it is appropriate to mention some of the goals of such a new frontier in the above three fields .
we recall in the special relativity field , for the first time , we observe phenomena occurring at lorentz gamma factors of approximately @xmath13 . in the field of relativistic quantum electro - dynamics we see for the first time the interchange between classical fields and
the created quantum matter - antimatter pairs . in the field of general relativity also for the first time we can test the blackholic energy which is the basic energetic physical variable underlying the entire grb phenomenon .
the most appealing aspect of this work is that , if indeed these three different fields are treated and approached with the necessary technical and scientific maturity , the model which results has a very large redundancy built - in .
the approach requires an unprecedented level of self - consistency .
any departures from the correct theoretical treatment in this very complex system lead to exponential departures from the correct solution and from the correct fit of the observations .
it is so that , as the model is being properly developed and verified , its solution will have existence and uniqueness .
the ongoing dialogue between our work and the one of the workers on grbs , rests still on some elementary considerations presented by einstein in his classic article of 1905 @xcite .
these considerations are quite general and even precede einstein s derivation , out of first principles , of the lorentz transformations .
we recall here einstein s words : `` we might , of course , content ourselves with time values determined by an observer stationed together with the watch at the origin of the co - ordinates , and co - ordinating the corresponding positions of the hands with light signals , given out by every event to be timed , and reaching him through empty space . but
this co - ordination has the disadvantage that it is not independent of the standpoint of the observer with the watch or clock , as we know from experience '' . and the laboratory time @xmath14 .
details in ruffini et al .
] the message by einstein is simply illustrated in fig .
[ ttasch_new_bn ] .
if we consider in an inertial frame a source ( solid line ) moving with high speed and emitting light signals ( dashed lines ) along the direction of its motion , a far away observer will measure a delay @xmath15 between the arrival time of two signals emitted at the origin and after a time interval @xmath16 in the laboratory frame .
the real velocity of the source is given by : @xmath17 and the apparent velocity is given by : @xmath18 as pointed out by einstein the adoption of coordinating light signals simply by their arrival time as in eq.([vapp ] ) , without an adequate definition of synchronization , is incorrect and leads to unsurmountable difficulties as well as to apparently `` superluminal '' velocities as soon as motions close to the speed of light are considered .
the use of @xmath15 as a time coordinate , often tacitly adopted by astronomers , should be done , if at all , with proper care .
the relation between @xmath15 and the correct time parameterization in the laboratory frame has to be taken into account : @xmath19 in other words , the relation between the arrival time and the laboratory time can not be done without a knowledge of the speed along the entire world - line of the source . in the case of grbs ,
such a worldline starts at the moment of gravitational collapse .
it is of course clear that the parameterization in the laboratory frame has to take into account the cosmological redshift @xmath20 of the source .
we then have , at the detector : @xmath21 in the current grb literature , eq.([tadef ] ) has been systematically neglected by addressing only the afterglow description neglecting the previous history of the source . often the integral equation has been approximated by a clearly incorrect instantaneous value : @xmath22 the attitude has been adopted that it should be possible to consider separately the afterglow part of the grb phenomenon , without the knowledge of the entire equation of motion of the source .
this point of view has reached its most extreme expression in the works reviewed by piran @xcite , where the so - called `` prompt radiation '' , lasting on the order of @xmath4 s , is considered as a burst emitted by the prolonged activity of an `` inner engine '' . in these models ,
generally referred to as the `` internal shock model '' , the emission of the afterglow is assumed to follow the `` prompt radiation '' phase @xcite .
as we outline in the following , such an extreme point of view originates from the inability of obtaining the time scale of the `` prompt radiation '' from a burst structure .
these authors consequently appeal to the existence of an `` ad hoc '' inner engine in the grb source to solve this problem .
we show in the following how this difficulty has been overcome in our approach by interpreting the `` prompt radiation '' as an integral part of the afterglow and _ not _ as a burst .
this explanation can be reached only through a relativistically correct theoretical description of the entire afterglow ( see section [ aft ] ) . within the framework of special relativity
we show that it is not possible to describe a grb phenomenon by disregarding the knowledge of the entire past worldline of the source .
we show that at @xmath4 seconds the emission occurs from a region of dimensions of approximately @xmath23 cm , well within the region of activity of the afterglow .
this point was not appreciated in the current literature due to the neglect of the apparent superluminal effects implied by the use of the `` pathological '' parametrization of the grb phenomenon by the arrival time of light signals .
an additional difference between our treatment and the ones in the current literature relates to the assumption of the existence of scaling laws in the afterglow phase : the power law dependence of the lorentz gamma factor on the radial coordinate is usually systematically assumed . from the proper use of the relativistic transformations and by the direct numerical and analytic integration of the special relativistic equations of motion we demonstrate ( see section [ eqaft ] ) that no simple power - law relation can be derived for the equations of motion of the system .
this situation is not new for workers in relativistic theories : scaling laws exist in the extreme ultrarelativistic regimes and in the newtonian ones but not in the intermediate fully relativistic regimes ( see e.g. ruffini @xcite ) .
three of the most important works in the field of general relativity have certainly been the discovery of the kerr solution @xcite , its generalization to the charged case ( newman et al .
@xcite ) and the formulation by brandon carter @xcite of the hamilton - jacobi equations for a charged test particle in the metric and electromagnetic field of a kerr - newman solution ( see e.g. landau and lifshitz @xcite ) .
the equations of motion , which are generally second order differential equations , were reduced by carter to a set of first order differential equations which were then integrated by using an effective potential technique by ruffini and wheeler for the kerr metric ( see e.g. landau and lifshitz @xcite ) and by ruffini for the reissner - nordstrm geometry ( ruffini @xcite , see fig .
[ effp ] ) .
all the above mathematical results were essential for understanding the new physics of gravitationally collapsed objects and allowed the publication of a very popular article : `` introducing the black hole '' ( ruffini and wheeler @xcite ) . in that paper
, we advanced the ansatz that the most general black hole is a solution of the einstein - maxwell equations , asymptotically flat and with a regular horizon : the kerr - newman solution , characterized only by three parameters : the mass @xmath24 , the charge @xmath25 and the angular momentum @xmath26 .
this ansatz of the `` black hole uniqueness theorem '' still today after thirty years presents challenges to the mathematical aspects of its complete proof ( see e.g. carter @xcite and bini et al .
in addition to these mathematical difficulties , in the field of physics this ansatz contains the most profound consequences .
the fact that , among all the possible highly nonlinear terms characterizing the gravitationally collapsed objects , only the ones corresponding solely to the einstein maxwell equations survive the formation of the horizon has , indeed , extremely profound physical implications .
any departure from such a minimal configuration either collapses on the horizon or is radiated away during the collapse process .
this ansatz is crucial in identifying precisely the process of gravitational collapse leading to the formation of the black hole and the emission of grbs .
indeed , in this specific case , the born - like nonlinear @xcite term of the heisenberg - euler - schwinger @xcite lagrangian are radiated away prior to the formation of the horizon of the black hole ( see e.g. ruffini et al .
@xcite ) . only the nonlinearity corresponding solely to the classical einstein - maxwell theory is left as the outcome of the gravitational collapse process . the same effective potential technique ( see landau and lifshitz @xcite ) , which allowed the analysis of circular orbits around the black hole , was crucial in reaching the equally interesting discovery of the reversible and irreversible transformations of black holes by christodoulou and ruffini @xcite , which in turn led to the mass - energy formula of the black hole : @xmath27 with @xmath28 where @xmath29 is the horizon surface area
, @xmath30 is the irreducible mass , @xmath31 is the horizon radius and @xmath32 is the quasi - spheroidal cylindrical coordinate of the horizon evaluated at the equatorial plane .
extreme black holes satisfy the equality in eq.([s1 ] ) . from eq.([em ] ) follows that the total energy of the black hole @xmath33 can be split into three different parts : rest mass , coulomb energy and rotational energy . in principle both coulomb energy and rotational energy can be extracted from the black hole ( christodoulou and ruffini @xcite ) .
the maximum extractable rotational energy is 29% and the maximum extractable coulomb energy is 50% of the total energy , as clearly follows from the upper limit for the existence of a black hole , given by eq.([s1 ] ) .
we refer in the following to both these extractable energies as the blackholic energy .
the existence of the black hole and the basic correctness of the circular orbits has been proven by the observations of cygnus - x1 ( see e.g. giacconi and ruffini @xcite ) . however ,
in binary x - ray sources , the black hole uniquely acts passively by generating the deep potential well in which the accretion process occurs .
it has become tantalizing to look for astrophysical objects in order to verify the other fundamental prediction of general relativity that the blackholic energy is the largest energy extractable from any physical object .
as we shall see in the next section , the feasibility of the extraction of the blackholic energy has been made possible by the quantum processes of creating , out of classical fields , a plasma of electron - positron pairs in the field of black holes .
the manifestation of such process of energy extraction from the black hole is astrophysically manifested by the occurrence of grbs . that a static electromagnetic field stronger than a critical value : @xmath34 can polarize the vacuum and create electron - positron pairs was clearly evidenced by heisenberg and euler @xcite .
the major effort in verifying the correctness of this theoretical prediction has been directed in the analysis of heavy ion collisions ( see ruffini et al . @xcite and references therein ) . from an order - of - magnitude estimate , it appears that around a nucleus with a charge : @xmath35 the electric field can be stronger than the electric field polarizing the vacuum .
a more accurate detailed analysis taking into account the bound states levels around a nucleus brings to a value of @xmath36 for the nuclear charge leading to the existence of a critical field . from the heisenberg uncertainty principle it follows that
, in order to create a pair , the existence of the critical field should last a time @xmath37 which is much longer then the typical confinement time in heavy ion collisions which is @xmath38 this is certainly a reason why no evidence for pair creation in heavy ion collisions has been obtained although remarkable effort has been spent in various accelerators worldwide .
similar experiments involving laser beams encounter analogous difficulties ( see e.g. ruffini et al .
@xcite and references therein ) .
the alternative idea was advanced in 1975 @xcite that the critical field condition given in eq.([ec ] ) could be reached easily , and for a time much larger than the one given by eq.([dt ] ) , in the field of a kerr - newman black hole in a range of masses @xmath39 . in that paper
we have generalized to the curved kerr - newman geometry the fundamental theoretical framework developed in minkowski space by heisenberg - euler @xcite and schwinger @xcite .
this result was made possible by the work on the structure of the kerr - newman spacetime previously done by carter @xcite and by the remarkable mathematical craftsmanship of thibault damour then working with me as a post - doc in princeton .
the maximum energy extractable in such a process of creating a vast amount of electron - positron pairs around a black hole is given by : @xmath40 we concluded in that paper that such a process `` naturally leads to a most simple model for the explanation of the recently discovered @xmath41-rays bursts '' .
at that time , grbs had not yet been optically identified and nothing was known about their distance and consequently about their energetics .
literally thousands of theories existed in order to explain them and it was impossible to establish a rational dialogue with such an enormous number of alternative theories .
we did not pursue further our model until the results of the bepposax mission , which clearly pointed to the cosmological origin of grbs , implying for the typical magnitude of their energy precisely the one predicted by our model .
it is interesting that the idea of using an electron - positron plasma as a basis of a grb model was independently introduced years later in a set of papers by cavallo and rees @xcite , cavallo and horstman @xcite and horstman and cavallo @xcite .
these authors did not address the issue of the physical origin of their energy source .
they reach their conclusions considering the pair creation and annihilation process occurring in the confinement of a large amount of energy in a region of dimension @xmath42 km typical of a neutron star .
no relation to the physics of black holes nor to the energy extraction process from a black hole was envisaged in their interesting considerations , mainly directed to the study of the opacity and the consequent dynamics of such an electron - positron plasma .
after the discovery of the afterglows and the optical identification of grbs at cosmological distances , implying exactly the energetics predicted in eq.([emax ] ) , we returned to the analysis of the vacuum polarization process around a black hole and precisely identified the region around the black hole in which the vacuum polarization process and the consequent creation of electron - positron pairs occur .
we defined this region , using the greek name dyad for pairs ( @xmath43 , @xmath44 ) , to be the `` dyadosphere '' of the black hole , bounded by the black hole horizon and the dyadosphere radius @xmath45 given by ( see ruffini @xcite , preparata et al . @xcite and fig.[dya ] ) : @xmath46 where we have introduced the dimensionless mass and charge parameters @xmath47 , @xmath48 .
the analysis of the dyadosphere was developed , at that time , around an already formed black hole . in recent months
we have been developing the dynamical formation of the black hole and correspondingly of the dyadosphere during the process of gravitational collapse , reaching some specific signatures which may be detectable in the structure of the short and long grbs ( cherubini et al .
@xcite , ruffini and vitagliano @xcite , ruffini et al .
many details of this topic have been presented in great details in ruffini et al .
@xcite .
after the vacuum polarization process around a black hole , one of the topics of the greatest scientific interest is the analysis of the dynamics of the electron - positron plasma formed in the dyadosphere .
this issue was addressed by us in a very effective collaboration with jim wilson at livermore .
the numerical simulations of this problem were developed at livermore , while the semi - analytic approach was developed in rome ( ruffini et al .
@xcite ) . the corresponding treatment in the framework of the cavallo et al . analysis was performed by piran et al . @xcite also using a numerical approach , by bisnovaty - kogan and murzina @xcite using an analytic approach and by mszros , laguna and rees @xcite using a numerical and semi - analytic approach . although some analogies exists between these treatments , they are significantly different in the theoretical details and in the final results . since the final result of the grb model is extremely sensitive to any departure from the correct treatment , it is indeed very important to detect at every step the appearance of possible fatal errors .
a conclusion common to all these treatments is that the electron - positron plasma is initially optically thick and expands till transparency reaching very high values of the lorentz gamma factor .
a second point , which is common , is the discovery of a new clear feature : the plasma shell expands but the lorentz contraction is such that its width in the laboratory frame appears to be constant .
there is however a major difference between our approach and the ones of piran , mszros and rees , in that the dyadosphere is assumed by us to be filled uniquely with an electron - positron plasma .
such a plasma expands in substantial agreement with the results presented in the work of bisnovati - kogan and murzina @xcite . in our model
the pulse of electron - positron pairs and photons ( pem pulse , see ruffini et al .
@xcite ) evolves and at a radius on the order of @xmath49 cm it encounters the remnant of the star progenitor of the newly formed black hole .
the pem pulse is then loaded with baryons .
a new pulse is formed of electron - positron - photons and baryons ( pemb pulse , see ruffini et al .
@xcite ) which expands all the way until transparency is reached . at transparency
the emitted photons give origin to what we define as the proper - grb ( see ruffini et al . @xcite and fig .
[ cip2 ] ) . ) .
details in ruffini et al .
] in our approach , the baryon loading is measured by a dimensionless quantity @xmath50 which gives direct information about the mass @xmath51 of the remnant . the corresponding treatment done by piran and collaborators ( shemi & piran @xcite , piran et al .
@xcite ) and by mszros , laguna and rees @xcite differs in one important respect : the baryonic loading is assumed to occur since the beginning of the electron - positron pair formation and no relation to the mass of the remnant of the collapsed progenitor star is attributed to it .
a marked difference also exists between our description of the rate equation for the electron - positron pairs and the ones by those authors . while our results are comparable with the ones obtained by piran under the same initial conditions , the set of approximations adopted by mszros , laguna and rees @xcite appears to be too radical and leads to very different results violating energy and momentum conservation ( see bianco et al .
@xcite ) . from our analysis ( ruffini et al .
@xcite ) it also becomes clear that such expanding dynamical evolution can only occur for values of @xmath52 .
this prediction , as we will show shortly in the three grb sources considered here , is very satisfactorily confirmed by observations . from the value of the @xmath53 parameter , related to the mass of the remnant
, it therefore follows that the collapse to a black hole leading to a grb is drastically different from the collapse to a neutron star . while in the case of a neutron star collapse
a very large amount of matter is expelled , in many instances well above the mass of the neutron star itself , in the case of black holes leading to a grb only a very small fraction of the initial mass ( @xmath54 or less ) is expelled .
the collapse to a black hole giving rise to a grb appears to be much smoother than any collapse process considered until today : almost 99.9% of the star has to be collapsing simultaneously !
we summarize in figs .
[ cip2][cip3 ] the optically thick phase of grbs in our model : we start from a given dyadosphere of energy @xmath55 ; the pair - electromagnetic pulse ( pem pulse ) self - accelerates outward typically reaching lorentz gamma factors @xmath56 at @xmath57 cm ; at this point the collision of the pem pulse with the remnant of the progenitor star occurs with an abrupt decrease in the value of the lorentz gamma factor ; a new pair - electromagnetic - baryon pulse ( pemb pulse ) is formed which self - accelerates outward until the system becomes transparent . which has been computed following the prescriptions of eq.([tadef ] ) .
details in ruffini et al .
] the photon emission at this transparency point is the proper - grb ( p - grb ) .
an accelerated beam of baryons with an initial lorentz gamma factor @xmath58 starts to interact with the interstellar medium at typical distances from the black hole of @xmath59 cm and at a photon arrival time at the detector on the earth surface of @xmath60 s. these values determine the initial conditions of the afterglow .
after reaching transparency and the emission of the p - grb , the accelerated baryonic matter ( the abm pulse ) interacts with the interstellar medium ( ism ) and gives rise to the afterglow ( see fig .
[ cip_tot ] ) .
also in the descriptions of this last phase many differences exist between our treatment and the other ones in the current literature . ) .
the value of the lorentz gamma factor is here given from the transparency point all the way to the ultrarelativisitc , relativistic and non relativistic regimes .
details in ruffini et al .
] the initial conditions of the afterglow era are determined at the end of the optically thick era when the p - grb is emitted .
as recalled in the last section , the transparency condition is determined by a time of arrival @xmath61 , a value of the gamma lorentz factor @xmath58 , a value of the radial coordinate @xmath62 , an amount of baryonic matter @xmath51 which are only functions of the two parameters @xmath55 and @xmath53 ( see eq.([bdef ] ) ) .
it is appropriate here to emphasize again that , in order to have the expansion leading to an observed grb , one must have @xmath52 .
this connection to the optically thick era is missing in the current approach in the literature which attributes the origin of the `` prompt radiation '' to an unspecified inner engine activity ( see piran @xcite and references therein ) .
the initial conditions at the beginning of the afterglow era are obtained by a best fit of the later parts of the afterglow .
this approach is quite unsatisfactory since , as we will explicitly show , the theoretical treatments currently adopted in the description of the afterglow are not correct .
the fit using an incorrect theoretical treatment leads necessarily to the wrong conclusions as well as , in turn , to the determination of incorrect initial conditions .
let us first summarize the commonalities between our approach and the ones in the current literature . in both cases
( see piran @xcite , chiang & dermer @xcite and ruffini et al .
@xcite ) a thin shell approximation is used to describe the collision between the abm pulse and the ism : [ taub_eq ] @xmath63 where @xmath64 , @xmath41 and @xmath24 are respectively the internal energy , the lorentz factor and the mass - energy of the expanding pulse , @xmath65 is the ism number density which is assumed to be constant , @xmath66 is the proton mass , @xmath67 is the emitted fraction of the energy developed in the collision with the ism and @xmath68 is the amount of ism mass swept up within the radius @xmath69 : @xmath70 , where @xmath62 is the starting radius of the shock front . in general
, an additional condition is needed in order to determine @xmath67 as a function of the radial coordinate . in the following , @xmath67
is assumed to be constant and such an approximation appears to be correct in the grb context . in both our work and in the current literature ( see piran @xcite , chiang & dermer @xcite and ruffini et al .
@xcite ) a first integral of these equations has been found , leading to expressions for the lorentz gamma factor as a function of the radial coordinate . in the `` fully adiabatic condition '' ( i.e. @xmath71 )
we have : @xmath72 while in the `` fully radiative condition '' ( i.e. @xmath73 ) we have : @xmath74}{\gamma_\circ^{-1}+\left(m_\mathrm{ism}/m_b\right)\left(1+\gamma_\circ^{-1 } \right)\left[1+\left(1/2\right)\left(m_\mathrm{ism}/m_b\right)\right]}\ , , \label{gamma_rad}\ ] ] where @xmath58 and @xmath51 are respectively the values of the lorentz gamma factor and of the mass of the accelerated baryons at the beginning of the afterglow phase and @xmath62 is the value of the radius @xmath69 at the beginning of the afterglow phase .
a major difference between our treatment and the other ones in the current literature is that we have integrated the above equations analytically , obtaining the explicit analytic form of the equations of motion for the expanding shell in the afterglow for a constant ism density . for the fully radiative case we have explicitly integrated the differential equation for @xmath75 in eq . , recalling that @xmath76 ^ 2 $ ] , where @xmath14 is the time in the laboratory reference frame .
we have then obtained a new explicit analytic solution of the equations of motion for the relativistic shell in the entire range from the ultra - relativistic to the non - relativistic regimes : @xmath77 ^ 3 \left(a^3 + 1\right)}}{{\left[a^3 + \left ( r / r_\circ \right)^3\right ] \left ( { a + 1 } \right)^3 } } } \right\ } - \tfrac{m_i^\circ r_\circ } { 8c\sqrt c } \\ & + t_\circ + \tfrac{m_i^\circ r_\circ } { 8c\sqrt c } \left ( { \tfrac{r}{{r_\circ } } } \right)^4 + \tfrac{{r_\circ \sqrt{3c}}}{{6 c m_i^\circ a^2 } } \left[\arctan \tfrac{{2\left(r / r_\circ\right ) - a}}{{a\sqrt 3 } } - \arctan \tfrac{{2 - a}}{{a\sqrt 3 } } \right ] \end{split } \label{analsol}\ ] ] where @xmath78{\left(m_b - m_i^\circ\right)/m_i^\circ}$ ] , @xmath79 and @xmath80 .
correspondingly , in the adiabatic case we have : @xmath81 in the current literature , following blandford and mckee @xcite , a so - called `` ultrarelativistic '' approximation @xmath82 has been widely adopted by many authors to solve eqs .
( see e.g. sari @xcite , waxman @xcite , rees & mszros @xcite , granot et al .
@xcite , panaitescu & mszros @xcite , piran @xcite , gruzinov & waxman @xcite , van paradijs et al .
@xcite , mszros @xcite and references therein ) .
this leads to simple constant - index power - law relations : [ gr0-rct ] @xmath83 with @xmath84 in the fully radiative case and @xmath85 in the fully adiabatic case .
this simple relation is in stark contrast to the complexity of eq . and
eq . respectively . in the same spirit , instead of eq . and
eq . , some authors have assumed the following much simpler approximation for the relation between the time and the radial coordinate of the expanding shell , both in the fully radiative and in the fully adiabatic cases : @xmath86 while others , like e.g. panaitescu & mszros @xcite , have integrated the approximate eq .
, obtaining : @xmath87\ , .
\label{t_app_pm98c}\ ] ] again , it is appropriate here to emphasize the stark contrast between eqs . , and the exact analytic solutions of eqs .
, expressed in eqs . , .
as pointed out long ago by couderc @xcite , in all relativistic expansion the crucial geometrical quantities with respect to a physical observer are the `` equitemporal surfaces '' ( eqtss ) , namely the locus of source points of the signals arriving at the observer at the same time . for a relativistically expanding spherically symmetric source
the eqtss are surfaces of revolution about the line of sight . the general expression for their profile , in the form @xmath88
, corresponding to an arrival time @xmath89 of the photons at the detector , can be obtained from ( see e.g. ruffini et al .
@xcite , bianco and ruffini @xcite and figs .
[ openang][opening ] ) : @xmath90 where @xmath91 is the initial size of the expanding source , @xmath92 is the angle between the radial expansion velocity of a point on its surface and the line of sight , and @xmath93 is its equation of motion , expressed in the laboratory frame , obtained by the integration of eqs.([taub_eq ] ) . from the definition of the lorentz gamma factor @xmath94 , we have in fact : @xmath95^{-1/2}dr'\ , , \label{tdir}\ ] ] where @xmath96 comes from the integration of eqs.([taub_eq ] ) .
are allowed .
only photons emitted at an angle such that @xmath97 can be viewed by the observer .
thus the maximum allowed @xmath92 value @xmath98 corresponds to @xmath99 . in this figure
we show @xmath98 ( i.e. the angular amplitude of the visible area of the abm pulse ) in degrees as a function of the arrival time at the detector for the photons emitted along the line of sight ( see text ) . in the earliest grb phases @xmath100 and so @xmath101 . on the other hand , in the latest phases of the afterglow the abm pulse velocity decreases and @xmath98 tends to the maximum possible value , i.e. @xmath102 .
details in ruffini et al .
@xcite ] ) @xmath98 is very small ( see left pane and fig .
[ opening ] ) , so the visible area is just a small fraction of the total abm pulse surface . on the other hand , in the final expansion phases @xmath103 and almost all
the abm pulse surface becomes visible .
details in ruffini et al .
@xcite ] and constant .
in fact , in grb 991216 the lorentz gamma factor ranges from @xmath104 to @xmath105 .
the eqtss represented here ( solid lines ) correspond respectively to values of the arrival time ranging from @xmath106 ( the smallest surface on the left of the plot ) to @xmath107 ( the largest one on the right ) .
each surface differs from the previous one by @xmath106 . to each eqts
contributes emission processes occurring at different values of the lorentz gamma factor .
the dashed lines are the boundaries of the visible area of the abm pulse and the black hole is located at position @xmath108 in this plot . note the different scales on the two axes , indicating the very high eqts `` effective eccentricity '' .
the time interval from @xmath109 to @xmath110 has been chosen to encompass the e - ape emission , ranging from @xmath111 to @xmath112 .
details in ruffini et al .
@xcite ] we have obtained the expressions in the adiabatic case and in the fully radiative cases respectively ( see bianco and ruffini @xcite ) : @xmath113 + \frac{ct_\circ}{r } \\[6pt ] & - \frac{ct_a}{r } + \frac{r^\star}{r } - \frac{\gamma_\circ-\left(m_i^\circ / m_b\right)}{\sqrt{\gamma_\circ^2 - 1}}\left[\frac{r_\circ}{r } - 1\right]\ , . \end{split } \label{eqts_g_dopo_ad}\ ] ] @xmath114 \\[6pt ] & + \frac{{r_\circ \sqrt{c } } } { { 12rm_i^\circ a^2 } } \ln \left\ { { \frac{{\left [ { a + \left(r / r_\circ\right ) } \right]^3 \left(a^3 + 1\right)}}{{\left[a^3 + \left ( r / r_\circ \right)^3\right ] \left ( { a + 1 } \right)^3 } } } \right\ } + \frac{ct_\circ}{r}-\frac{ct_a}{r } \\[6pt ] & + \frac{r^\star}{r } + \frac{{r_\circ \sqrt{3c } } } { { 6rm_i^\circ a^2 } } \left [ \arctan \frac{{2\left(r / r_\circ\right ) - a}}{{a\sqrt{3 } } } - \arctan \frac{{2 - a}}{{a\sqrt{3 } } } \right]\ , .
\end{split } \label{eqts_g_dopo}\ ] ] the two eqtss are represented at selected values of the arrival time @xmath89 in fig .
[ eqts_comp ] , where the illustrative case of grb 991216 has been used as a prototype .
the initial conditions at the beginning of the afterglow era are in this case given by @xmath115 , @xmath116 cm , @xmath117 s , @xmath118 cm ( see ruffini et al .
@xcite ) .
s , @xmath119 s , @xmath120 s and @xmath121 s , respectively from the inner to the outer one .
the right plot shows the eqts at an arrival time of 2 days .
details in bianco and ruffini @xcite.,title="fig:",width=309 ] s , @xmath119 s , @xmath120 s and @xmath121 s , respectively from the inner to the outer one .
the right plot shows the eqts at an arrival time of 2 days .
details in bianco and ruffini @xcite.,title="fig:",width=309 ] we assume that the internal energy due to kinetic collision is instantly radiated away and that the corresponding emission is isotropic . as in section
[ eqaft ] , let @xmath122 be the internal energy density developed in the collision . in the comoving frame the energy per unit of volume and per solid angle is simply @xmath123 due to the fact that the emission is isotropic in this frame . the total number of photons emitted is an invariant quantity independent of the frame used .
thus we can compute this quantity as seen by an observer in the comoving frame ( which we denote with the subscript `` @xmath124 '' ) and by an observer in the laboratory frame ( which we denote with no subscripts ) . doing this
we find : @xmath125 where @xmath126 comes from the projection of the elementary surface of the shell on the direction of propagation and @xmath127 is the doppler factor introduced in the two following differential transformation @xmath128 for the solid angle transformation and @xmath129 for the time transformation . the integration in @xmath130 is performed over the visible area of the abm pulse at laboratory time @xmath14 , namely with @xmath131 and @xmath98 defined in section [ eqts ] ( see figs .
[ openang][opening ] ) .
an extra @xmath132 factor comes from the energy transformation : @xmath133 see also chiang and dermer @xcite .
thus finally we obtain : @xmath134 doing this we clearly identify @xmath135 as the energy density in the comoving frame up to a factor @xmath136 ( see eq.([deo ] ) ) . then we have : @xmath137 where the integration in @xmath130 is performed over the abm pulse visible area at laboratory time @xmath14 , namely with @xmath131 and @xmath98 defined in section [ eqts ] . eq.([fluxlab ] ) gives us the energy emitted toward the observer per unit solid angle and per unit laboratory time @xmath14 in the laboratory frame . what we really need is the energy emitted per unit solid angle and per unit detector arrival time @xmath61 , so we must use the complete relation between @xmath61 and @xmath14 given in eq.([ta_g ] ) .
first we have to multiply the integrand in eq.([fluxlab ] ) by the factor @xmath138 to transform the energy density generated per unit of laboratory time @xmath14 into the energy density generated per unit arrival time @xmath61
. then we have to integrate with respect to @xmath130 over the _ equitemporal surface _ ( eqts , see section [ eqts ] ) of constant arrival time @xmath61 instead of the abm pulse visible area at laboratory time @xmath14 .
the analog of eq.([fluxlab ] ) for the source luminosity in detector arrival time is then : @xmath139 it is important to note that , in the present case of grb 991216 , the doppler factor @xmath140 in eq.([fluxarr ] ) enhances the apparent luminosity of the burst , as compared to the intrinsic luminosity , by a factor which at the peak of the afterglow is in the range between @xmath49 and @xmath141 !
we are now able to reproduce in fig .
[ bolum ] the general behavior of the luminosity starting from the p - grb to the latest phases of the afterglow as a function of the arrival time .
it is generally agreed that the grb afterglow originates from an ultrarelativistic shell of baryons with an initial lorentz factor @xmath142@xmath13 with respect to the interstellar medium ( see e.g. ruffini et al .
@xcite , bianco & ruffini @xcite and references therein ) .
using grb 991216 as a prototype , in ruffini et al .
@xcite we have shown how from the time varying bolometric intensity of the afterglow it is possible to infer the average density @xmath143 particle/@xmath144 of the interstellar medium ( ism ) in a region of approximately @xmath145 cm surrounding the black hole giving rise to the grb phenomenon .
it was shown in ruffini et al .
@xcite that the theoretical interpretation of the intensity variations in the prompt phase in the afterglow implies the presence in the ism of inhomogeneities of typical scale @xmath146 cm .
such inhomogeneities were there represented for simplicity as spherically symmetric over - dense regions with @xmath147 separated by under - dense regions with @xmath148 also of typical scale @xmath149 cm in order to keep @xmath150 constant .
the summary of these general results are shown in fig .
[ grb991216 ] , where the p - grb , the emission at the peak of the afterglow in relation to the `` prompt emission '' and the latest part of the afterglow are clearly identified for the source grb 991216 .
details in ruffini et al .
having obtained a general agreement between the observed luminosity variability and our treatment of the bolometric luminosity , we have further developed the model in order to explain + a ) the details of the observed luminosity in fixed energy bands , which are the ones actually measured by the detectors on the satellites + b ) the instantaneous as well as the average spectral distribution in the entire afterglow and + c ) the observed hard to soft drift observed in grb spectra . in order to do
so we have developed ( ruffini et al .
@xcite ) a more detailed theory of the structure of the shock front giving rise to the afterglow .
we have modeled the interaction between the ultrarelativistic shell of baryons and the ism by a shock front with three well - defined layers ( see e.g. secs . 8589 , 135 of landau & lifshitz @xcite , ch . 2 and sec . 1315 of zeldovich & rayzer @xcite and sec .
iv , 1113 of sedov @xcite ) . from the back end to the leading edge of this shock front there is : + * a ) * a compressed high - temperature layer , of thickness @xmath151 , in front of the relativistic baryonic shell , created by the accumulated material swept up in the ism . +
* b ) * a thin shock front , with a jump @xmath152 in the temperature which has been traditionally estimated in the comoving frame by the rankine - hugoniot adiabatic equations : @xmath153 ^ 2 k\ , , \label{eqt}\ ] ] where @xmath154 is the velocity jump , @xmath66 is the proton mass and @xmath155 is boltzmann s constant .
of course such a treatment , valid for @xmath156 , has to be modified ( see below ) in our novel treatment for the @xmath56 case relevant to grbs .
+ * c ) * a pre - shock layer of ism swept - up matter at much lower density and temperature , both of which change abruptly at the thin shock front behind it . at larger distances
ahead of the expanding fireball the ism is at still smaller densities .
the upper limit to the temperature jump at the thin shock front , given in eq.([eqt ] ) , is due to the transformation of kinetic energy to thermal energy , since the particle mean free path is assumed to be less then the thickness of the layer ( a ) .
the thermal emission of the observed x- and gamma ray radiation , which as seen from the observations reveals a high level of stability , is emitted in the above region ( a ) due to the sharp temperature gradient at the thin shock front described in the above region ( b ) .
the optical and radio emission comes in our model from the extended region ( c ) .
the description of such a region , unlike the sharp and well - defined temperature gradient occurring in region ( b ) , requires magnetohydrodynamic simulations of the evolution of the electron energy distribution of the synchrotron emission .
such analysis has been performed using 3-d eulerian mhd codes for the particle acceleration models to produce the energy spectrum of cosmic rays at supernova envelope fronts ( see e.g. mckee and cowie @xcite , tenorio - tagle et al .
@xcite , stone and norman @xcite , jun & jones @xcite ) .
other challenges are the magnetic field and the instabilities .
we mention two key phenomena : first , the importance of the development of kelvin - helmholtz and rayleigh - taylor instabilities ahead of the thin shock front .
the second is the dual effect that the shock front has on the ism initial magnetic field , first through the compression of the swept - up matter containing the field and secondly the amplification of the radial magnetic field component due to the rayleigh - taylor instability .
simulations of both effects ( see e.g. jun and jones @xcite and references therein ) , modeling the synchrotron radio emission for an expanding supernova shell at various initial magnetic field and ism parameter values , shows for example that the presence of an initial tangential magnetic field component may essentially affect the resulting magnetic field configuration and hence the outgoing radio flux and spectrum . among the additional effects to be taken into account
are the initial inhomogeneity of the ism and the contribution of magnetohydrodynamic turbulence .
in our approach we focus uniquely on the x- and gamma ray radiation , which appears to be conceptually much simpler than the optical and radio emission .
it is perfectly predictable by a set of constitutive equations ( see next section ) , which leads to directly verifiable and very stable features in the spectral distribution of the observed grb afterglows . in line with the observations of grb 991216 and other grb sources , we assume in the following that the x- and gamma ray luminosity represents approximately 90% of the energy flux of the afterglow , while the optical and radio emission represents only the remaining 10% .
this approach differs significantly from the other ones in the current literature , where attempts are made to explain at once all the multi - wavelength emission in the radio , optical , x and gamma ray as coming from a common origin which is linked to boosted synchrotron emission .
such an approach has been shown to have a variety of difficulties ( ghirlanda et al .
@xcite , preece et al .
@xcite ) and can not anyway have the instantaneous variability needed to explain the structure in the `` prompt radiation '' in an external shock scenario , which is indeed confirmed by our model . here the fundamental new assumption is adopted ( see also ruffini et al .
@xcite ) that the x- and gamma ray radiation during the entire afterglow phase has a thermal spectrum in the co - moving frame .
the temperature is then given by : @xmath157^{1/4}\ , , \label{tdir}\ ] ] where @xmath158 is the internal energy developed in the collision with the ism in a time interval @xmath159 in the co - moving frame , @xmath160 is the stefan - boltzmann constant and @xmath161 is the ratio between the `` effective emitting area '' of the afterglow and the surface area of radius @xmath69 . in grb 991216 such a factor
is observed to be decreasing during the afterglow between : @xmath162 ( ruffini et al .
@xcite ) .
the temperature in the comoving frame corresponding to the density distribution described in ruffini et al .
@xcite is shown in fig .
[ tcom_fig ] . .
the dashed line corresponds to an homogeneous distribution with @xmath163 .
details in ruffini et al .
] we are now ready to evaluate the source luminosity in a given energy band .
the source luminosity at a detector arrival time @xmath61 , per unit solid angle @xmath164 and in the energy band @xmath165 $ ] is given by ( see ruffini et al .
@xcite ) : @xmath166}}{dt_a^d d \omega } = \int_{eqts } \frac{\delta \varepsilon}{4 \pi } \ ; v \ ; \cos \vartheta \ ; \lambda^{-4 } \ ; \frac{dt}{dt_a^d } w\left(\nu_1,\nu_2,t_{arr}\right ) d \sigma\ , , \label{fluxarrnu}\ ] ] where @xmath167 is the energy density released in the interaction of the abm pulse with the ism inhomogeneities measured in the comoving frame , @xmath168 is the doppler factor , @xmath169 is an `` effective weight '' required to evaluate only the contributions in the energy band @xmath165 $ ] , @xmath170 is the surface element of the eqts at detector arrival time @xmath61 on which the integration is performed ( see also ruffini et al .
@xcite ) and @xmath171 is the observed temperature of the radiation emitted from @xmath170 : @xmath172\ , .
\label{tarr}\ ] ] the `` effective weight '' @xmath169 is given by the ratio of the integral over the given energy band of a planckian distribution at a temperature @xmath171 to the total integral @xmath173 : @xmath174 where @xmath175 is the planckian distribution at temperature @xmath171 : @xmath176
we turn now to the much debated issue of the origin of the observed hard - to - soft spectral transition during the grb observations ( see e.g. frontera et al .
@xcite , ghirlanda et al .
@xcite , piran @xcite , piro et al .
we consider the instantaneous spectral distribution of the observed radiation for three different eqtss : * @xmath177 s , in the early radiation phase near the peak of the luminosity , * @xmath178 s , in the last observation of the afterglow by the chandra satellite , and * @xmath179 s , chosen in between the other two ( see fig . [ spectrum ] ) .
the observed hard - to - soft spectral transition is then explained and traced back to : 1 . a time decreasing temperature of the thermal spectrum measured in the comoving frame , 2 .
the grb equations of motion , 3 .
the corresponding infinite set of relativistic transformations .
a clear signature of our model is the existence of a common low - energy behavior of the instantaneous spectrum represented by a power - law with index @xmath180 .
this prediction will be possibly verified in future observations .
s , @xmath179 s and @xmath181 s. these diagrams have been computed assuming a constant @xmath182 particle/@xmath144 and clearly explains the often quoted hard - to - soft spectral evolution in grbs .
details in ruffini et al .
] starting from these instantaneous values , we integrate the spectra in arrival time obtaining what is usually fit in the literature by the `` band relation '' ( band et al .
indeed we find for our integrated spectra a low energy spectral index @xmath183 and an high energy spectral index @xmath184 when interpreted within the framework of a band relation ( see fig .
[ spectband ] ) .
this theoretical result can be submitted to a direct confrontation with the observations of grb 991216 and , most importantly , the entire theoretical framework which we have developed can now be applied to any grb source .
the theoretical predictions on the luminosity in fixed energy bands so obtained can be then straightforwardly confronted with the observational data .
kev is fit by a power - law with index @xmath185 and the high energy part above @xmath186 kev is fit by a power - law with an index @xmath184
. details in ruffini et al .
having outlined the main features of our model and shown its application to grb 991216 used as a prototype , before addressing the two new sources which are going to be the focus of this presentation , we recall the three paradigms for the interpretation of grbs we had previously introduced . the first paradigm , the relative space - time transformation ( rstt ) paradigm ( ruffini et al .
@xcite ) emphasizes the importance of a global analysis of the grb phenomenon encompassing both the optically thick and the afterglow phases . since all the data are received in the detector arrival time it is essential to know the equations of motion of all relativistic phases with @xmath187 of the grb sources in order to reconstruct the time coordinate in the laboratory frame , see eq.([tadef ] ) .
contrary to other phenomena in nonrelativistic physics or astrophysics , where every phase can be examined separately from the others , in the case of grbs all the phases are inter - related by their signals received in arrival time @xmath61 .
there is the need , in order to describe the physics of the source , to derive the laboratory time @xmath14 as a function of the arrival time @xmath61 along the entire past worldline of the source using eq.([taddef ] ) . the second paradigm , the interpretation of the burst structure ( ibs ) paradigm ( ruffini et al .
@xcite ) covers three fundamental issues : + a ) the existence , in the general grb , of two different components : the p - grb and the afterglow related by precise equations determining their relative amplitude and temporal sequence ( see ruffini et al .
@xcite ) ; + b ) what in the literature has been addressed as the `` prompt emission '' and considered as a burst , in our model is not a burst at all instead it is just the emission from the peak of the afterglow ( see fig .
[ grb991216 ] ) ; + c ) the crucial role of the parameter @xmath53 in determining the relative amplitude of the p - grb to the afterglow and discriminating between the short and the long bursts ( see fig . [ bcross ] ) .
both short and long bursts arise from the same physical phenomena : the dyadosphere .
the absence of baryonic matter in the remnant leads to the short bursts and no afterglow .
the presence of baryonic matter with @xmath52 leads to the afterglow and consequently to its peak emission which gives origin to the so - called long bursts . ) , are plotted as functions of the @xmath53 parameter . ]
the third paradigm , the grb - supernova time sequence ( gsts ) paradigm ( ruffini et al .
@xcite ) , deals with the relation of the grb and the associated supernova process , and acquires a special meaning in relation to the sources grb 980425 and grb 030329 as we will show in the following .
we now shortly illustrate some consequences of these three paradigms .
the order of magnitude estimate usually quoted for the the characteristic time scale to be expected for a burst emitted by a grb at the moment of transparency at the end of the expansion of the optically thick phase is given by @xmath188 , which for a @xmath189 black hole will give @xmath190 s. there are reasons today not to take seriously such an order of magnitude estimate ( see e.g. ruffini et al .
@xcite ) . in any case
this time is much shorter then the ones typically observed in `` prompt radiation '' of the long bursts , from a few seconds all the way to @xmath4 s. in the current literature ( see e.g. piran @xcite and references therein ) , in order to explain the `` prompt radiation '' and overcome the above difficulty it has been generally assumed that its origin should be related to a prolonged `` inner engine '' activity preceding the afterglow which is not well identified .
to us this explanation has always appeared logically inconsistent since there remain to be explained not one but two very different mechanisms , independent of each other , of similar and extremely large energetics .
this approach has generated an additional very negative result : it has distracted everybody working in the field from the earlier very interesting work on the optically thick phase of grbs .
the way out of this dichotomy in our model is drastically different : 1 ) indeed the optically thick phase exists , is crucial to the grb phenomenon and terminates with a burst : the p - grb ; 2 ) the `` prompt radiation '' follows the p - grb ; 3 ) the `` prompt radiation '' is not a burst : it is actually the temporally extended peak emission of the afterglow ( e - ape ) .
the observed structures of the prompt radiation can all be traced back to inhomogeneities in the interstellar medium ( see fig .
[ grb991216 ] and ruffini et al .
@xcite ) .
the fundamental diagram determining the relative intensity of the p - grb and the afterglow as a function of the dimensionless parameter @xmath53 has been shown in fig .
[ bcross ] .
the underlying machine generating the short and the long grbs is identical : in both cases is the dyadosphere .
the main difference relates to the amount of baryonic matter engulfed by the electron - positron plasma in their optically thick phase prior to transparency . in the limit of small
@xmath191 the intensity of the p - grb is larger and dominates the afterglow .
this corresponds to the short bursts . for @xmath192
the afterglow dominates the grbs and we have the so - called `` long bursts '' . for @xmath193 we may observe a third class of `` bursts '' , eventually related to a turbulent process occurring prior to transparency ( ruffini et al .
this third family should be characterized by smaller values of the lorentz gamma factors than in the case of the short or long bursts .
the relation between the grbs and the supernovae is one of the most complex aspects to be addressed by our model , which needs the understanding of new fields of general relativistic physics in relation to yet unexplored many - body solutions in a substantially new astrophysical scenario . as we will show in the two systems grb980425/sn1998bw and grb030329/sn2003dh which we are going to discuss next , there is in each one the possibility of an astrophysical `` triptych '' ] formed by : + 1 ) the formation of the black hole and the emission of the grb , + 2 ) the gravitational collapse of an evolved companion star , leading to a supernova , + 3 ) a clearly identified urca source whose nature appears to be of the greatest interest .
this new astrophysical scenario presents new challenges : + a ) the identification of the physical reasons of the instability leading to the gravitational collapse of a @xmath194 star , giving origin to the black hole .
such an implosion must occur radially with negligible mass of the remnant ( @xmath52 ) .
+ b ) the identification of the physical reasons for the instability leading to the gravitational collapse of an evolved companion star , giving origin to the supernova .
+ c ) the theoretical issues related to the urca sources , which range today in many possible directions : from the physics of black holes , to the physical processes occurring in the expanding supernova remnants , and finally to the very exciting possibility that we are observing for the first time a newly born neutron star .
the main effort in the next sections is to show that the detailed understanding we have reached for the grb phenomenon and its afterglow allows us to state , convincingly , that the urca source , contrary to what established in the current literature , is not part of the grb nor of its afterglow .
we will draw in the conclusions some considerations on the possible nature of the urca sources .
we illustrate the application of our grb model to two different systems , which are quite different in the energetics but are both related to supernovae : grb 980425 and grb 030329
. we will let the gradual theoretical understanding of the system to unveil the underlying astrophysical scenario .
approaches in the current literature have always attempted to explain both the supernova and the grb as two aspects of a single phenomenon assuming that the grb takes his origin from a specially strong and yet unobserved supernova process : a hypernova ( see paczyski @xcite , kulkarni @xcite , iwamoto @xcite ) . we have taken a very different approach , following cicero s classic aphorism `` divide et impera '' , which was adopted as the motto of the roman empire : `` divide and conquer '' . in this specific case of grbs , which are indeed a very complex system
, we plan to divide and identify the truly independent physical constituents and conquer the understanding of the underlying astrophysical process .
as we will see , this approach will lead to an unexpected and much richer scenario .
in addition to the source grb 980425 and the supernova sn1998bw , two x - ray sources have been found by bepposax in the error box for the location of grb 980425 : a source _ s1 _ and a source _ s2 _ ( pian et al .
@xcite ) , which have been traditionally interpreted either as a background source or as a part of the grb afterglow . see fig .
our approach has been : to first comprehend the entire afterglow of grb 980425 within our theory .
this allows the computation of the luminosity in given energy bands , the spectra , the lorentz gamma factors , and more generally of all the dynamical aspects of the source .
having characterized the features of grb 980425 , we can gradually approach the remaining part of the scenario , disentangling the grb observations from those of the supernova and then disentangling both the grb and the supernova observations from those of the sources _
s1 _ and _ s2_. this leads to a natural identification of distinct events and to their autonomous astrophysical characterization .
our best fit for grb 980425 corresponds to @xmath195 , @xmath196 and the ism average density is found to be @xmath197 .
the plasma temperature and the total number of pairs in the dyadosphere are respectively @xmath198 and @xmath199 .
the light curve of the grb is shown in figs .
[ d15][d17 ] .
the p - grb is under the threshold and in the case of this source is not observable ( see ruffini et al .
@xcite , fraschetti et al .
@xcite ) .
@xmath200 kev energy band is fitted by our model . ]
the characteristic parameter @xmath201 , defining the filamentary structure of the ism , monotonically decreases from @xmath202 to @xmath203 ) .
the results are given in fig .
[ d19 ] where the bolometric luminosity is represented together with the optical data of sn1998bw , the source _
s1 _ and the source _ s2_. it is then clear that grb 980425 is separated both from the supernova data and from the sources _
s2_. @xmath204 kev band of s1 and s2 ( pian et al .
@xcite ) and the optical data of sn 1998bw ( iwamoto @xcite ) . here
we also report some very qualitative curves to be expected for the urca-1 luminosity .
details in ruffini et al .
] while the occurrence of the supernova in relation to the grb has already been discussed within the grb - supernova time sequence ( gsts ) paradigm ( ruffini et al .
@xcite ) , we like to address here a different fundamental issue : the nature of the source _ s1 _ which we have named , in celebration of the work of gamow and shoenberg , urca-1 .
it is clear , from the theoretical predictions of the afterglow luminosity , that the urca-1 can not be part of the afterglow ( see figs .
[ d15 ] , [ d19 ] ) .
there are three different possibilities for the explanation of such source : + 1 ) its possible relation to the black hole formed during the process of gravitational collapse leading to the grb emission .
+ 2 ) its possible relation to emission originating in the early phases of the expansion of the supernova remnant .
+ 3 ) the very exciting possibility that for the first time we are observing a newly born neutron star out of the supernova phenomenon .
while some general considerations will be discussed in the conclusions , we would like to stress here the paramount importance of following the further time history of urca-1 and of the source _
s2_. if , as we propose , _ s2 _ is a background source , its flux should be practically constant in time and this source has nothing to do with the grb 980425 / sn1998bw system .
the drastic behavior of the urca-1 luminosity reported in the talk by elena pian in this meeting , showing the latest urca-1 observations by the xmm and chandra satellites , is crucial for the understanding of the nature of this source .
some very qualitative luminosity curves are sketched in fig .
[ d19 ] , illustrating the possible time evolution of urca-1 .
they are still very undetermined today due to a lack of attention to these observational data and , consequently , to the lack of a detailed theoretical model of the phenomenon .
we therefore propose to have a dedicated attention to the astrophysical `` triptych '' grb 980425 / sn 1998bw / urca-1 .
we have adopted for our modeling of grb030329 a spherically symmetric distribution for the source and , as initial conditions at @xmath205 s , an @xmath206-@xmath207-photon neutral plasma lying between the radii @xmath208 cm and @xmath209 cm .
the temperature of such a plasma is 2.1 mev , the total energy @xmath210 erg and the total number of pairs @xmath211 .
the baryonic matter component @xmath51 is the second free parameter of our theory : @xmath212 . at the emission of the p - grb ,
the lorentz gamma factor is @xmath213 and the radial coordinate is @xmath214 cm .
the ism average density is best fit by @xmath215 particle/@xmath144 .
the third free parameter of our theory is given by @xmath216 .
we then obtain ( see also bernardini et al .
@xcite ) for the grb 030329 the luminosities in given energy bands , computed in the range @xmath217@xmath218 kev with very high accuracy .
[ 030329global][fig1 ] shows the results for the luminosities in the @xmath219@xmath218 kev and @xmath217@xmath204 kev bands , including the `` prompt radiation '' .
subsequently , the theoretically predicted grb spectra have been evaluated at selected values of the arrival time ( ruffini et al .
@xcite ) .
@xmath204 kev and in the @xmath219@xmath218 kev energy bands predicted by our model are fitted to the data of r - xte ( gcn circ .
1996 @xcite ) and hete-2 ( gcn circ .
1997 @xcite ) respectively .
the sn 2003dh optical luminosity is given by the crosses ( hjorth et al .
@xcite ) . details in bernardini et al .
] @xmath204 kev energy bands .
the data refers to the r - xte observations ( gcn circ .
1996 @xcite ) .
details in bernardini et al .
] -rays ( 30 - 400 kev ) with the horizontal bar corresponding to the mean peak flux from hete-2 ( gcn circ .
1997 @xcite ) .
the solid line represents the corresponding one in x - rays ( 2 - 10 kev ) with the experimental data obtained by r - xte ( gcn circ .
1996 @xcite ) .
the remaining points refer respectively to the optical vlt data ( hjorth et al .
@xcite ) of sn2003bw and to the x - ray xmm data ( tiengo et al .
@xcite ) of urca-2 .
the dash - dotted lines corresponds to qualitative luminosity curves expected for urca-2 .
it is interesting to compare and contrast these results with the ones for grb980425/sn1998bw ( see fig . 3 in ruffini et al .
@xcite ) . details in ruffini et al . @xcite . ]
the splendid news received the evening before the presentation of this talk is graphically represented by the xmm observations shown in fig .
again , the xmm observations , like the corresponding ones of grb 980425 , occur after the decaying part of the afterglow and , in analogy to the one occurring in the system grb 980425 / sn 1998bw / urca-1 , we call this source urca-2
. further observations by xmm are highly recommended to follow the urca-2 temporal evolution .
also in this system we are dealing with an astrophysical `` triptych '' : grb 030329 / sn 2003dh / urca-2 .
by the analysis of the first and second batse catalogs tavani in 1998 @xcite ( see fig .
[ tavani ] ) confirmed the previous results by kouveliotou et al .
@xcite on the existence of two families of grbs : the so - called `` long - bursts '' with a soft spectrum and duration @xmath220sec and the `` short - bursts '' with harder spectrum and duration @xmath221sec . in 2001
we have proposed the theory @xcite that both short - bursts and long - bursts originate from the same underlying physical process : the vacuum polarization of electromagnetic overcritical gravitational collapse leading to the creation of @xmath222 pairs at the expenses of the `` blackholic '' energy @xcite .
the difference between the short - bursts and long - bursts in this theory is mainly due to the amount of baryonic matter , described by the dimensionless parameter @xmath53 previously mentioned , encountered by the @xmath223 pairs in their relativistic expansion ( see fig .
[ longshort ] ) .
short - bursts occur in a range of b : ) in the limiting case of no baryonic matter present ( @xmath224 ) .
the optically thick electron - positron pairs expand to ultrarelativistic lorentz gamma factors reaching , then , transparency .
the emission at the transparency point originates the short grbs . ]
@xmath225 and the long - bursts occur for : @xmath226 compare and contrast fig .
[ cip_tot ] with fig .
[ longshort ] .
an indirect support of our theory was given by schmidt @xcite who has shown that short - bursts and long - bursts have the same isotropic - equivalent characteristic peak luminosity . in recent work
we have systematically developed the theoretical background of the process of gravitational collapse of matter involving an electromagnetic field with field strength higher than the critical value for @xmath223 pair creation @xcite .
the goal has been to clarify the physical nature of the process of extracting the `` blackholic '' energy by the creation of @xmath223 matter pairs @xcite and to analyze the electromagnetic radiation emission process during the transient dynamical phases of the gravitational collapse , leading to the final formation of the black hole .
all the considerations presented in the description of the long grbs were based on a dyadosphere of an already formed black hole , presented in section [ dyadosphere ] .
this approximate treatment is very satisfactory in estimating the general dependence of the energy of the p - grb , the kinetic energy of the baryonic matter pulse generating the afterglow and consequently the intensity of the afterglow itself .
in particular it is possible to obtain the overall time structure of the grb and especially the time of the release of the p - grb in respect to the moment of gravitational collapse and its relative intensity with respect to the afterglow .
if , however , we address the issue of the detailed temporal structure of the p - grb and its detailed spectral distribution , the dynamical considerations on the dyadosphere formation , which we are going to present in the following sections , are needed ( see also @xcite ) . in turn , this detailed analysis is needed in order to describe the general relativistic effects close to the horizon formation . as expressed already in section [ dyadosphere ] , all general relativistic quantum field theory effects are encoded in the fine structure of the p - grb . as emphasized in section [ fp ] , the only way to differentiate between solutions with same @xmath55 but different black hole mass and charge is to observe the p - grbs in the limit @xmath224 , namely , to observe the short grbs ( see fig .
[ longshort ] ) .
while the formation in time of the dyadosphere is the fundamental phenomena we are interested in , we can get an insight on the issue of gravitational collapse of an electrically charged star core studying in details a simplified model , namely a thin shell of charged dust .
in @xcite it is shown that the problem of a collapsing charged shell in general relativity can be reduced to a set of ordinary differential equations .
we reconsider here the following relativistic system : a spherical shell of electrically charged dust which is moving radially in the reissner - nordstrm background of an already formed nonrotating black hole of mass @xmath227 and charge @xmath228 , with @xmath229 .
the world surface spanned by the shell divides the space - time into two regions : an internal one @xmath230 and an external one @xmath231 .
the line element in schwarzschild like coordinate is @xcite @xmath232{l } -f_{+}dt_{+}^{2}+f_{+}^{-1}dr^{2}+r^{2}d\omega^{2}\qquad\text{in }
\mathcal{m}_{+}\\ -f_{-}dt_{-}^{2}+f_{-}^{-1}dr^{2}+r^{2}d\omega^{2}\qquad\text{in } \mathcal{m}_{- } \end{array } \right . , \label{e0}\ ] ] where @xmath233 , @xmath234 and @xmath235 and @xmath236 are the schwarzschild - like time coordinates in @xmath230 and @xmath231 respectively .
@xmath24 is the total mass - energy of the system formed by the shell and the black hole , measured by an observer at rest at infinity and @xmath237 is the total charge : sum of the charge @xmath238 of the shell and the charge @xmath228 of the internal black hole .
indicating by @xmath239 the radius of the shell and by @xmath240 its time coordinate , the equations of motion of the shell become @xcite @xmath241 where @xmath242 is the rest mass of the shell and @xmath243 is its proper time .
eqs.([equya],[equyb ] ) ( together with eq.([e0 ] ) ) completely describe a 5-parameter ( @xmath24 , @xmath25 , @xmath227 , @xmath228 , @xmath242 ) family of solutions of the einstein - maxwell equations .
note that eqs.([equya],[equyb ] ) imply that @xmath244 holds for @xmath245 if @xmath246 and for @xmath247 if @xmath248 . for astrophysical applications @xcite the trajectory of the shell @xmath249
is obtained as a function of the time coordinate @xmath250 relative to the space - time region @xmath231 . in the following we drop the @xmath251 index from @xmath250 . from eqs.([equya],[equyb ] ) we have @xmath252 where @xmath253 since we are interested in an imploding shell , only the minus sign case in ( [ equaisrdlc ] ) will be studied
. we can give the following physical interpretation of @xmath254 .
if @xmath255 , @xmath254 coincides with the lorentz @xmath41 factor of the imploding shell at infinity ; from eq.([equaisrdlc ] ) it satisfies @xmath256 when @xmath257 then there is a _ turning point _ @xmath258 , defined by @xmath259 . in this case @xmath254 coincides with the `` effective potential '' at @xmath258 : @xmath260 the solution of the differential equation ( [ equaisrdlc ] ) is given by : @xmath261 the functional form of the integral ( [ gryd ] ) crucially depends on the degree of the polynomial @xmath262 , which is generically two , but in special cases has lower values . we therefore distinguish the following cases : 1 .
@xmath263 ; @xmath264 ; @xmath265 : @xmath266 is equal to @xmath267 , we simply have @xmath268 2 . @xmath263 ; @xmath269 ; @xmath270 : @xmath266 is a constant , we have @xmath271 .\label{caso1}%\end{aligned}\ ] ] 3 .
@xmath263 ; @xmath272 : @xmath266 is a first order polynomial and @xmath273 \nonumber\\ & -\tfrac{1}{\sqrt{m^{2}-q^{2}}}\left [ r_{+}^{2}\mathrm{arctanh}\left ( \tfrac{r}{r_{+}}\tfrac{\sqrt{\omega^{2}-f}}{\omega_{+}}\right ) \right .
\nonumber\\ & \left .
-r_{-}^{2}\mathrm{arctanh}\left ( \tfrac{r}{r_{-}}\tfrac { \sqrt{\omega^{2}-f}}{\omega_{-}}\right ) \right ] , \label{caso2}%\end{aligned}\ ] ] + where @xmath274 .
@xmath275 : @xmath266 is a second order polynomial and @xmath276 \nonumber\\ & -r_{-}^{2}\log\left [ \tfrac{r\sqrt{\omega^{2}-f}}{r - r_{-}}+\tfrac { r^{2}\left ( \omega^{2}-f\right ) + r_{-}^{2}\omega_{-}^{2}-\left ( \gamma ^{2}-1\right ) \left ( r - r_{-}\right ) ^{2}}{2\left ( r - r_{-}\right ) r\sqrt{\omega^{2}-f}}\right ] \nonumber\\ & -\tfrac{\left [ 2mm_{0}\left ( 2\gamma^{3}-3\gamma\right ) + m_{0}^{2}% + q^{2}-q_{1}^{2}\right ] \sqrt{m^{2}-q^{2}}}{m_{0}\left ( \gamma^{2}-1\right ) ^{3/2}}\log\left [ \tfrac{r\sqrt{\omega^{2}-f}}{m}\right .
\nonumber\\ & \left .
+ \tfrac{2m_{0}\left ( \gamma^{2}-1\right ) r-\left ( m_{0}^{2}+q^{2}-q_{1}^{2}\right ) \gamma+2m_{0}m}{2m_{0}m\sqrt{\gamma^{2}-1}% } \right ] \right\ } .\label{caso3}%\end{aligned}\ ] ] of particular interest is the time varying electric field @xmath277 on the external surface of the shell . in order to study the variability of @xmath278 with time
it is useful to consider in the tridimensional space of parameters @xmath279 the parametric curve @xmath280 . in astrophysical applications
@xcite we are specially interested in the family of solutions such that @xmath281 is 0 when @xmath282 which implies that @xmath283 . in fig .
[ elec3d ] we plot the collapse curves in the plane @xmath284 for different values of the parameter @xmath285 , @xmath286 .
the initial data @xmath287 are chosen so that the integration constant in equation ( [ caso2 ] ) is equal to 0 . in all the cases we can follow the details of the approach to the horizon which is reached in an infinite schwarzschild time
coordinate . in fig .
[ elec3d ] we plot the parametric curves @xmath288 in the space @xmath279 for different values of @xmath289 .
again we can follow the exact asymptotic behavior of the curves @xmath288 , @xmath278 reaching the asymptotic value @xmath290 .
the detailed knowledge of this asymptotic behavior is of great relevance for the observational properties of the black hole formation ( see e.g. @xcite ) . for @xmath291 and for different values of the parameter @xmath289 .
the asymptotic behavior is the clear manifestation of general relativistic effects as the horizon of the black hole is approached . *
right ) * electric field behavior at the surface of the shell for @xmath291 and for different values of the parameter @xmath289 .
the asymptotic behavior is the clear manifestation of general relativistic effects as the horizon of the black hole is approached.,title="fig:",width=321 ] for @xmath291 and for different values of the parameter @xmath289 .
the asymptotic behavior is the clear manifestation of general relativistic effects as the horizon of the black hole is approached . *
right ) * electric field behavior at the surface of the shell for @xmath291 and for different values of the parameter @xmath289 .
the asymptotic behavior is the clear manifestation of general relativistic effects as the horizon of the black hole is approached.,title="fig:",width=321 ] in the case of a shell falling in a flat background ( @xmath292 ) eq.([equya ] ) reduces to @xmath293 introducing the total radial momentum @xmath294 of the shell , we can express the kinetic energy of the shell as measured by static observers in @xmath230 as @xmath295 .
then from equation ( [ equy2 ] ) we have @xmath296 where we choose the positive root solution due to the constraint ( [ constraint ] )
. eq.([eqc ] ) is the _ mass formula _ of the shell , which depends on the time - dependent radial coordinate @xmath239 and kinetic energy @xmath297 .
if @xmath298 , a black hole is formed and we have @xmath299 where @xmath300 and @xmath301 is the radius of external horizon of the black hole . we know from the christodoulou - ruffini black hole mass formula that @xmath302
so it follows that @xmath303 namely that @xmath304 is the sum of only three contributions : the rest mass @xmath242 , the gravitational potential energy and the kinetic energy of the rest mass evaluated at the horizon .
@xmath304 is independent of the electromagnetic energy , a fact noticed by bekenstein @xcite .
we have taken one further step here by identifying the independent physical contributions to @xmath304 .
next we consider the physical interpretation of the electromagnetic term @xmath305 , which can be obtained by evaluating the conserved killing integral @xmath306 where @xmath307 is the space - like hypersurface in @xmath231 described by the equation @xmath308 , with @xmath309 as its surface element vector and where @xmath310 is the energy - momentum tensor of the electromagnetic field .
the quantity in eq.([eqr ] ) differs from the purely electromagnetic energy @xmath311 where @xmath312 is the unit normal to the integration hypersurface and @xmath313 .
this is similar to the analogous situation for the total energy of a static spherical star of energy density @xmath314 within a radius @xmath239 , @xmath315 , which differs from the pure matter energy @xmath316 @xmath317 by the gravitational energy ( see @xcite ) .
therefore the term @xmath318 in the mass formula ( [ eqc ] ) is the _ total _ energy of the electromagnetic field and includes its own gravitational binding energy .
this energy is stored throughout the region @xmath307 , extending from @xmath239 to infinity .
we now turn to the problem of extracting the blackholic energy from a black hole ( see @xcite ) .
we can distinguish between two conceptually physically different processes , depending on whether the electric field strength @xmath319 is smaller or greater than the critical value @xmath320 . here
@xmath321 and @xmath322 are the mass and the charge of the electron .
as already mentioned in this paper an electric field @xmath323 polarizes the vacuum creating electron - positron pairs ( see @xcite ) .
the maximum value @xmath324 of the electric field around a black hole is reached at the horizon .
we then have the following : 1 .
for @xmath325 the leading energy extraction mechanism consists of a sequence of discrete elementary decay processes of a particle into two oppositely charged particles .
the condition @xmath325 implies @xmath326{r}% \tfrac{gm / c^{2}}{\lambda_{\mathrm{c}}}\tfrac{\sqrt{g}m_{e}}{e}\sim 10^{-6}\tfrac{m}{m_{\odot}}\quad\text{if } \tfrac{m}{m_{\odot}}\leq10^{6}\\ 1\quad\quad\quad\quad\quad\quad\quad\quad\text{if } \tfrac{m}{m_{\odot}}>10^{6}%
\end{array } \right . , \label{critical3}%\end{aligned}\ ] ] where @xmath327 is the compton wavelength of the electron .
@xcite and @xcite have defined as the _ effective ergosphere _ the region around a black hole where the energy extraction processes occur .
this region extends from the horizon @xmath31 up to a radius @xmath328 \nonumber\\ & \simeq\tfrac{e}{m_{e}}\tfrac{q}{c^{2}}\,.\label{efferg}%\end{aligned}\ ] ] the energy extraction occurs in a finite number @xmath329 of such discrete elementary processes , each one corresponding to a decrease of the black hole charge .
we have @xmath330 since the total extracted energy is ( see eq.([irrmass ] ) ) @xmath331 , we obtain for the mean energy per accelerated particle @xmath332 @xmath333 which gives @xmath334{r}% \left ( \tfrac{m}{m_{\odot}}\right ) \times10^{21}ev\quad\text{if } \tfrac { m}{m_{\odot}}\leq10^{6}\\ 10^{27}ev\quad\quad\text{if } \tfrac{m}{m_{\odot}}>10^{6}%
\end{array } \right . .
\label{uhecr}%\ ] ] one of the crucial aspects of the energy extraction process from a black hole is its back reaction on the irreducible mass expressed in @xcite .
although the energy extraction processes can occur in the entire effective ergosphere defined by eq .
( [ efferg ] ) , only the limiting processes occurring on the horizon with zero kinetic energy can reach the maximum efficiency while approaching the condition of total reversibility ( see fig . 2 in @xcite for details ) .
the farther from the horizon that a decay occurs , the more it increases the irreducible mass and loses efficiency . only in the complete reversibility limit @xcite
can the energy extraction process from an extreme black hole reach the upper value of @xmath335 of the total black hole energy .
2 . for @xmath336
the leading extraction process is a _ collective _ process based on an electron - positron plasma generated by the vacuum polarization , ( see fig .
[ dya ] ) as discussed in section iii in @xcite the condition @xmath337 implies @xmath338 this vacuum polarization process can occur only for a black hole with mass smaller than @xmath339 .
the electron - positron pairs are now produced in the dyadosphere of the black hole , ( note that the dyadosphere is a subregion of the effective ergosphere ) whose radius @xmath45 is given in eq.([rc ] ) .
we have @xmath340 .
the number of particles created and the total energy stored in dyadosphere are given in eqs.(17,18 ) of ref .
@xcite respectively and we have approximately @xmath341 the mean energy per particle produced in the dyadosphere @xmath342 is then @xmath343 which can be also rewritten as @xmath344 such a process of vacuum polarization , occurring not at the horizon but in the extended dyadosphere region ( @xmath345 ) around a black hole , has been observed to reach the maximum efficiency limit of @xmath335 of the total mass - energy of an extreme black hole ( see e.g. @xcite ) .
the conceptual justification of this result follows from the present work : the @xmath346 creation process occurs at the expense of the coulomb energy given by eq .
( [ eqr ] ) and does not affect the irreducible mass given by eq .
( [ eqm ] ) , which indeed , as we have proved , does not depend of the electromagnetic energy . in this sense ,
@xmath347 and the transformation is fully reversible .
this result will be further validated by the study of the dynamical formation of the dyadosphere , which we have obtained using the present work and @xcite ( see @xcite ) .
let us now compare and contrast these two processes .
we have @xmath348 moreover we see ( eqs . ( [ uhecr ] ) , ( [ grb ] ) ) that @xmath349 is in the range of energies of uhecr , while for @xmath350 and @xmath351 , @xmath352 is in the gamma ray range . in other words ,
the discrete particle decay process involves a small number of particles with ultra high energies ( @xmath353 ) , while vacuum polarization involves a much larger number of particles with lower mean energies ( @xmath354 ) .
having so established and clarified the basic conceptual processes of the energetic of the black hole , we are now ready to approach , using the new analytic solution obtained , the dynamical process of vacuum polarization occurring during the formation of a black hole as qualitatively represented in fig .
[ dyaform ] .
the study of the dyadosphere dynamical formation as well as of the electron - positron plasma dynamical evolution will lead to the first possibility of directly observing the general relativistic effects approaching the black hole horizon .
the aim of this section is to point out how the knowledge obtained from the black hole model is of relevance also for the basic theory of black holes and further how very high precision verification of general relativistic effects in the very strong field near the formation of the horizon should be expected in the near future .
we shall first see how eq.([eqm ] ) for @xmath304 , @xmath355 leads to a deeper physical understanding of the role of the gravitational interaction in the maximum energy extraction process of a black hole .
this formula can also be of assistance in clarifying some long lasting epistemological issue on the role of general relativity , quantum theory and thermodynamics .
it is well known that if a spherically symmetric mass distribution without any electromagnetic structure undergoes free gravitational collapse , its total mass - energy @xmath24 is conserved according to the birkhoff theorem : the increase in the kinetic energy of implosion is balanced by the increase in the gravitational energy of the system . if one considers the possibility that part of the kinetic energy of implosion is extracted then the situation is very different : configurations of smaller mass - energy and greater density can be attained without violating birkhoff theorem .
we illustrate our considerations with two examples : one has found confirmation from astrophysical observations , the other promises to be of relevance for gamma ray bursts ( grbs ) ( see @xcite ) .
concerning the first example , it is well known from the work of @xcite that at the endpoint of thermonuclear evolution , the gravitational collapse of a spherically symmetric star can be stopped by the fermi pressure of the degenerate electron gas ( white dwarf ) .
a configuration of equilibrium can be found all the way up to the critical number of particles @xmath356 where the factor @xmath357 comes from the coefficient @xmath358 of the solution of the lane - emden equation with polytropic index @xmath359 , and @xmath360 is the planck mass , @xmath361 is the nucleon mass and @xmath362 the average number of electrons per nucleon . as the kinetic energy of implosion is carried away by radiation the star settles down to a configuration of mass @xmath363 where the gravitational binding energy @xmath364 can be as high as @xmath365 .
similarly gamov ( see @xcite ) has shown that a gravitational collapse process to still higher densities can be stopped by the fermi pressure of the neutrons ( neutron star ) and oppenheimer @xcite has shown that , if the effects of strong interactions are neglected , a configuration of equilibrium exists also in this case all the way up to a critical number of particles @xmath366 where the factor @xmath367 comes now from the integration of the tolman - oppenheimer - volkoff equation ( see e.g. @xcite ) . if the kinetic energy of implosion is again carried away by radiation of photons or neutrinos and antineutrinos the final configuration is characterized by the formula ( [ be ] ) with @xmath368 .
these considerations and the existence of such large values of the gravitational binding energy have been at the heart of the explanation of astrophysical phenomena such as red - giant stars and supernovae : the corresponding measurements of the masses of neutron stars and white dwarfs have been carried out with unprecedented accuracy in binary systems @xcite . from a theoretical physics point of view
it is still an open question how far such a sequence can go : using causality nonviolating interactions , can one find a sequence of braking and energy extraction processes by which the density and the gravitational binding energy can increase indefinitely and the mass - energy of the collapsed object be reduced at will ?
this question can also be formulated in the mass - formula language of a black hole given in @xcite ( see also @xcite ) : given a collapsing core of nucleons with a given rest mass - energy @xmath242 , what is the minimum irreducible mass of the black hole which is formed ? following @xcite and @xcite , consider a spherical shell of rest mass @xmath242 collapsing in a flat space - time . in the neutral case
the irreducible mass of the final black hole satisfies the equation ( see @xcite ) @xmath369 where @xmath24 is the total energy of the collapsing shell and @xmath250 the kinetic energy at the horizon @xmath31 . recall that the area @xmath370 of the horizon is @xcite @xmath371 where @xmath372 is the horizon radius .
the minimum irreducible mass @xmath373 is obtained when the kinetic energy at the horizon @xmath250 is @xmath267 , that is when the entire kinetic energy @xmath250 has been extracted .
we then obtain the simple result @xmath374 we conclude that in the gravitational collapse of a spherical shell of rest mass @xmath242 at rest at infinity ( initial energy @xmath375 ) , an energy up to @xmath335 of @xmath376 can in principle be extracted , by braking processes of the kinetic energy . in this limiting case
the shell crosses the horizon with @xmath377 .
the limit @xmath378 in the extractable kinetic energy can further increase if the collapsing shell is endowed with kinetic energy at infinity , since all that kinetic energy is in principle extractable . in order to illustrate the physical reasons for this result , using the formulas of @xcite
, we have represented in fig .
[ fig1l2 ] the world lines of spherical shells of the same rest mass @xmath242 , starting their gravitational collapse at rest at selected radii @xmath258 .
these initial conditions can be implemented by performing suitable braking of the collapsing shell and concurrent kinetic energy extraction processes at progressively smaller radii ( see also fig . [ fig3l2 ] ) .
the reason for the existence of the minimum ( [ mirrmin ] ) in the black hole mass is the `` self closure '' occurring by the formation of a horizon in the initial configuration ( thick line in fig .
[ fig1l2 ] ) .
starting at rest at selected radii @xmath258 computed by using the exact solutions given in @xcite .
a different value of @xmath304 ( and therefore of @xmath31 ) corresponds to each curve .
the time parameter is the schwarzschild time coordinate @xmath14 and the asymptotic behavior at the respective horizons is evident .
the limiting configuration @xmath379 ( solid line ) corresponds to the case in which the shell is trapped , at the very beginning of its motion , by the formation of the horizon.,width=377 ] is the limit @xmath380 actually attainable without violating causality ?
let us consider a collapsing shell with charge @xmath25 . if @xmath298 a black hole is formed . as pointed out in @xcite the irreducible mass of the final black hole does not depend on the charge @xmath25
. therefore eqs.([mirr2 ] ) and ( [ mirrmin ] ) still hold in the charged case with @xmath301 . in fig .
[ fig3l2 ] we consider the special case in which the shell is initially at rest at infinity , i.e. has initial energy @xmath375 , for three different values of the charge @xmath25 .
we plot the initial energy @xmath381 , the energy of the system when all the kinetic energy of implosion has been extracted as well as the sum of the rest mass energy and the gravitational binding energy @xmath382 of the system ( here @xmath239 is the radius of the shell ) . in the extreme case @xmath383 ,
the shell is in equilibrium at all radii ( see @xcite ) and the kinetic energy is identically zero . in all three cases , the sum of the extractable kinetic energy @xmath297 and the electromagnetic energy @xmath305 reaches @xmath335 of the rest mass energy at the horizon , according to eq.([mirrmin ] ) . , for selected values of the charge . in the first diagram @xmath384 ; the dashed line represents the total energy for a gravitational collapse without any braking process as a function of the radius @xmath239 of the shell ; the solid , stepwise line represents a collapse with suitable braking of the kinetic energy of implosion at selected radii ; the dotted line represents the rest mass energy plus the gravitational binding energy . in the second and third diagram @xmath385 , @xmath386 respectively ; the dashed and the dotted lines have the same meaning as above ; the solid lines represent the total energy minus the kinetic energy .
the region between the solid line and the dotted line corresponds to the stored electromagnetic energy .
the region between the dashed line and the solid line corresponds to the kinetic energy of collapse . in all the cases
the sum of the kinetic energy and the electromagnetic energy at the horizon is 50% of @xmath242 .
both the electromagnetic and the kinetic energy are extractable .
it is most remarkable that the same underlying process occurs in the three cases : the role of the electromagnetic interaction is twofold : a ) to reduce the kinetic energy of implosion by the coulomb repulsion of the shell ; b ) to store such an energy in the region around the black hole . the stored electromagnetic energy is extractable as shown in @xcite.,width=377 ] what is the role of the electromagnetic field here ? if we consider the case of a charged shell with @xmath387 , the electromagnetic repulsion implements the braking process and the extractable energy is entirely stored in the electromagnetic field surrounding the black hole ( see @xcite ) .
in @xcite we have outlined two different processes of electromagnetic energy extraction .
we emphasize here that the extraction of @xmath335 of the mass - energy of a black hole is not specifically linked to the electromagnetic field but depends on three factors : a ) the increase of the gravitational energy during the collapse , b ) the formation of a horizon , c ) the reduction of the kinetic energy of implosion .
such conditions are naturally met during the formation of an extreme black hole but are more general and can indeed occur in a variety of different situations , e.g. during the formation of a schwarzschild black hole by a suitable extraction of the kinetic energy of implosion ( see fig .
[ fig1l2 ] and fig .
[ fig3l2 ] ) . now consider a test particle of mass @xmath388 in the gravitational field of an already formed schwarzschild black hole of mass @xmath24 and go through such a sequence of braking and energy extraction processes .
kaplan @xcite found for the energy @xmath389 of the particle as a function of the radius @xmath69 @xmath390 it would appear from this formula that the entire energy of a particle could be extracted in the limit @xmath391 .
such @xmath392 efficiency of energy extraction has often been quoted as evidence for incompatibility between general relativity and the second principle of thermodynamics ( see @xcite and references therein ) .
j. bekenstein and s. hawking have gone as far as to consider general relativity not to be a complete theory and to conclude that in order to avoid inconsistencies with thermodynamics , the theory should be implemented through a quantum description @xcite .
einstein himself often expressed the opposite point of view ( see e.g. @xcite ) .
the analytic treatment presented in @xcite can clarify this fundamental issue .
it allows to express the energy increase @xmath389 of a black hole of mass @xmath227 through the accretion of a shell of mass @xmath242 starting its motion at rest at a radius @xmath239 in the following formula which generalizes eq.([pointtest ] ) : @xmath393 where @xmath394 is clearly the mass - energy of the final black hole .
this formula differs from the kaplan formula ( [ pointtest ] ) in three respects : a ) it takes into account the increase of the horizon area due to the accretion of the shell ; b ) it shows the role of the gravitational self energy of the imploding shell ; c ) it expresses the combined effects of a ) and b ) in an exact closed formula .
the minimum value @xmath395 of @xmath389 is attained for the minimum value of the radius @xmath396 : the horizon of the final black hole .
this corresponds to the maximum efficiency of the energy extraction .
we have @xmath397 or solving the quadratic equation and choosing the positive solution for physical reasons @xmath398 the corresponding efficiency of energy extraction is @xmath399 which is strictly _ smaller than _
100% for _ any _ given @xmath400 .
it is interesting that this analytic formula , in the limit @xmath401 , properly reproduces the result of equation ( [ mirrmin ] ) , corresponding to an efficiency of @xmath335 . in the opposite limit @xmath402
we have @xmath403 only for @xmath404 , eq.([efficiency ] ) corresponds to an efficiency of 100% and correctly represents the limiting reversible transformations introduced in @xcite .
it seems that the difficulties of reconciling general relativity and thermodynamics are ascribable not to an incompleteness of general relativity but to the use of the kaplan formula in a regime in which it is not valid .
the generalization of the above results to stationary black holes is being considered .
we are now ready to analyze the dynamical properties of an electron positron
photon plasma created by the vacuum polarization process occurring around a charged gravitationally collapsing core of an initially neutral star are examined within the framework of general relativity and quantum field theory . the reissner
nordstrm geometry is assumed to apply between the collapsing core and the oppositely charged remnant of the star .
the appearance of a separatrix at radius @xmath405 , well outside the asymptotic approach to the horizon , is evidenced . the neutral electron positron photon plasma created at radii @xmath406 self - propels outwards to infinity , following the classical pem
pulse analysis @xcite .
the plasma created at @xmath407 remains trapped and follows the gravitational collapse of the core only contributing to the reduction of the electromagnetic energy of the black hole and to the increase of its irreducible mass .
this phenomenon has consequences for the observational properties of gamma
ray bursts and is especially relevant for the theoretical prediction of the temporal and spectral structure of the short bursts .
the formulation of the physics of the _ dyadosphere _ of an electromagnetic black hole ( black hole ) has been until now approached by assuming the vacuum polarization process _
_ sauter heisenberg
schwinger @xcite in the field of an already formed kerr newmann @xcite or reissner nordstrm black hole @xcite .
this acausal approach is certainly valid in order to describe the overall energetics and the time development of the gamma ray bursts ( grbs ) reaching a remarkable agreement between the observations and the theoretical prediction , in particular with respect to : a ) the existence of a proper gamma ray burst ( p grb ) @xcite , b ) the afterglow detailed luminosity function and spectral properties @xcite and c ) the relative intensity of the p - grb to the afterglow @xcite .
this acausal approach has to be improved by taking into account the causal dynamical process of the formation of the dyadosphere as soon as the detailed description on timescales of @xmath408s of the p grb are considered .
such a description leads to theoretical predictions on the time variability of the p
grb spectra which may become soon testable by a new class of specially conceived space missions .
we report progress in this theoretically challenging process which is marked by distinctive and precise quantum and general relativistic effects .
these new results have been made possible by the recent progress in refs .
@xcite , @xcite and especially @xcite .
there it was demonstrated the intrinsic stability of the gravitational amplification of the electromagnetic field at the surface of a charged star core collapsing to a black hole .
the @xmath346 plasma generated by the vacuum polarization process around the core is entangled in the electromagnetic field @xcite .
the @xmath346 pairs do thermalize in an electron positron photon plasma on a time scale @xmath409 times larger than @xmath410 @xcite , where @xmath411 is the speed of light and @xmath321 the electron mass .
as soon as the thermalization has occurred , a dynamical phase of this electrically neutral plasma starts following the considerations already discussed in @xcite . while the temporal evolution of the @xmath412 plasma takes place , the gravitationally collapsing core moves inwards , giving rise to a further amplified supercritical field , which in turn generates a larger amount of @xmath346 pairs leading to a yet higher temperature in the newly formed @xmath413 plasma .
we report , in the following , progress in the understanding of this crucial dynamical process : the main difference from the previous treatments is the fact that we do not consider an already formed black hole but we follow the dynamical phase of the formation of dyadosphere and of the asymptotic approach to the horizon by examining the time varying process at the surface of the gravitationally collapsing core . the space
time external to the surface of the spherically symmetric collapsing core is described by the reissner - nordstrm geometry @xcite with line element @xmath414 with @xmath415 , @xmath416 , where @xmath24 and @xmath25 are the total energy and charge of the core as measured at infinity . on the core surface , which at the time @xmath417 has radial coordinate @xmath418 , the electromagnetic field strength is @xmath419 .
the equation of core s collapse is ( see @xcite ) : @xmath420 where @xmath421 and @xmath242 is the core rest mass .
analytic expressions for the solution of eq.([motion ] ) were given in @xcite .
we here recall that the dyadosphere radius is defined by @xmath422 @xmath423 @xcite as @xmath424 , where @xmath322 is the electron charge . in the following
we assume that the dyadosphere starts to be formed at the instant @xmath425 .
having formulated the core collapse in general relativity in eq.([motion ] ) , in order to describe the quantum phenomena , we consider , at each value of @xmath418 and @xmath417 , a slab of constant coordinate thickness @xmath426 small in comparison with @xmath427 and larger than @xmath428 .
all the results will be shown to be independent on the choice of the value of @xmath426 . in each slab
the process of vacuum polarization leading to @xmath346 pair creation is considered . as shown in @xcite
the pairs created oscillate @xcite with ultrarelativistic velocities and partially annihilate into photons ; the electric field oscillates around zero and the amplitude of such oscillations decreases with a characteristic time of the order of @xmath409 @xmath428 .
the electric field is effectively screened to the critical value @xmath429 and the pairs thermalize to an @xmath413 plasma .
while the average of the electric field @xmath430 over one oscillation is @xmath267 , the average of @xmath431 is of the order of @xmath432 , therefore the energy density in the pairs and photons , as a function of @xmath418 , is given by @xcite @xmath433 = \tfrac{\mathcal{e}% _ { c}^{2}}{8\pi}\left [ \left ( \tfrac{r_{\mathrm{ds}}}{r_{0}}\right ) ^{4}-1\right ] .\label{eps0}%\ ] ] for the number densities of @xmath346 pairs and photons at thermal equilibrium we have @xmath434 ; correspondingly the equilibrium temperature @xmath435 , which is clearly a function of @xmath418 and is different for each slab , is such that @xmath436 with @xmath314 and @xmath437 given by fermi ( bose ) integrals ( with zero chemical potential ) : @xmath438 where @xmath155 is the boltzmann constant . from the conditions set by eqs.([eq0 ] ) , ( [ integrals1 ] ) , ( [ integrals2 ] ) , we can now turn to the dynamical evolution of the @xmath412 plasma in each slab .
we use the covariant conservation of energy momentum and the rate equation for the number of pairs in the reissner
nordstrm geometry external to the star core : @xmath439 , \label{na}%\ ] ] where @xmath440 is the energy
momentum tensor of the plasma with proper energy density @xmath314 and proper pressure @xmath441 , @xmath442 is the fluid @xmath443velocity , @xmath444 is the number of pairs , @xmath445 is the equilibrium number of pairs and @xmath446 is the mean of the product of the @xmath346 annihilation cross - section and the thermal velocity of pairs .
we follow closely the treatment which we developed for the consideration of a plasma generated in the dyadosphere of an already formed black hole @xcite .
it was shown in @xcite that the plasma expands as a pair electromagnetic pulse ( pem pulse ) of constant thickness in the laboratory frame . since the expansion ,
hydrodynamical timescale is much larger than the pair creation ( @xmath428 ) and the thermalization ( @xmath447 ) time - scales , in each slab the plasma remains at thermal equilibrium in the initial phase of the expansion and the right hand side of the rate eq.([na ] ) is effectively @xmath267 , see fig .
24 ( second panel ) of @xcite for details .
if we denote by @xmath448 the static killing vector field normalized at unity at spacial infinity and by @xmath449 the family of space - like hypersurfaces orthogonal to @xmath448 ( @xmath14 being the killing time ) in the reissner
nordstrm geometry , from eqs.([na ] ) , the following integral conservation laws can be derived ( see for instance @xcite ) @xmath450 where @xmath451 is the vector surface element , @xmath389 the total energy and @xmath452 the total number of pairs which remain constant in each slab .
we then have @xmath453 r^{2}=\mathfrak { e},\quad n_{e^{+}e^{-}}\gamma\alpha^{-1}r^{2}=\mathfrak{n}_{e^{+}e^{-}% } , \label{ne}%\ ] ] where @xmath454 and @xmath455 are constants and @xmath456 ^{-1/2}%\ ] ] is the lorentz @xmath41 factor of the slab as measured by static observers .
we can rewrite eqs.([ne ] ) for each slab as @xmath457 where pedex @xmath458 refers to quantities evaluated at selected initial times @xmath459 , having assumed @xmath460 , @xmath461 , @xmath462 .
eq.([eq17 ] ) is only meaningful when @xmath463 . from the structural analysis of such equation
it is clearly identifiable a critical radius @xmath405 such that : * for any slab initially located at @xmath464 we have @xmath465 for any value of @xmath466 and @xmath467 for @xmath468 ; therefore a slab initially located at a radial coordinate @xmath464 moves outwards , * for any slab initially located at @xmath469 we have @xmath465 for any value of @xmath470 and @xmath471 for @xmath472 ; therefore a slab initially located at a radial coordinate @xmath469 moves inwards and is trapped by the gravitational field of the collapsing core .
we define the surface @xmath473 , the _ dyadosphere trapping surface _ ( dts ) .
the radius @xmath405 of dts is generally evaluated by the condition @xmath474 .
@xmath405 is so close to the horizon value @xmath31 that the initial temperature @xmath435 satisfies @xmath475 and we can obtain for @xmath405 an analytical expression .
namely the ultrarelativistic approximation of all fermi integrals , eqs.([integrals1 ] ) and ( [ integrals2 ] ) , is justified and we have @xmath476 and therefore @xmath477 ( @xmath478 ) . the defining equation of @xmath405 , together with ( [ eq19 ] ) , then gives @xmath479 > r_{+}.\ ] ] in the case of a black hole with @xmath291 , @xmath480 , we compute : * the fraction of energy trapped in dts : @xmath481 * the world lines of slabs of plasma for selected @xmath418 in the interval @xmath482 ( see fig .
[ f1 ] ) ; * the world
lines of slabs of plasma for selected @xmath418 in the interval @xmath483 ( see fig .
[ f2 ] ) . at time
@xmath484 when the dts is formed , the plasma extends over a region of space which is almost one order of magnitude larger than the dyadosphere and which we define as the _ effective dyadosphere_. the values of the lorentz @xmath41 factor , the temperature and @xmath346 number density in the effective dyadosphere are given in fig .
[ f3 ] . in conclusion
we see how the causal description of the dyadosphere formation can carry important messages on the time variability and spectral distribution of the p
grb due to quantum effects as well as precise signature of general relativity . ) for a black hole with @xmath291 , @xmath480 ; world lines of slabs of plasma for selected radii @xmath418 in the interval @xmath485 . at time
@xmath486 the expanding plasma extends over a region which is almost one order of magnitude larger than the dyadosphere .
the small rectangle in the right bottom is enlarged in fig .
[ f2].,width=377 ] .
lines of slabs of plasma for selected radii @xmath418 in the interval @xmath487.,width=491 ] factor , proper temperature and proper @xmath346 number density as functions at time @xmath486 for a black hole with @xmath291 and @xmath480.,width=491 ]
we describe the creation and evolution of electron - positron pairs in a strong electric field as well as the pairs annihilation into photons . the formalism is based on generalized vlasov equations , which are numerically integrated .
we recover previous results about the oscillations of the charges , discuss the electric field screening and the relaxation of the system to a thermal equilibrium configuration .
the timescale of the thermalization is estimated to be @xmath488 .
three different earth - bound experiments and one astrophysical observation have been proposed for identifying the polarization of the electronic vacuum due to a supercritical electric field ( @xmath489 , where @xmath321 and @xmath322 are the electron mass and charge ) postulated by sauter - heisenberg - euler - schwinger @xcite : 1 . in central collisions of heavy ions near the coulomb barrier , as first proposed in @xcite ( see also @xcite ) . despite some
apparently encouraging results @xcite , such efforts have failed so far due to the small contact time of the colliding ions @xcite .
typically the electromagnetic energy involved in the collisions of heavy ions with impact parameter @xmath490 cm is @xmath491erg and the lifetime of the diatomic system is @xmath492s .
2 . in collisions of an electron beam with optical laser pulses
: a signal of positrons above background has been observed in collisions of a 46.6 gev electron beam with terawatt pulses of optical laser in an experiment at the final focus test beam at slac @xcite ; it is not clear if this experimental result is an evidence for the vacuum polarization phenomenon .
the energy of the laser pulses was @xmath493erg , concentrated in a space - time region of spacial linear extension ( focal length ) @xmath494 cm and temporal extension ( pulse duration ) @xmath495s @xcite .
3 . at the focus of an x - ray free electron laser ( xfel ) ( see @xcite and references therein ) .
proposals for this experiment exist at the tesla collider at desy and at the lcls facility at slac @xcite .
typically the electromagnetic energy at the focus of an xfel can be @xmath496erg , concentrated in a space - time region of spacial linear extension ( spot radius ) @xmath497 cm and temporal extension ( coherent spike length ) @xmath498s @xcite . and from astrophysics 1 .
around an electromagnetic black hole ( black hole ) @xcite , giving rise to the observed phenomenon of gamma - ray bursts ( grb ) @xcite .
the electromagnetic energy of an black hole of mass @xmath351 and charge @xmath499 is @xmath500erg and it is deposited in a space - time region of spacial linear extension @xmath501 cm @xcite and temporal extension ( collapse time ) @xmath502s @xcite .
in addition to their marked quantitative difference in testing the same basic physical phenomenon , there is a very important conceptual difference among these processes : the first three occur in a transparency condition in which the created electron - positron pairs and , possibly , photons freely propagate to infinity , while the one in the black hole occurs in an opacity condition @xcite . under the opacity condition
a thermalization effect occurs and a final equipartition between the @xmath346 and @xmath41 is reached . far from being just an academic issue , this process and its characteristic timescale is of the greatest importance in physics and astrophysics .
it has been shown by a numerical simulation done in livermore and an analytic work done in rome @xcite , that , as soon as the thermalization of @xmath503 and @xmath41 created around a black hole has been reached , the plasma self propels outwards and this process is at the very heart of the gamma - ray burst ( grb ) phenomenon .
a critical step was missing up to now : how to bridge the gap between the creation of pairs in the supercritical field of the black hole and the thermalization of the system to a plasma configuration .
we report some progress on this topic with special attention to the timescale needed for the thermalization of the newly created @xmath346 pairs in the background field .
the comparison of the thermalization timescale to the one of gravitational collapse , which occurs on general relativistic timescale , is at the very ground of the comprehension of grbs @xcite .
the evolution of a system of particle - antiparticle pairs created by the schwinger process has been often described by a transport vlasov equation ( see , for example , @xcite ) .
more recently it has been showed that such an equation can be derived from quantum field theory @xcite . in the homogeneous case , the equations have been numerically integrated taking into account the back reaction on the external electric field @xcite . in many papers ( see @xcite and references therein ) a phenomenological term describing equilibrating collisions is introduced in the transport equation which is parameterized by an effective relaxation time @xmath243 . in @xcite one further step
is taken by allowing time variability of @xmath243 ; the ignorance on the collision term is then parameterized by a free dimensionless constant .
the introduction of a relaxation time corresponds to the assumption that the system rapidly evolves towards thermal equilibrium . in this paper
we focus on the evolution of a system of @xmath346 pairs , explicitly taking into account the scattering processes @xmath504 .
since we are mainly interested in a system in which the electric field varies on macroscopic length scale ( @xmath505 cm , above ) , we can limit ourselves to a homogeneous electric field .
also , we will use transport equations for electrons , positrons and photons , with collision terms , coupled to maxwell equations .
there is no free parameter here : the collision terms can be exactly computed , since the qed cross sections are known . starting from a regime which is far from thermal equilibrium , we find that collisions do not prevent plasma oscillations in the initial phase of the evolution and analyse the issue of the timescale of the approach to a @xmath412 plasma equilibrium configuration , which is the most relevant quantity in the process of gravitational collapse @xcite .
the motion of positrons ( electrons ) is the resultant of three contributions : the pair creation , the electric acceleration and the annihilation damping .
the homogeneous system consisting of electric field , electrons , positrons and photons can be described by the equations @xmath506 where @xmath507 is the distribution function in the phase - space of positrons ( electrons ) , @xmath508 is the distribution function in the phase - space of photons , @xmath509 is the electric field , @xmath510 is the energy of an electron of 3-momentum @xmath511 ( @xmath321 is the mass of the electron ) and @xmath512 is the energy of a photon of 3-momentum @xmath513 . @xmath514 and @xmath515 are normalized so that @xmath516 , @xmath517 , where @xmath518 and @xmath519 are number densities of positrons ( electrons ) and photons , respectively . the term @xmath520
\delta(p_{\parallel } ) \label{s}%\ ] ] is the schwinger source for pair creation ( see @xcite ) : @xmath521 and @xmath522 are the components of the 3-momentum @xmath511 parallel and orthogonal to @xmath509 .
we assume that the pairs are produced at rest in the direction parallel to the electric field @xcite .
we also have , in eqs .
( [ pairs ] ) , ( [ photons ] ) and ( [ maxwell ] ) , @xmath523 , \label{ce}\\ \mathcal{c}_{\gamma}\left ( t,\mathbf{k}\right ) & \simeq\int\tfrac { d^{3}\mathbf{p}_{1}}{\epsilon_{\mathbf{p}_{1}}}\tfrac{d^{3}\mathbf{p}_{2}% } { \epsilon_{\mathbf{p}_{2}}}\tfrac{d^{3}\mathbf{k}_{1}}{\epsilon _ { \mathbf{k}_{1}}}\delta^{\left ( 4\right ) } \left ( p_{1}+p_{2}% -k - k_{1}\right ) \nonumber\\ \times & \left| \mathcal{m}\right| ^{2}\left [ f_{e}\left ( \mathbf{p}% _ { 1}\right ) f_{e}\left ( \mathbf{p}_{2}\right ) -f_{\gamma}\left ( \mathbf{k}\right ) f_{\gamma}\left ( \mathbf{k}_{1}\right ) \right ] , \label{cf}%\end{aligned}\ ] ] which describe probability rates for pair creation by photons and pair annihilation into photons , @xmath524 being the matrix element for the process @xmath525 . note that the collisional terms ( [ ce ] ) and ( [ cf ] ) are either inapplicable or negligible in the case of the above three earth - bound experiments where the created pairs do not originate a dense plasma . they have been correctly neglected in previous works ( see e. g. @xcite ) .
collisional terms have also been considered in the different physical context of vacuum polarization by strong chromoelectric fields . unlike the present qed case , where expressions for the cross sections are known exactly , in the qcd case the cross sections are yet unknown and such collisional terms are of a phenomenological type and useful uniquely near the equilibrium regime @xcite .
finally @xmath526 and @xmath527 are polarization and conduction current respectively ( see @xcite ) . in eqs .
( [ ce ] ) and ( [ cf ] ) we neglect , as a first approximation , pauli blocking and bose enhancement ( see e.g. @xcite ) . by suitably integrating ( [ pairs ] ) and ( [ photons ] ) over the phase spaces of positrons ( electrons ) and photons , we find the following exact equations for mean values : @xmath528 where , for any function of the momenta @xmath529 furthermore @xmath530 is the relative velocity between electrons and positrons , @xmath531 is the relative velocity between photons , @xmath532 is the total cross section for the process @xmath533 and @xmath534 is the total cross section for the process @xmath535 ( here @xmath536 is the energy of a particle in the reference frame of the center of mass ) . in order to evaluate the mean values in system ( [ system1 ] ) we need some further hypotheses on the distribution functions .
let us define @xmath537 , @xmath538 and @xmath539 such that @xmath540 .
we assume @xmath541 since in the scattering @xmath533 the coincidence of the scattering direction with the incidence direction is statistically favored , we also assume @xmath542 , \label{fgamma}%\ ] ] where @xmath543 and @xmath544 have analogous meaning as @xmath521 and @xmath522 and the terms @xmath545 and @xmath546 account for the probability of producing , respectively , forwardly scattered and backwardly scattered photons . since the schwinger source term ( [ s ] ) implies that the positrons ( electrons ) have initially fixed @xmath521 , @xmath547 , assumption ( [ fe ] ) ( ( [ fgamma ] ) ) means that the distribution of @xmath521 ( @xmath548 ) does not spread too much with time and , analogously , that the distribution of energies is sufficiently peaked to be describable by a @xmath549function .
the dependence on the momentum of the distribution functions has been discussed in @xcite
. approximations ( [ fe ] ) , ( [ fgamma ] ) reduce eqs .
( [ system1 ] ) to a system of ordinary differential equations . in average , since the inertial reference frame we fix coincides with the center of mass frame for the processes @xmath550 , @xmath551 for each species . substituting ( [ fe ] ) and ( [ fgamma ] ) into ( [ system1 ] ) we find @xmath552 where @xmath553 , @xmath554 , @xmath555 are the energy density of positrons ( electrons ) , the energy density of photons and the density of `` parallel momentum '' of positrons ( electrons ) , @xmath430 is the electric field strength and @xmath556 the unique component of @xmath557 parallel to @xmath509 .
@xmath558 and @xmath559 are evaluated at @xmath560 for each species .
note that eqs.([system2 ] ) are `` classical '' in the sense that the only quantum information is encoded in the terms describing pair creation and scattering probabilities .
eqs.([system2 ] ) are consistent with energy density conservation : @xmath561 the initial conditions for eqs.([system2 ] ) are @xmath562 . in fig .
[ fig1luca ] the results of the numerical integration for @xmath563 is showed .
the integration stops at @xmath564 ( where @xmath565 ) .
each variable is represented in units of @xmath321 and @xmath566 .
the numerical integration confirms @xcite that the system undergoes plasma oscillations : a ) the electric field oscillates with decreasing amplitude rather than abruptly reaching the equilibrium value ; b ) electrons and positrons oscillates in the electric field direction , reaching ultrarelativistic velocities ; c ) the role of the @xmath567 @xmath568 scatterings is marginal in the early time of the evolution , the electrons are too extremely relativistic and consequently the density of photons builds up very slowly ( see .
details in fig .
[ fig1luca ] ) . at late times
the system is expected to relax to a plasma configuration of thermal equilibrium and assumptions ( [ fe ] ) and ( [ fgamma ] ) have to be generalized to take into account quantum spreading of the distribution functions .
it is nevertheless interesting to look at the solutions of eqs.([system2 ] ) in this regime . in fig .
[ fig2 ] we plot the numerical solution of eqs.([system2 ] ) but
the integration extends here all the way up to @xmath569 ( the time scale of oscillations is not resolved in these plots ) .
it is interesting that the leading term recovers the expected asymptotic behavior : a ) the electric field is screened to about the critical value : @xmath570 for @xmath571 ; b ) the initial electromagnetic energy density is distributed over electron - positron pairs and photons , indicating energy equipartition ; c ) photons and electron - positron pairs number densities are asymptotically comparable , indicating number equipartition . at such late times
a regime of thermalized electrons - positrons - photons plasma begins and the system is describable by hydrodynamic equations @xcite .
we provided a very simple formalism apt to describe simultaneously the creation of electron - positron pairs by a strong electric field @xmath572 and the pairs annihilation into photons .
as discussed in literature , we find plasma oscillations .
in particular the collisions do not prevent such a feature .
this is because the momentum of electrons ( positrons ) is very high , therefore the cross section for the process @xmath533 is small and the annihilation into photons is negligible in the very first phase of the evolution . as a result , the system takes some time ( @xmath573 ) to thermalize to a @xmath412 plasma equilibrium configuration .
we finally remark that , at least in the case of electromagnetic schwinger mechanism , the picture could be quite different from the one previously depicted in literature , where the system is assumed to thermalize in a very short time ( see @xcite and references therein ) .
it is conceivable that in the race to first identify the vacuum polarization process _
la _ sauter - euler - heisenberg - schwinger , the astrophysical observations will reach a positive result before earth - bound experiments , much like in the case of the discovery of lines in the sun chromosphere by j. n. lockyer in 1869 , later identified with the helium spectral lines by w. ramsay in 1895 @xcite . ,
@xmath574 and plot : a ) electromagnetic field strength ; b ) electrons energy density ; c ) electrons number density ; d ) photons energy density ; e ) photons number density as functions of time.,width=321 ] , @xmath575 and plot : a ) electromagnetic field strength ; b ) electrons energy density ; c ) electrons number density ; d ) photons energy density ; e ) photons number density as functions of time - the oscillation period is not resolved in these plots .
the model used should have a breakdown at a time much earlier than @xmath576 and therefore this plot contains no more than qualitative informations.,width=321 ]
we finally present theoretical predictions for the spectral , temporal and intensity signatures of the electromagnetic radiation emitted during the process of the gravitational collapse of a stellar core to a black hole , during which electromagnetic field strengths rise over the critical value for @xmath223 pair creation .
the last phases of this gravitational collapse are studied , using the result presented in the previous sections , leading to the formation of a black hole with a subcritical electromagnetic field , likely with zero charge , and an outgoing pulse of initially optically thick @xmath223-photon plasma .
such a pulse reaches transparency at lorentz gamma factors of @xmath4@xmath2 .
we find a clear signature in the outgoing electromagnetic signal , drifting from a soft to a hard spectrum , on very precise time - scales and with a very specific intensity modulation .
we outline finally the relevance of these theoretical results for the understanding of short gamma - ray bursts .
the dynamics of the collapse of an electrically - charged stellar core , separating itself from an oppositely charged remnant in an initially neutral star , was first modeled by an exact solution of the einstein - maxwell equations corresponding to a shell of charged matter in ref .
the fundamental dynamical equations and their analytic solutions were obtained , revealing the amplification of the electromagnetic field strength during the process of collapse and the asymptotic approach to the final static configuration .
the results , which properly account for general relativistic effects , are summarized in fig .
1 and fig . 2 of ref .
@xcite .
a first step toward the understanding of the process of extracting energy from a black hole was obtained in ref .
@xcite , where it was shown how the extractable electromagnetic energy is not stored behind the horizon but is actually distributed all around the black hole .
such a stored energy is in principle extractable , very efficiently , on time - scales @xmath577 , by a vacuum polarization process _
l _ sauter - heisenberg - euler - schwinger @xcite .
such a process occurs if the electromagnetic field becomes larger than the critical field strength @xmath429 for @xmath223 pair creation . in ref .
@xcite we followed the approach of damour and ruffini @xcite in order to evaluate the energy density and the temperature of the created @xmath223-photon plasma .
as a byproduct , a formula for the irreducible mass of a black hole was also derived solely in terms of the gravitational , kinetic and rest mass energies of the collapsing core .
this surprising result allowed us in ref .
@xcite to obtain a deeper understanding of the maximum limit for the extractable energy during the process of gravitational collapse , namely 50% of the initial energy of the star : the well known result of a 50% maximum efficiency for energy extraction in the case of a reissner - nordstrm black hole @xcite then becomes a particular case of a process of much more general validity .
the crucial issue of the survival of the electric charge of the collapsing core in the presence of a copious process of @xmath223 pair creation was addressed in refs .
@xcite . by using theoretical techniques borrowed from plasma physics and statistical mechanics @xcite based on a generalized vlasov equation
, it was possible to show that while the core keeps collapsing , the created @xmath223 pairs are entangled in the overcritical electric field . the electric field itself , due to the back reaction of the created @xmath223 pairs , undergoes damped oscillations in sign finally settling down to the critical value @xmath429 .
the pairs fully thermalize to an @xmath223-photon plasma on time - scales typically of the order of @xmath578@xmath579 . during this characteristic damping time , which we recall is much larger than the pair creation time - scale @xmath428 , the core moves inwards , collapsing with a speed @xmath580@xmath581 , further amplifying the electric field strength at its surface and enhancing the pair creation process .
turning now to the dynamical evolution of such an @xmath223 plasma we recall that , after some original attempt to consider a steady state emission @xcite , the crucial progress was represented by the understanding that during the optically thick phase such a plasma expands as a thin shell .
there exists a fundamental relation between the width of the expanding shell and the lorentz gamma factor .
the shell expands , but the lorentz contraction is such that its width in laboratory frame appears to be constant .
such a result was found in @xcite on the basis of a numerical approach , further analyzed in bisnovatyi - kogan and murzina @xcite on the basis of an analytic approach .
attention to the role of the rate equations governing the @xmath223 annihilation were given in @xcite , where approximations to the full equation were introduced .
these results were improved in two important respects in 1999 and 2000 @xcite : the initial conditions were made more accurate by the considerations of the dyadosphere as well as the dynamics of the shell was improved by the self - consistent solution of the hydrodynamical equation and the rate equation for the @xmath223 plasma following both an analytic and numerical approach .
we are now ready to report the result of using the approach in @xcite in this general framework describing the dynamical formation of the dyadosphere . the first attempt to analyze the expansion of the newly generated and thermalized @xmath223-photon plasma was made in ref .
the initial dynamical phases of the expansion were analyzed , using the general relativistic equations of ref .
@xcite for the gravitational collapse of the core .
separatrix _ was found in the motion of the plasma at a critical radius @xmath405 : the plasma created at radii larger than @xmath405 expands to infinity , while the one created at radii smaller than @xmath405 is trapped by the gravitational field of the collapsing core and implodes towards the black hole .
the value of @xmath405 was found in ref .
@xcite to be @xmath582 $ ] , where @xmath24 and @xmath25 are the mass and the charge of the core , respectively .
we now pursue further the evolution of such a system , describing the dynamical phase of the expansion of the pulse of the optically thick plasma all the way to the point where the transparency condition is reached .
some pioneering work in this respect were presented in goodman in 1986 @xcite . in this process
the pulse reaches ultrarelativistic regimes with lorentz factor @xmath583@xmath584 .
the spectra , the luminosities and the time - sequences of the electromagnetic signals captured by a far - away observer are analyzed here in detail for the first time . the relevance of these theoretical results for short - bursts is then discussed .
we discretize the gravitational collapse of a spherically symmetric core of mass @xmath24 and charge @xmath25 by considering a set of events along the world line of a point of fixed angular position on the collapsing core surface . between each of these events
we consider a spherical shell slab of plasma of constant coordinate thickness @xmath426 so that : 1 .
@xmath426 is assumed to be a constant which is small with respect to the core radius ; 2 .
@xmath426 is assumed to be large with respect to the mean free path of the particles so that the statistical description of the @xmath412 plasma can be used ; 3 .
there is no overlap among the slabs and their union describes the entirety of the process .
we check that the final results are independent of the special value of the chosen @xmath426 . in order to describe the dynamics of the expanding plasma pulse the energy - momentum conservation law and the rate equation for the number of pairs in the reissner - nordstrm geometry external to the collapsing core have
to be integrated : @xmath585 , \label{nabis}%\end{aligned}\ ] ] where @xmath586 is the energy - momentum tensor of the plasma with proper energy density @xmath314 and proper pressure @xmath441 , @xmath442 is the fluid 4-velocity , @xmath444 is the pair number density , @xmath587 is the equilibrium pair number density at the temperature @xmath297 of the plasma and @xmath446 is the mean of the product of the @xmath346 annihilation cross - section and the thermal velocity of the pairs .
we use eqs.([tab ] ) and ( [ nabis ] ) to study the expansion of each slab , following closely the treatment developed in refs @xcite where it was shown how a homogeneous slab of plasma expands as a pair - electromagnetic pulse ( pem pulse ) of constant thickness in the laboratory frame .
two regimes can be identified in the expansion of the slabs : 1 . in the initial phase of expansion
the plasma experiences the strong gravitational field of the core and a fully general relativistic description of its motion is needed .
the plasma is sufficiently hot in this first phase that the @xmath346 pairs and the photons remain at thermal equilibrium in it .
as shown in ref .
@xcite , under these circumstances , the right hand side of eqs.([nabis ] ) is effectively @xmath267 and eqs.([tab ] ) and ( [ nabis ] ) are equivalent to : @xmath588{c}% \left ( \tfrac{dr}{cdt}\right ) ^{2}=\alpha^{4}\left [ 1-\left ( \tfrac{n_{e^{+}e^{-}}}{n_{e^{+}e^{-}0}}\right ) ^{2}\left ( \tfrac{\alpha_{0}% } { \alpha}\right ) ^{2}\left ( \tfrac{r}{r_{0}}\right ) ^{4}\right ] , \\
\left ( \tfrac{r}{r_{0}}\right ) ^{2}=\left ( \tfrac{\epsilon+p}{\epsilon_{0}% } \right ) \left ( \tfrac{n_{e^{+}e^{-}0}}{n_{e^{+}e^{-}}}\right ) ^{2}\left ( \tfrac{\alpha}{\alpha_{0}}\right ) ^{2}-\tfrac{p}{\epsilon_{0}}\left ( \tfrac{r}{r_{0}}\right ) ^{4 } , \end{array } \right .
\label{eq2}%\ ] ] where @xmath69 is the radial coordinate of a slab of plasma , @xmath589 @xmath590 is the gravitational redshift factor and the subscript @xmath591 " refers to quantities evaluated at the initial time .
2 . at asymptotically late times
the temperature of the plasma drops below an equivalent energy of @xmath592 mev and the @xmath346 pairs and the photons can no longer be considered to be in equilibrium : the full rate equation for pair annihilation needs to be used .
however , the plasma is so far from the central core that gravitational effects can be neglected . in this new regime , as shown in ref .
@xcite , eqs.([tab ] ) and ( [ nabis ] ) reduce to : @xmath593 , \nonumber\end{aligned}\ ] ] where @xmath594 , @xmath595 is the volume of a single slab as measured in the laboratory frame by an observer at rest with the black hole , @xmath596 is the pair number density as measured in the laboratory frame by an observer at rest with the black hole , and @xmath597 is the equilibrium laboratory pair number density .
eqs.([eq2 ] ) and ( [ eq3 ] ) must be separately integrated and the solutions matched at the transition between the two regimes .
the integration stops when each slab of plasma reaches the optical transparency condition given by @xmath598 where @xmath599 is the thomson cross - section and the integral extends over the radial thickness @xmath426 of the slab .
the evolution of each slab occurs without any collision or interaction with the other slabs ; see the upper diagram in fig .
the outer layers are colder than the inner ones and therefore reach transparency earlier ; see the lower diagram in fig .
[ ff1 ] . and @xmath600 .
upper diagram : world lines of the plasma .
lower diagram : lorentz @xmath41 factor as a function of the radial coordinate @xmath69.,title="fig:",width=377 ] and @xmath600 .
upper diagram : world lines of the plasma .
lower diagram : lorentz @xmath41 factor as a function of the radial coordinate @xmath69.,title="fig:",width=377 ] in fig .
[ ff1 ] , eqs.([eq2 ] ) and ( [ eq3 ] ) have been integrated for a core with @xmath601 the upper diagram represents the world lines of the plasma as functions of the radius , while the lower diagram shows the corresponding lorentz @xmath41 factors .
the overall independence of the result of the dynamics on the number @xmath602 of the slabs adopted in the discretization process or analogously on the value of @xmath426 has also been checked .
we have repeated the integration for @xmath603 , @xmath604 reaching the same result to extremely good accuracy .
the results in fig .
[ ff1 ] correspond to the case @xmath603 . , @xmath600 at @xmath605 as functions of the arrival time @xmath606.,width=377 ] we now turn to the results in fig .
[ ff4 ] , where we plot both the theoretically predicted luminosity @xmath26 and the spectral hardness of the signal reaching a far - away observer as functions of the arrival time @xmath606 .
since all three of these quantities depend in an essential way on the cosmological redshift factor @xmath20 , see refs .
@xcite , we have adopted a cosmological redshift @xmath605 for this figure . as a function of the mass @xmath24 of the core.,width=377 ] as the plasma becomes transparent , gamma ray photons are emitted .
the energy @xmath607 of the observed photon is @xmath608 , where @xmath155 is the boltzmann constant , @xmath297 is the temperature in the comoving frame of the pulse and @xmath41 is the lorentz factor of the plasma at the transparency time .
we also recall that if the initial zero of time is chosen as the time when the first photon is observed , then the arrival time @xmath606 of a photon at the detector in spherical coordinates centered on the black hole is given by @xcite : @xmath609\ ] ] where @xmath610 labels the laboratory emission event along the world line of the emitting slab and @xmath418 is the initial position of the slab .
the projection of the plot in fig .
[ ff4 ] onto the @xmath606-@xmath26 plane gives the total luminosity as the sum of the partial luminosities of the single slabs .
the sudden decrease of the intensity at the time @xmath611 s corresponds to the creation of the _ separatrix _ introduced in ref .
we find that the duration of the electromagnetic signal emitted by the relativistically expanding pulse is given in arrival time by @xmath612 the projection of the plot in fig .
[ ff4 ] onto the @xmath613 , @xmath606 plane describes the temporal evolution of the spectral hardness .
we observe a precise soft - to - hard evolution of the spectrum of the gamma ray signal from @xmath614 kev monotonically increasing to @xmath615 mev .
we recall that @xmath616 .
the above quantities are clearly functions of the cosmological redshift @xmath20 , of the charge @xmath25 and the mass @xmath24 of the collapsing core .
we present in fig .
3 the arrival time interval for @xmath24 ranging from @xmath351 to @xmath617 , keeping @xmath600 .
the arrival time interval is very sensitive to the mass of the black hole : @xmath618 similarly the spectral hardness of the signal is sensitive to the ratio @xmath619 @xcite .
moreover the duration , the spectral hardness and luminosity are all sensitive to the cosmological redshift @xmath20 ( see ref .
all the above quantities can also be sensitive to a possible baryonic contamination of the plasma due to the remnant of the progenitor star which has undergone the process of gravitational collapse .
we first stress some general considerations originating from comparing and contrasting the three grb sources we have discussed : 1 .
the value of the @xmath53 parameter for all three sources occurs , as theoretically expected , in the allowed range ( see fig .
[ bcross ] ) @xmath620 we have in fact : + + [ cols="^,^,^ " , ] 2 .
the enormous difference in the grb energy of the sources simply relates to the electromagnetic energy of the black hole given in eq.([em ] ) which turns out to be smaller than the critical value given by eq.([s1 ] ) .
the fact that the theory is valid over @xmath621 orders of magnitude is indeed very satisfactory .
3 . also revealing is the fact that in both sources grb 980425 and grb 030329 the associated supernova energies are similar .
we have , in fact , for both sn 1998bw and sn 2003dh an energy @xmath622 erg .
details in fraschetti et al .
@xcite and bernardini et al .
the further comparison between the sn luminosity and the grb intensity is crucial . in the case of grb 980425 the grb and the sn energies are comparable , and no dominance of one source over the other can be ascertained . in the case of grb 030329
the energy of the grb source is @xmath5 larger than the sn : in no way the grb can originate from the sn event . the above stringent energetics considerations and the fact that grbs occur also without an observed supernova give a strong evidence that grbs can not originate from supernovae .
we turn now to the most exciting search for the nature of urca-1 and urca-2 .
we have already mentioned above that a variety of possibilities naturally appear .
the first possibility is that the urca sources are related to the black hole originating the grb phenomenon . in order to probe such an hypothesis
, it would be very important to find even a single case in which an urca source occurs in association with a grb and in absence of an associated supernova .
such a result , theoretically unexpected , would open an entire new problematic in relativistic astrophysics and in the physics of black holes . if indeed , as we expect , the clear association between urca sources and the supernovae occurring together with the grbs , then it is clear that the analysis of the other two possibilities will be favored .
namely , an emission from processes occurring in the early phases of the expansion of the supernova remnant or the very exciting possibility that for the first time we are observing a newly born neutron star out of the supernova phenomenon .
of course , this last hypothesis is the most important one , since it would offer new fundamental information about the outcome of the gravitational collapse , about the equations of state at supranuclear densities and about a variety of fundamental issues of relativistic astrophysics of neutron stars .
we shall focus in the following only on this last topic .
we have already recalled how the need for a rapid cooling process due to neutrino anti - neutrino emission in the process of gravitational collapse leading to the formation of a neutron star was considered for the first time by george gamow and mario schoenberg in 1941 @xcite .
it was gamow who gave this process the name `` urca process '' , see [ gamow ] and [ urca ] .
since then , a systematic analysis of the theory of neutron star cooling was advanced by tsuruta @xcite , tsuruta and cameron @xcite , tsuruta et al .
@xcite and by canuto @xcite .
the coming of age of x - ray observatories such as einstein ( 1978 - 1981 ) , exosat ( 1983 - 1986 ) , rosat ( 1990 - 1998 ) , and the contemporary missions of chandra and xmm - newton since 1999 dramatically presented an observational situation establishing very embarrassing and stringent upper limits to the surface temperature of neutron stars in well known historical supernova remnants ( see e.g. romani @xcite ) .
it was so that , for some remnants , notably sn 1006 and the tycho supernova , the upper limits to the surface temperatures were significantly lower than the temperatures given by standard cooling times ( see e.g. romani @xcite ) .
much of the theoretical works has been mainly directed , therefore , to find theoretical arguments in order to explain such low surface temperature @xmath623@xmath624 k embarrassingly low , when compared to the initial hot ( @xmath625 k ) birth of a neutron star in a supernova explosion ( see e.g. romani @xcite ) . some very important steps in this direction of research
have been represented by the works of van riper @xcite , lattimer and his group @xcite and by the most extensive work of yakovlev and his group @xcite .
the youngest neutron star to be searched for using its thermal emission in this context has been the pulsar psr j0205 + 6449 in 3c 58 ( see e.g. yakovlev and pethick @xcite ) , which is @xmath626 years old ! recently , evidence for the detection of thermal emission from the crab nebula pulsar was reported by trumper @xcite which is , again , @xmath627 years old . in the case of urca-1 and urca-2
, we are exploring a totally different regime : the x - ray emission possibly from a recently born neutron star in the first days months of its existence , where no observations have yet been performed and no embarrassing constraints upper limits on the surface temperature exist .
the reason of approaching first the issue of the thermal emission from the neutron star surface is extremely important , since in principle it can give information on the equations of state in the core at supranuclear densities and on the detailed mechanism of the formation of the neutron star itself and the related neutrino emission .
it is of course possible that the neutron star is initially fast rotating and its early emission is dominated by the magnetospheric emission or by accretion processes from the remnant which would overshadow the thermal emission . in that case a periodic signal related to the neutron star rotational period should in principle be observable in a close enough grb source provided the suitable instrumentation from the earth .
the literature on young born neutron star is relatively scarse today .
there are some very interesting contributions which state : `` the time for a neutron star s center to cool by the direct urca process to a temperature @xmath297 has been estimated to be @xmath628^{-4}$ ] s. the direct urca process and all the exotic cooling mechanisms only occur at supranuclear densities .
matter at subnuclear densities in neutron star crust cools primarily by diffusion of heat to the interior .
thus the surface temperature remains high , in the vicinity of @xmath629 k or more , until the crust s heat reservoir is consumed . after this diffusion time , which is on the order of @xmath105@xmath630 years
, the surface temperature abruptly plunges to values below @xmath631 k '' ( lattimer et al .
`` soon after a supernova explosion , the young neutron star has large temperature gradients in the inner part of the crust .
while the powerful neutrino emission quickly cools the core , the crust stays hot .
the heat gradually flows inward on a conduction time scale and the whole process can be thought of as a cooling wave propagation from the center toward the surface '' ( gnedin et al .
@xcite ) .
the two considerations we have quoted above are developed in the case of spherical symmetry and we would like to keep the mind open , in this new astrophysical field , to additional factors , some more traditional than others , to be taken into account . among the traditional ones we recall : 1 ) the presence of rotation and magnetic field which may affect the thermal conductivity and the structure of the surface , as well as the above mentioned magnetospheric emission ; 2 ) there could be accretion of matter from the expanding nebula ; and , among the nontraditional ones , we recall 3 ) some exciting theoretical possibilities advanced by dyson on volcanoes on neutron stars @xcite as well as iron helide on neutron star @xcite , as well as the possibility of piconuclear reactions on neutron star surface discussed in lai & salpeter @xcite .
all the above are just scientific arguments to attract attention on the abrupt fall in luminosity reported in this meeting on urca-1 by elena pian which is therefore , in this light , of the greatest scientific interest and further analysis should be followed to check if a similar behavior will be found in future xmm and chandra observations also in urca-2 .
in addition to these very rich problematics in the field of theoretical physics and theoretical astrophysics , there are also more classical astronomical and astrophysical issues , which will need to be answered if indeed the observations of a young neutron star will be confirmed .
an important issue to be addressed will be how the young neutron star can be observed , escaping from being buried under the expelled matter of the collapsing star .
a possible explanation can originate from the binary nature of the newly born neutron star : the binary system being formed by the newly formed black hole and the triggered gravitational collapse of a companion evolved star leading , possibly , to a `` kick '' on and ejection of the newly born neutron star .
another possibility , also related to the binary nature of the system , is that the supernova progenitor star has been depleted of its outer layer by dynamic tidal effects .
in addition , there are other topics in which our scenario can open new research directions in fundamental physics and astrophysics : + 1 ) the problem of the instability leading to the complete gravitational collapse of a @xmath194 star needs the introduction of a new critical mass for gravitational collapse , which is quite different from the one for white dwarfs and neutron stars which has been widely discussed in the current literature ( see e.g. giacconi & ruffini @xcite ) .
+ 2 ) the issue of the trigger of the instability of gravitational collapse induced by the grb on the progenitor star of the supernova or , vice versa , by the supernova on the progenitor star of the grb needs accurate timing and the considerations of new relativistic phenomena .
+ 3 ) the general relativistic instability induced on a nearby star by the formation of a black hole needs some very basic new developments in the field of general relativity .
only a very preliminary work exists on this subject , by jim wilson and his collaborators , see e.g. the paper by mathews and wilson @xcite .
the reason for the complexity in answering such a question is simply stated : unlike the majority of theoretical work on black holes , which deals mainly with one - body solutions , we have to address here a many - body problem in general relativity .
we are starting in these days to reconsider , in this framework , some classic work by fermi @xcite , hanni and ruffini @xcite , majumdar @xcite , papapetrou @xcite , parker et al .
@xcite , bini et al .
@xcite which may lead to a new understanding of general relativistic effects relevant to these astrophysical `` triptychs '' . after concluding the problematic of the long grbs and their vast astrophysical implications , we have turned to the physics of short grbs .
we first report some progress in the understanding the dynamical phase of collapse , the mass - energy formula and the extraction of blackholic energy which have been motivated by the analysis of the short grbs . in this context progress
has also been accomplished on establishing an absolute lower limit to the irreducible mass of the black hole as well as on some critical considerations about the relations of general relativity and the second law of thermodynamics .
we recall how this last issue has been one of the most debated in theoretical physics in the past thirty years due to the work of bekenstein and hawking . following these conceptual progresses
we analyze the vacuum polarization process around an overcritical collapsing shell .
we evidence the existence of a separatrix and a dyadosphere trapping surface in the dynamics of the electron - positron plasma generated during the process of gravitational collapse .
we then analyze , using recent progress in the solution of the vlasov - boltzmann - maxwell system , the oscillation regime in the created electron - positron plasma and their rapid convergence to a thermalized spectrum .
we conclude by making precise predictions for the spectra , the energy fluxes and characteristic time - scales of the radiation for short - bursts .
the characteristic spectra , time variabilities and luminosities of the electromagnetic signals from collapsing overcritical stellar cores have been derived from first principles , and they agree with preliminary observations of short - bursts @xcite . hopefully new space missions will be planned , with temporal resolution down to fractions of @xmath362s and higher collecting area and spectral resolution than at present , in order to verify the detailed agreement between our model and the observations .
it is now clear that if our theoretical predictions will be confirmed , we will have a very powerful tool for cosmological observations : the independent information about luminosity , time - scale and spectrum can uniquely determine the mass , the electromagnetic structure and the distance from the observer of the collapsing core , see e.g. fig .
[ ff5 ] and ref .
@xcite . in that case
short - bursts , in addition to give a detailed information on all general relativistic and relativistic field theory phenomena occurring in the approach to the horizon , may also become the best example of standard candles in cosmology @xcite .
we are currently analyzing the introduction of baryonic matter in the optically thick phase of the expansion of the @xmath223 plasma , within this detailed time - varying description of the gravitational collapse , which may affect the structure of the proper - grb ( p - grb ) @xcite as well as the structure of the long - bursts @xcite .
an interesting proposal was advanced in 2002 @xcite that the @xmath223 plasma may have a fundamental role as well in the physical process generating jets in the extragalactic radio sources .
the concept of dyadosphere originally introduced in reissner - nordstrm black hole in order to create the @xmath223 plasma relevant for grbs can also be generalized to the process of vacuum polarization originating in a kerr - newman black hole due to magneto - hydrodynamical process of energy extraction ( see e.g. @xcite and references therein ) .
the concept therefore introduced here becomes relevant for both the extraction of rotational and electromagnetic energy from the most general black hole @xcite .
after the completion of these lectures , we have become aware that ghirlanda et al .
@xcite have given evidence for the existence of an exponential cut off at high energies in the spectra of short bursts .
we are currently comparing and contrasting these observational results with the predicted cut off in fig .
[ ff4 ] which results from the existence of the separatrix introduced in @xcite .
the observational confirmation of the results presented in fig . [ ff4 ] would lead for the first time to the identification of a process of gravitational collapse and its general relativistic self - closure as seen from an asymptotic observer .
we are thankful to charles dermer , hagen kleinert , tsvi piran , andreas ringwald , rashid sunyaev , lev titarchuk , jim wilson and dima yakovlev for many interesting theoretical discussions , as well as to lorenzo amati , lucio angelo antonelli , enrico costa , filippo frontera , luciano nicastro , elena pian , luigi piro , marco tavani and all the bepposax team for assistance in the data analysis .
the evolution of the @xmath223-pair plasma generated in the dyadosphere has been treated in two papers @xcite .
we recall here the basic governing equations in the most general case in which the plasma fluid is composed of @xmath223-pairs , photons and baryonic matter .
the plasma is described by the stress - energy tensor @xmath632 where @xmath633 and @xmath441 are respectively the total proper energy density and pressure in the comoving frame of the plasma fluid and @xmath634 is its four - velocity , satisfying @xmath635 where @xmath636 and @xmath637 are the radial and temporal contravariant components of the 4-velocity and @xmath638 where @xmath639 and @xmath640 .
the conservation law for baryon number can be expressed in terms of the proper baryon number density @xmath641 @xmath642 the radial component of the energy - momentum conservation law of the plasma fluid reduces to @xmath643 = 0 ~. \label{cmom2}\ ] ] the component of the energy - momentum conservation law of the plasma fluid equation along a flow line is @xmath644=0 ~. \label{conse1}\end{aligned}\ ] ] we define also the total proper internal energy density @xmath314 and the baryonic mass density @xmath645 in the comoving frame of the plasma fluid , @xmath646 a computer code @xcite has been used to evolve the spherically symmetric general relativistic hydrodynamic equations starting from the dyadosphere @xcite . we define the generalized gamma factor @xmath41 and the radial 3-velocity in the laboratory frame @xmath647 @xmath648 from eqs.([s ] , [ tt ] ) , we then have @xmath649 following eq.([cpp ] ) , we also define @xmath650 so that the conservation law of baryon number ( [ contin ] ) can then be written as @xmath651 eq.([conse1 ] ) then takes the form , @xmath652 .
\label{jay2}\ ] ] defining the radial momentum density in the laboratory frame @xmath653 we can express the radial component of the energy - momentum conservation law given in eq.([cmom2 ] ) by @xmath654\nonumber\\ & = & - { \alpha \over r^2 } { \partial \over \partial r } ( { r^2 \over \alpha } s_r v^r ) - \alpha
{ \partial p \over \partial r}\nonumber\\ & - & \alpha\left({m \over r^2}-{q^2 \over r^3}\right ) \biggl({d + \gamma e \over \gamma } \biggr ) \biggl [ \left({\gamma \over \alpha}\right)^2 + { ( u^r)^2 \over \alpha^4 } \biggr ] ~. \label{jay3}\end{aligned}\ ] ] in order to determine the number - density of @xmath223 pairs , we use the pair rate equation .
we define the @xmath223-pair density in the laboratory frame @xmath655 and @xmath656 , where @xmath657 is the total proper number density of pairs in comoving frame at thermodynamic equilibrium with temperature @xmath297 in the process @xmath658 @xmath659 , @xmath660 is the total proper number density of pairs in comoving frame at a generic time before reaching the equilibrium .
we write the rate equation in the form @xmath661 these equations are integrated starting from the dyadosphere distributions given in fig .
17 ( right ) in @xcite and assuming as usual ingoing boundary conditions on the horizon of the black hole .
after the explosion from the dyadosphere a thermal plasma of @xmath223 pairs and photons optically thick with respect to scattering processes begins to expand at ultrarelativistic velocity . in this era the expansion takes place in a region of very low baryonic contamination .
recalling that the local number density of electron and positron pairs created as a function of radius is given by @xmath662 ~ , \label{nd}\ ] ] the limit on such baryonic contamination , where @xmath663 is the mass - energy density of baryons , is given by @xmath664(g / cm^3 ) .
\label{nb}\ ] ] near the horizon @xmath665 , this gives @xmath666 and near the radius of the dyadosphere @xmath45 : @xmath667_{r\rightarrow r_{ds}}(g / cm^3)\ , .
\label{nb2}\ ] ] such conditions can be easily satisfied in the collapse to a black hole , but not necessarily in a collapse to a neutron star .
consequently we have solved the equations governing a plasma composed solely of @xmath223-pairs and electromagnetic radiation , starting at time zero from the dyadosphere configurations corresponding to constant density in fig .
[ 3dens ] . and
@xmath289 are given .
the three different configurations are markedly different in their spatial extent as well as in their energy - density distribution ( see text)_.,width=340 ] the plasma of @xmath223 pairs and photons is described by the covariant energy - momentum tensor : @xmath668 where @xmath633 and @xmath441 are respectively total proper energy density and pressure in the comoving system ; @xmath669 are contravariant components of 4-velocity and @xmath670 takes into account of dissipative effects due to heat conduction and viscosity , but in this treatment it has been neglected .
in general we have @xmath671 . for a spherically symmetric motion this reduces to @xmath672 , where @xmath637 and @xmath636 are respectively temporal and radial controvariant components of 4-velocity @xmath669 .
it is assumed that the gravitational interaction with central black hole is negligible with respect to the total energy of pem - pulse such that a fluid expansion with special relativistic equations can be considered
. moreover it is assumed that photons remain trapped inside fireball until complete transparency , i.e. the emission of electromagnetic radiation is negligible during the first phases of expansion , being therefore adiabatic @xcite .
this assumption is valid until the photon mean free path is negligible with respect to the thickness of pulse .
the thermodynamic quantities used to describe the process are the total proper internal energy density of pulse @xmath314 , given by @xmath673 , where @xmath674 ( @xmath675 ) is total proper internal energy density of electrons ( positrons ) and @xmath676 of photons .
the proper number density of pairs @xmath660 , if the system is in thermodynamic equilibrium initially at temperature @xmath297 of order @xmath677 , enough for @xmath223 pair creation , equals the proper number density of photons @xmath519 .
this is not valid at lower temperature @xcite .
the pressure is @xmath678 , where @xmath679 are electrons and positrons pressures and @xmath680 is photons pressure .
the system is highly relativistic , so the equation of state @xmath681 can be considered valid .
this equation of state is represented with thermal index @xmath254 : @xmath682 thermodynamical quantities before introduced are expressed in terms of integrals over bose distribution for photons and fermi distribution for @xmath223 pairs with zero chemical potentials @xmath683 and @xmath684 .
we begin from the reaction @xmath658 . from statistical mechanics
it is known that given a thermodynamic system at temperature @xmath297 kept inside a volume @xmath685 and made of a number of particle variable @xmath602 , the thermodynamic equilibrium is expressed by the condition that the potential free energy of helmholtz @xmath686 is stationary with respect to @xmath602 variations : @xmath687 by definition chemical potential @xmath362 is given by @xmath688 so that for a system made by a photon gas at equilibrium with matter with respect to creation and adsorption processes , we have @xmath689 @xcite .
therefore the chemical potential of electrons and positrons associated to reaction @xmath658 is equal and opposite : @xmath690 ; moreover also @xmath362 must be zero since the total electric charge of fireball is zero : if @xmath25 is total electric charge of fireball , we have @xmath691 = 0 \label{charge}\ ] ] where @xmath692 is given by @xmath693 so @xmath694 . in the following the expressions of thermodynamical quantities as fermi integrals are listed .
the proper number density of electrons @xcite is given by @xmath695 where @xmath696 , @xmath388 is the electron mass , @xmath297[mev ] is the temperature of fireball in comoving frame , @xmath697 is a constant given by @xmath698 , @xmath155 is the boltzmann constant and @xmath699 is a numerical constant introduced for convenience . +
since the thermodynamic equilibrium is assumed and in all cases considered the initial temperature is larger than @xmath223 pairs creation threshold ( @xmath700 mev ) , the proper number density of electrons is roughly equal to that one of photons : @xmath701 in these conditions the number of particles is conserved : @xmath702 later on , for @xmath703mev ( see fig .
[ tem2 ] ) , @xmath223 pairs go on in annihilation but can not be created anymore , therefore @xmath704 as shown in fig .
[ pairs_fig ] . _
.,width=377 ] @xmath660 and number density of photons @xmath705 as a function of emission time for different values of black hole mass @xmath362_.,width=377 ] the total proper internal energy density for photons is given by @xmath706 where @xmath707 .
the total proper internal energy density for electrons is given by : @xmath708 where @xmath696 and the integral is computed numerically .
therefore the total proper internal energy density of the pem - pulse , summing up all the contributions of photons and @xmath223 pairs , is given by @xmath709 \label{entot1}\ ] ] where the factor @xmath217 in front of the integral takes into account of electrons and positrons .
+ about the pressure of the photons it holds @xmath710 and about the pressure of electrons @xmath711 therefore the total pressure of pem - pulse is given by @xmath712 .
\label{ptot}\ ] ] in the following we recall a zeroth order approximation of the fully relativistic equations of the previous section @xcite : + ( i ) since we are mainly interested in the expansion of the @xmath223 plasma away from the black hole , we neglect the gravitational interaction .
+ ( ii ) we describe the expanding plasma by a special relativistic set of equations .
+ in the pem - pulse phase the expansion in vacuum is described by a set of equation expressing : * entropy conservation , because of the assumption that emission of electromagnetic radiation is negligible up to transparency ; * energy conservation , because the increase of kinetic energy is compensated by a decrease of total internal energy . for the expansion of a single shell , the adiabaticity is given by @xmath713 where @xmath685 is the volume of the shell in the comoving frame and
@xmath714 is the total proper internal energy of plasma . by using the equation of state [ state ]
we find @xmath715 and , by integrating , we find @xmath716 recalling that the volume of the fireball in the comoving frame is given by @xmath717 , where @xmath718 is the volume in the laboratory frame , we find @xmath719 the total energy conservation of the shell implies @xcite : @xmath720 and this gives the evolution for @xmath721 : @xmath722 substituting this expression for @xmath721 in ( [ scale ] ) the final equation for proper internal energy density is found @xmath723 the evolution of a plasma of @xmath223 pairs and photons should be treated by relativistic hydrodynamics equations describing the variation of the number of particles in the process .
the 4-vector number density of pairs is defined @xmath724 , which in the comoving frame reduces to the 4-vector @xmath725 .
the law of number conservation for pairs is @xmath726 where @xmath727 is the determinant of reissner - nordstrm metric . in the system processes of creation and annihilation of particles
occur due to collisions between particles .
if the number of particles is conserved , it holds @xmath728 ; if instead it is not conserved , in the assumptions that only binary collisions between particles occur and in the hypothesis of molecular caos , the eq .
becomes @xmath729 \label{rate1}\ ] ] where @xmath160 is the cross section for the process of creation and annihilation of pairs , given by @xmath730 , \label{cross}\ ] ] with @xmath731 and @xmath389 total energy of positrons in the laboratory frame , and @xmath732 the classical radius of electron , @xmath733 is the sound velocity in the fireball : @xmath734 and @xmath446 is the mean value of @xmath735 ; for @xmath160 we use as a first approximation the thomson cross section , @xmath736 ; @xmath657 is the total proper number density of electrons and positrons in comoving frame at thermodynamic equilibrium in the process @xmath658 @xmath659 , @xmath660 is the total proper number density of electrons and positrons in comoving frame at a generic time before reaching the equilibrium .
+ using the approximation of special relativity , the 4-velocity is written @xmath737 ; substituting to @xmath738 the @xmath739 and to @xmath740 the @xmath741 , eq . in hybrid form
becomes @xmath742 valid for electrons and positrons .
+ now we have a complete set of equations for numerical integration : ( [ scale2 ] ) , ( [ gammavar ] ) e la ( [ paira ] ) .
if we now turn from a single shell to a finite distribution of shells , we can introduce the average values of the proper internal energy and pair number densities ( @xmath743 ) for the pem - pulse , where the average @xmath721-factor is defined by @xmath744 and @xmath718 is the total volume of the shell in the laboratory frame @xcite .
in principle we could have an infinite number of possible schemes to define geometry of the expanding shell .
three different possible schemes have been proposed @xcite : * sphere .
an expansion with radial component of 4-velocity proportional to the distance to the black hole @xmath745 , where @xmath364 is the radial component of 4-velocity on the external surface of pem - pulse ( having radius @xmath746 ) , the factor @xmath721 from ( [ ga ] ) is @xmath747 ; \label{gammamean_1sch}\ ] ] this distribution corresponds to a uniform and time decreasing density , like in friedmann model for the universe ; * slab 1 .
an expansion with thickness of fireball constant @xmath748 in laboratory frame in which the black hole is at rest , with @xmath749 and @xmath750 ; this distribution does not require an average ; * slab 2 .
an expansion with thickness of fireball constant in comoving frame of pem - pulse .
the result has been compared with the one of hydrodynamic equation in general relativity @xcite ( see fig . [ shells ] ) .
excellent agreement has been found with the scheme in which the thickness of fireball is constant in laboratory frame : what happens is that the thickness in comoving frame increases , but due to the lorentz contraction , it is kept constant in laboratory frame and equal to @xmath751 . in this case
@xmath752 , where @xmath721 is computed by conservation equations . +
a similar situation occurs for the temperature of pem - pulse . in the comoving frame
the temperature decreases as @xmath753 , in accordance with results in literature @xcite .
since @xmath41 monotonically increases as @xmath754 @xcite , in laboratory frame @xmath755 @xcite ; photons are blue - shifted in laboratory frame in such away that , at least in the first phase , the temperature measured by an observer at infinity is constant .
the numerical value of the temperature of equilibrium at each instant is found by imposing the equivalence , within a certain precision , of ( [ entot1 ] ) numerically computed and ( [ scale2 ] ) .
factor as a function of radial coordinate .
three schemes of expansion of pem - pulse ( see text ) are compared with solution of hydrodynamics relativistic equations numerically integrated for a black hole with @xmath756 and @xmath757 .
the result is in accordance with the scheme of a fireball with constant thickness in laboratory frame_.,width=566 ] even if the pem - pulse is optically thick in the expansion before transparency , photons located at a distance from the external surface less their mean free path can escape and reach the observer at infinity .
the mean free path in the comoving frame is given by @xmath758 while in laboratory frame is given by @xmath759 cm .
however the luminosity emitted at this stage is negligible , since the ratio between @xmath760 and the thickness of the fireball @xmath761 in the laboratory frame ( with @xmath762 cm ) is of the order of @xmath763 .
the pem pulse expands initially in a region of very low baryonic contamination created by the process of gravitational collapse . as it moves outside the baryonic remnant of the progenitor star
is swept up .
the existence of such a remnant is necessary in order to guarantee the overall charge neutrality of the system : the collapsing core has the opposite charge of the remnant and the system as a whole is clearly neutral .
the number of extra charges in the baryonic remnant negligibly affects the overall charge neutrality of the pem pulse .
the baryonic matter remnant is assumed to be distributed well outside the dyadosphere in a shell of thickness @xmath764 between an inner radius @xmath765 and an outer radius @xmath766 at a distance from the black hole not so big that the pem pulse expanding in vacuum has not yet reached transparency and not so small that the system will reach enoughly high value of lorentz @xmath41 in order to not be stopped in the collision ( see fig .
[ cip2 ] ) . for the sake of an example we choose @xmath767 the total baryonic mass @xmath768
is assumed to be a fraction of the dyadosphere initial total energy @xmath769 .
the total baryon - number @xmath770 is then expressed as a function of the dimensionless parameter @xmath53 given by @xmath771 where @xmath53 is a parameter in the range @xmath772 and @xmath66 is the proton mass .
we shall see below the role of @xmath53 in the determination of the features of the grbs .
we saw in section [ fp ] the sense in which @xmath53 and @xmath55 can be considered to be the only two free parameters of the black hole theory for the entire grb family , the so called `` long bursts '' .
for the so called `` short bursts '' the black hole theory depends on the two other parameters @xmath362 , @xmath289 , since in that case @xmath773 since most of the energy , unless the whole energy , in the pulse is emitted at transparency .
the baryon number density @xmath774 is assumed to be a constant @xmath775 as the pem pulse reaches the region @xmath776 , it interacts with the baryonic matter which is assumed to be at rest . in our model
we make the following assumptions to describe this interaction : * the pem pulse does not change its geometry during the interaction ; * the collision between the pem pulse and the baryonic matter is assumed to be inelastic , * the baryonic matter reaches thermal equilibrium with the photons and pairs of the pem pulse .
these assumptions are valid if : ( i ) the total energy of the pem pulse is much larger than the total mass - energy of baryonic matter @xmath51 , @xmath777 , ( ii ) the ratio of the comoving number density of pairs and baryons at the moment of collision @xmath778 is very high ( e.g. , @xmath779 ) and ( iii ) the pem pulse has a large value of the gamma factor ( @xmath780 ) . in the collision between the pem pulse and the baryonic matter at @xmath781
, we impose total conservation of energy and momentum .
we consider the collision process between two radii @xmath782 satisfying @xmath783 and @xmath784 .
the amount of baryonic mass acquired by the pem pulse is @xmath785 where @xmath786 is the mean - density of baryonic matter at rest . as for energy density of dyadosphere , here
also we choose a simplification for the energy density : in fact during the passage of the shell a deposition of material on the external surface of the fireball creates ; however we neglected this effect and assumed that this material after collision diffuses instantaneously in the pulse with a constant density : @xmath787 where @xmath788 is the number of particle of the remnant shell swept up by the pulse and @xmath685 is the comoving volume of the fireball .
the conservation of total energy leads to the estimate of the corresponding quantities before ( with `` @xmath124 '' ) and after such a collision @xmath789 where @xmath790 is the corresponding increase of internal energy due to the collision .
similarly the momentum - conservation gives @xmath791 where the radial component of the four - velocity of the pem pulse is @xmath792 and @xmath254 is the thermal index .
we then find @xmath793,\label{heat_2}\\ \bar\gamma & = & { a\over\sqrt{a^2 - 1}},\hskip0.5 cm a\equiv { \bar\gamma_\circ \over u^\circ_r}+ { \delta m\over ( \gamma\bar\epsilon_\circ + \bar\rho^\circ_b)\bar\gamma_\circ u^\circ_r{\mathcal v}_\circ}. \label{dgamma_2}\end{aligned}\ ] ] these equations determine the gamma factor @xmath721 and the internal energy density @xmath794 in the capture process of baryonic matter by the pem pulse .
the effect of the collision of the pem pulse with the remnant leads to the following consequences : * a reheating of the plasma in the comoving frame but not in the laboratory frame ; an increase of the number of @xmath223 pairs and of free electrons originated from the ionization of those atoms remained in the baryonic remnant ; correspondingly this gives an overall increase of the opacity of the pulse ; * the more the amount of baryonic matter swept up , the more internal energy of the pemb pulse is converted in kinetic energy of baryons . by describing the interaction of pem pulse with remnant as completely inelastic collision of two particles
, one can compute by the energy - momentum conservation equation the decrease of lorentz @xmath41 and the increase of internal energy as function of @xmath53 parameter and also the ultrarelativistic approximation ( @xmath795 ) : 1 .
an abrupt decrease of the gamma factor given by @xmath796 where @xmath58 is the gamma factor of the pem pulse before the collision , 2 .
an increase of the internal energy in the comoving frame @xmath797 developed in the collision given by @xmath798 this approximation applies when the final gamma factor at the end of the pem pulse era is larger than @xmath799 , right panel in fig .
[ cip2 ] . in this phase of expansion ,
another thermodynamic quantity has not been considered : the chemical potential @xmath362 of the electrons from ionization of baryonic remnant .
we remind that the total proper number density of electrons of ionization is given by @xmath800 four equations are imposed to find a formula useful for numerical computation : the first one is the thermodynamical equilibrium of fireball , or @xmath801 the second one is @xmath802 where @xmath803 , with @xmath804 for hydrogen atoms and @xmath805 for baryonic matter in general ; the third one derives from the definition of @xmath53 , and states a relation between the two densities @xmath806 and @xmath741 : from definition of @xmath53 , we have @xmath807 where @xmath808 is the initial temperature of fireball and @xmath809 is a parameter ( @xmath810 ) defined by ( [ def_di_b_1 ] ) ; so if @xmath811 is the initial volume of dyadosphere and @xmath812 the initial volume of the baryonic shell @xmath813 finally the fourth one is the conservation law of baryonic matter @xmath814 therefore the chemical potential @xmath362 is numerically determined at a certain time of expansion if the initial temperature @xmath808 of fireball and the initial volume of baryonic shell @xmath812 are known and , at that time , the volume @xmath685 , the temperature @xmath297 and the lorentz factor @xmath721 of the fireball , the volume of the baryonic shell swept up @xmath815 and the ratio @xmath816 : @xmath817 where the factor in brackets @xmath818 must be considered only for @xmath819 , while the proportionality factor is the function zeta of riemann @xmath820 for computation of @xmath519 , with @xmath821 . +
therefore the equations for this phase are ( [ heat_2 ] ) , ( [ dgamma_2 ] ) , ( [ nb ] ) , ( [ paira ] ) and ( [ pot_chi_fin ] ) .
after the engulfment of the baryonic matter of the remnant the plasma formed of @xmath223-pairs , electromagnetic radiation and baryonic matter expands again as a sharp pulse , namely the pemb pulse .
the calculation is continued as the plasma fluid expands , cools and the @xmath223 pairs recombine until it becomes optically thin : @xmath822 where @xmath823 is the thomson cross - section and the integration is over the radial interval of the pemb pulse in the comoving frame . in order to study the pemb pulse expansion
the validity of the slab approximation adopted for the pem pulse phase has to be verified ; otherwise the full hydrodynamics relativistic equations should be integrated .
the pemb pulse evolution firstly has been simulated by integrating the general relativistic hydrodynamical equations with the livermore codes , for a total energy in the dyadosphere of @xmath824 erg and a baryonic shell of thickness @xmath825 at rest at a radius of @xmath826 and @xmath827 . in analogy with the special relativistic treatment for the pem pulse , presented in section [ era1 ]
( see also @xcite ) , for the adiabatic expansion of the pemb pulse in the constant - slab approximation described by the rome codes the following hydrodynamical equations with @xmath828 has been found @xmath829 in these equations ( @xmath830 ) the comoving baryonic mass and number densities are @xmath831 and @xmath832 , where @xmath685 is the comoving volume of the pemb pulse .
factor as a function of radial coordinate from the pemb - pulse simulation is compared with the @xmath721 factor as solution of hydrodynamics relativistic equations numerically integrated ( open squares ) for @xmath833erg and @xmath834 , @xmath835 and @xmath836 .
the result is in accordance with the scheme of a fireball with constant thickness in laboratory frame which is valid up to @xmath837_.,width=566 ] the result is shown in fig .
[ shells_2 ] @xcite where the bulk gamma factor as computed from the rome and livermore codes are compared and very good agreement has been found .
this validates the constant - thickness approximation in the case of the pemb pulse as well .
on this basis we easily estimate a variety of physical quantities for an entire range of values of @xmath53 .
for the same black hole different cases have been considered @xcite .
the results of the integration show that for the first parameter range the pemb pulse propagates as a sharp pulse of constant thickness in the laboratory frame , but already for @xmath838 the expansion of the pemb pulse becomes much more complex , turbulence phenomena can not be neglected any more and the constant - thickness approximation ceases to be valid .
it is also interesting to evaluate the final value of the gamma factor of the pemb pulse when the transparency condition given by eq.([thin_1 ] ) is reached as a function of @xmath53 , see fig .
[ 3gamma ] .
for a given black hole , there is a _
maximum _ value of the gamma factor at transparency . by further increasing the value of @xmath53
the entire @xmath55 is transferred into the kinetic energy of the baryons ( see also @xcite ) .
for the typical case @xmath839 erg .
the asymptotic values @xmath840 are also plotted .
the collision of the pem pulse with the baryonic remnant occurs at @xmath841 where the jump occurs.*right ) * the @xmath41 factor ( the solid line ) at the transparency point is plotted as a function of the @xmath53 parameter .
the asymptotic value ( the dashed line ) @xmath842 is also plotted_.,title="fig:",width=275 ] for the typical case @xmath839 erg .
the asymptotic values @xmath840 are also plotted .
the collision of the pem pulse with the baryonic remnant occurs at @xmath841 where the jump occurs.*right ) * the @xmath41 factor ( the solid line ) at the transparency point is plotted as a function of the @xmath53 parameter .
the asymptotic value ( the dashed line ) @xmath842 is also plotted_.,title="fig:",width=321 ] in fig .
[ 3gamma]left we plot the gamma factor of the pemb pulse as a function of radial distance for different amounts of baryonic matter .
the diagram extends to values of the radial coordinate at which the transparency condition given by eq.([thin_1 ] ) is reached .
the `` asymptotic '' gamma factor @xmath843 is also shown for each curve . the closer the gamma value approaches the `` asymptotic '' value ( [ asymp ] ) at transparency , the smaller the intensity of the radiation emitted in the burst and the larger the amount of kinetic energy left in the baryonic matter ( see fig . [ 3gamma]right ) .
from g. gamow @xcite : the summer of 1939 i spent with my family vacation on the copacabana beach in rio de janeiro . one evening , visiting the famous casino da urca to watch the gamblers , i was introduced to a young theoretical physicist born on an amazon river plantation , named mario schoenberg .
we became friends , and i arranged for him a guggenheim fellowship to spend a year in washington to work with me in nuclear astrophysics .
his visit was very successful , and we hit upon a process which could be responsible for the vast stellar explosions known as supernovae .
the trick is done by alternative absorption and reemission of one of the thermal electrons in the very hot ( billions of degrees ! ) stellar interior by various atomic nucleai .
both processes are accompanied by the emission of neutrinos and antineutrinos which , possessing tremendous penetrating power , pass through the body of a star like a swarm of musquitoes through chicken wire and carry with them large amount of energy .
thus , the stellar interior cools rapidly , the pressure drops , and the stellar body collapse with a great explosion of light and heat .
all this is too complicate to explain in non technical words , and i am mentioning it only as background for how we came to give that process is name .
we called it the urca process , partially to commemorate the casino in which we first met , and partially because the urca process results in a rapid disappearance of thermal energy from the interior of the star , similar to the rapid disappearance of money from the pockets of the gamblers of the casino d urca . sending our article `` on the urca process '' for publication on the _ physical review _
i was worried that the editor would ask why we called the process `` urca '' .
after much thought i decided to say that this is short for `` unrecordable cooling agent '' , but they never asked . today , there are other known cooling processes involving neutrinos which work even faster than the urca process . for example
, a neutrino pair can be formes instead of two gamma quanta in the annihilation of a positive and negative electrons .
99 r. ruffini , m.g .
bernardini , c.l .
bianco , p. chardonnet , f. fraschetti , s .- s .
xue , in `` proceedings of the tenth marcel grossmann meeting '' , m. novello , s.e .
perez - bergliaffa ( eds . ) , world scientific , singapore , in press .
f. fraschetti , m.g .
bernardini , c.l .
bianco , p. chardonnet , r. ruffini , s .- s.xue , in `` proceedings of the tenth marcel grossmann meeting '' , m. novello , s.e .
perez - bergliaffa ( eds . ) , world scientific , singapore , in press .
r. ruffini , c.l .
bianco , p. chardonnet , f. fraschetti , l. vitagliano , s .- s .
xue , `` new perspectives in physics and astrophysics from the theoretical understanding of gamma - ray bursts '' , in _ cosmology and gravitation : x@xmath844 brazilian school of cosmology and gravitation ; 25@xmath844 anniversary ( 1977 - 2002 ) _ , m. novello , s.e .
perez - bergliaffa ( eds . ) , _ aip conf .
_ , * 668 * , 16 ( 2003 ) . r. ruffini , `` beyond the critical mass : the dyadosphere of black holes '' , in _ black holes and high energy astrophysics , proceedings of the @xmath845 yamada conference _ , h. sato and n. sugiyama ( eds . ) , universal ac .
press ( tokyo , 1998 ) .
bernardini , c.l .
bianco , p. chardonnet , f. fraschetti , r. ruffini , s .- s.xue , in `` proceedings of the tenth marcel grossmann meeting '' , m. novello , s.e .
perez - bergliaffa ( eds . ) , world scientific , singapore , in press . | we outline the confluence of three novel theoretical fields in our modeling of gamma - ray bursts ( grbs ) : 1 ) the ultrarelativistic regime of a shock front expanding with a lorentz gamma factor @xmath0 ; 2 ) the quantum vacuum polarization process leading to an electron - positron plasma originating the shock front ; and 3 ) the general relativistic process of energy extraction from a black hole originating the vacuum polarization process .
there are two different classes of grbs : the long grbs and the short grbs .
we here address the issue of the long grbs .
the theoretical understanding of the long grbs has led to the detailed description of their luminosities in fixed energy bands , of their spectral features and made also possible to probe the astrophysical scenario in which they originate .
we are specially interested , in this report , to a subclass of long grbs which appear to be accompanied by a supernova explosion .
we are considering two specific examples : grb980425/sn1998bw and grb030329/sn2003dh .
while these supernovae appear to have a standard energetics of @xmath1 ergs , the grbs are highly variable and can have energetics @xmath2
@xmath3 times larger than the ones of the supernovae .
moreover , many long grbs occurs without the presence of a supernova .
it is concluded that in no way a grb can originate from a supernova .
the precise theoretical understanding of the grb luminosity we present evidence , in both these systems , the existence of an independent component in the x - ray emission , usually interpreted in the current literature as part of the grb afterglow .
this component has been observed by chandra and xmm to have a strong decay on scale of months .
we have named here these two sources respectively urca-1 and urca-2 , in honor of the work that george gamow and mario shoenberg did in 1939 in this town of urca identifying the basic mechanism , the urca processes , leading to the process of gravitational collapse and the formation of a neutron star and a supernova .
the further hypothesis is considered to relate this x - ray source to a neutron star , newly born in the supernova .
this hypothesis should be submitted to further theoretical and observational investigation .
some theoretical developments to clarify the astrophysical origin of this new scenario are outlined .
we turn then to the theoretical developments in the short grbs : we first report some progress in the understanding the dynamical phase of collapse , the mass - energy formula and the extraction of blackholic energy which have been motivated by the analysis of the short grbs . in this context progress
has also been accomplished on establishing an absolute lower limit to the irreducible mass of the black hole as well as on some critical considerations about the relations of general relativity and the second law of thermodynamics .
we recall how this last issue has been one of the most debated in theoretical physics in the past thirty years due to the work of bekenstein and hawking . following these conceptual progresses
we analyze the vacuum polarization process around an overcritical collapsing shell .
we evidence the existence of a separatrix and a dyadosphere trapping surface in the dynamics of the electron - positron plasma generated during the process of gravitational collapse .
we then analyze , using recent progress in the solution of the vlasov - boltzmann - maxwell system , the oscillation regime in the created electron - positron plasma and their rapid convergence to a thermalized spectrum .
we conclude by making precise predictions for the spectra , the energy fluxes and characteristic time - scales of the radiation for short - bursts . if the precise luminosity variation and spectral hardening of the radiation we have predicted will be confirmed by observations of short - bursts , these systems will play a major role as standard candles in cosmology .
these considerations will also be relevant for the analysis of the long - bursts when the baryonic matter contribution will be taken into account .
address = icra international center for relativistic astrophysics , altaddress = dipartimento di fisica , universit di roma `` la sapienza '' , piazzale aldo moro 5 , i-00185 roma , italy . address = icra international center for relativistic astrophysics , altaddress = dipartimento di fisica , universit di roma `` la sapienza '' , piazzale aldo moro 5 , i-00185 roma , italy .
address = icra international center for relativistic astrophysics , altaddress = dipartimento di fisica , universit di roma `` la sapienza '' , piazzale aldo moro 5 , i-00185 roma , italy .
address = icra international center for relativistic astrophysics , altaddress = universit de savoie , lapth - lapp , bp 110 , f-74941 annecy - le - vieux cedex , france .
address = icra international center for relativistic astrophysics , altaddress = universit di trento , via sommarive 14 , i-38050 povo ( trento ) , italy .
address = icra international center for relativistic astrophysics , altaddress = yerevan physics institute , alikhanian brothers street 2 , 375036 , yerevan-36 , armenia .
address = icra international center for relativistic astrophysics , altaddress = dipartimento di fisica , universit di roma `` la sapienza '' , piazzale aldo moro 5 , i-00185 roma , italy .
address = icra international center for relativistic astrophysics , altaddress = dipartimento di fisica , universit di roma `` la sapienza '' , piazzale aldo moro 5 , i-00185 roma , italy . |
the number of real solutions to a system of real equations is congruent to the number of complex solutions modulo two . in @xcite , we established a congruence modulo four for many symmetric schubert problems given by osculating flags , leaving as a conjecture a stronger form of that result .
we prove this conjecture for symmetric schubert problems given by flags that are isotropic with respect to a symplectic form , giving a simpler proof of a stronger and more basic result than that obtained in @xcite .
this congruence modulo four follows from a result on the real points in fibers of a map between real varieties equipped with an involution .
when the fixed point set of the involution has codimension at least two , the number of real points satisfies a congruence modulo four .
there is an involution acting on symmetric schubert problems given by isotropic flags and we can compute the dimension of the fixed point locus in a universal family of schubert problems .
our inability to compute this dimension when the flags are osculating was the obstruction to establishing the conjecture in @xcite .
the congruence modulo four often implies a non - trivial lower bound on the number of real solutions to a symmetric schubert problem given by isotropic flags .
similar lower bounds and congruences in real algebraic geometry have been of significant interest @xcite .
another topological study was recently made of this phenomena in the schubert calculus @xcite , and delicate lower bounds @xcite were given by computing the signature of a hermitian matrix arising in the proof of the shapiro conjecture @xcite . in section [
s : one ] we state our main result , whose proof occupies section [ s : two ] .
let @xmath0 be a complex vector space of dimension @xmath1 equipped with a nondegenerate alternating form @xmath2 .
write @xmath3 for the complex conjugate of a point , vector , subspace , or variety @xmath4 .
a variety @xmath4 is if it is defined by real equations ; equivalently , if @xmath5 .
write for the real points of a real variety @xmath4 , those that are fixed by complex conjugation .
write for the symmetric group of permutations of @xmath6 .
the set of @xmath7-dimensional linear subspaces of @xmath0 forms the grassmannian , @xmath8 , which is a manifold of dimension @xmath9 .
a is a sequence @xmath10 of linear subspaces of @xmath0 with @xmath11 .
a is a weakly decreasing sequence of integers @xmath12 .
a flag @xmath13 and a partition @xmath14 determine a schubert subvariety of @xmath8 , @xmath15 this has codimension @xmath16 in @xmath8 .
let @xmath17 be a list of partitions and @xmath18 be general flags . by kleiman
s transversality theorem @xcite the intersection @xmath19 is either empty or has dimension @xmath20 .
call @xmath21 a if this expected dimension is zero so that is either empty or consists of finitely many points .
the number of points in is independent of the choice of general flags .
we will assume that @xmath22 .
a choice of flags is an of the schubert problem @xmath21 ; its are the points in .
the instance is if for all @xmath23 , there is some @xmath24 with @xmath25 and @xmath26 , for then is stable under complex conjugation .
[ r : osculating ] osculating flags provide a rich source of isotropic flags .
as explained in section 3 of @xcite , a rational normal curve @xmath27 induces a symplectic form on @xmath0 and a symplectic form on @xmath0 gives rise to a rational normal curve , and we may assume that @xmath28 is real in that @xmath29 .
if @xmath28 is a rational normal curve corresponding to the symplectic form @xmath30 , then every osculating flag is isotropic .
( for @xmath31 , the is the flag whose @xmath23-plane @xmath32 is spanned by @xmath33 and its derivatives @xmath34 . )
the study of real solutions to schubert problems given by flags osculating at real points in grassmannians and flag manifolds has been quite rich and fruitful @xcite .
a partition @xmath14 is represented by its young diagram , which is a left - justified array of boxes with @xmath35 boxes in row @xmath23 .
we display some partitions with their young diagrams , @xmath36 a partition @xmath14 is if it is symmetric about its main diagonal , that is , if @xmath37 , where is the transpose of @xmath14 .
the partitions @xmath38 and @xmath39 are symmetric while @xmath40 is not .
a schubert problem @xmath21 is if every partition in @xmath21 is symmetric .
recall that our vector space @xmath0 was equipped with a nondegenerate alternating bilinear form @xmath30 .
a linear subspace @xmath4 of @xmath0 has annihilator @xmath41 under @xmath30 , @xmath42 and we have @xmath43 .
this induces a map @xmath44 on @xmath8 called the .
given a flag @xmath13 , we get the flag whose @xmath23-plane is @xmath45 .
a flag @xmath13 is if @xmath46 .
the of a symmetric partition is the number of boxes on its main diagonal , so @xmath47 while @xmath48 .
we state our main theorem .
[ th : main ] suppose that @xmath17 is a symmetric schubert problem on @xmath8 and that @xmath18 are isotropic flags defining a real instance of the schubert problem @xmath21 such that @xmath49 is finite . if @xmath50 , then the number ( counted with multiplicity ) of real points in @xmath49 is congruent to the number @xmath51 of complex points , modulo four .
we show in remark [ r : last ] that @xmath52 and this sum has the same parity as @xmath7 , so that the condition in theorem [ th : main ] for this congruence modulo four is that @xmath53 is not equal to @xmath7 or to @xmath54 , which is very mild .
we use the observation of remark [ r : osculating ] that osculating flags are isotropic to deduce a corollary about schubert problems given by osculating flags . fix a real rational normal curve @xmath27 with corresponding osculating flags @xmath55 for @xmath56 , and a symmetric schubert problem @xmath17 on @xmath8 .
an of this schubert problem is a list of distinct complex numbers @xmath57 which give corresponding osculating flags @xmath58 .
this osculating instance is if @xmath59 implies that @xmath60 .
we deduce a corollary to theorem [ th : main ] that implies conjecture 21 of @xcite , which was the strongest result one could reasonably expect to hold concerning this congruence modulo four for symmetric schubert problems .
this is strictly stronger than all congruence results obtained in @xcite .
suppose that @xmath17 is a symmetric schubert problem on @xmath8 and that @xmath58 are osculating flags defining a real instance of @xmath21 .
if @xmath50 , then the number ( counted with multiplicity ) of real points in @xmath61 is congruent to the number @xmath51 of complex points , modulo four .
we do not need to assert that the intersection consists of finitely many points , for it always does @xcite .
when @xmath51 is congruent to two modulo four and @xmath50 , there will always be at least two real solutions to a real instance of a symmetric schubert problem .
such lower bounds implied by theorem [ th : main ] occur frequently .
table [ ta : one ] gives the total number of symmetric schubert problems in @xmath8 for small values of @xmath7 , together with the number of those for which theorem [ th : main ] implies a lower bound of two .
the timings are reported in ghz - seconds ( s ) , ghz - hours ( h ) , and ghz - years ( y ) .
.numbers of symmetric schubert problem with a lower bound of two . [ cols=">,>,>,>,>,>,>",options="header " , ]
we follow the main line of argument for the results of @xcite . we observe that the lagrangian involution @xmath44 permutes the solutions to an instance of a symmetric schubert problem @xmath21 given by isotropic flags and then construct a family @xmath62 whose base parameterizes instances of the schubert problem @xmath21 given by isotropic flags and whose fibers are the solutions to those instances .
we then estimate the codimension of the @xmath63-fixed point locus of the family @xmath62 , which shows that the numerical condition @xmath50 implies that the fixed points have codimension at least two . finally , we invoke a key lemma from @xcite to complete the proof .
an @xmath7-dimensional subspace @xmath64 of @xmath0 is if @xmath65 .
the set of all lagrangian subspaces of @xmath0 forms the .
this is smooth of dimension @xmath66 and is a homogeneous space for the symplectic group of linear transformations of @xmath0 which preserve @xmath30 .
an isotropic flag @xmath13 and a symmetric partition @xmath14 determine a schubert subvariety @xmath67 of @xmath68 , which is the intersection @xmath69 , @xmath70 this has codimension @xmath71 in @xmath68 .
let @xmath21 be a symmetric schubert problem .
we construct families whose bases parameterize all instances of @xmath21 given by isotropic flags and whose fibers are the solutions to the corresponding instance .
the set of isotropic flags in @xmath0 is a flag manifold for @xmath76 of dimension @xmath9 .
define @xmath77 we have the two projections @xmath78 for isotropic flags @xmath18 , the fiber @xmath79 consists of the solutions @xmath80 to the instance of the schubert problem @xmath21 given by the flags @xmath18 . as @xmath76
does not act transitively on @xmath8 , we can not use kleiman s theorem @xcite to conclude that an intersection given by general flags is transverse .
transversality follows instead from the main result of @xcite .
consequently , there is a nonempty zariski open subset @xmath81 consisting of @xmath82-tuples of isotropic flags for which the intersection is transverse and therefore consists of @xmath51 points .
we seek a family @xmath83 of instances of @xmath21 where @xmath84 and @xmath85 is irreducible with @xmath86 parameterizing all real instances of @xmath21 .
since we can not easily compute the dimension of @xmath87 , we replace it by a possibly smaller set .
define to be the closure of @xmath88 in @xmath87 .
restricting @xmath89 to @xmath90 gives the dominant map @xmath91 where a fiber @xmath92 is a subset of the intersection and is equal to it when the intersection is finite .
thus @xmath93 .
let @xmath95 be the group of permutations @xmath96 of @xmath97 with @xmath98 for all @xmath99 .
then @xmath100 where @xmath21 consists of @xmath101 distinct partitions @xmath102 with @xmath103 occurring @xmath104 times .
then @xmath105 acts on the families @xmath106 , preserving fibers , @xmath107 define to be the quotient of @xmath108 by the group @xmath105 .
[ l : a ] the map @xmath109 is a proper dominant map of real varieties of the same dimension with @xmath110 smooth and @xmath111 connected . the lagrangian involution preserves fibers of @xmath89 and the codimension in @xmath112 of the @xmath63-fixed points @xmath113 is at least @xmath114 .
[ p : key ] let @xmath115 be a proper dominant map of real varieties of the same dimension with @xmath116 smooth .
suppose that @xmath117 has an involution @xmath63 preserving the fibers of @xmath118 such that the image in @xmath116 of the set of @xmath63-fixed points has codimension at least @xmath119 .
lemma 5 in @xcite requires that there are no @xmath63-fixed points in either fiber @xmath123 or @xmath124 .
this may be relaxed to only one fiber avoiding @xmath63-fixed points , which may be seen using a limiting argument along the lines of the proof of corollary 7 in @xcite . by lemma
[ l : a ] , the hypotheses of proposition [ p : key ] hold , as the inequality @xmath50 implies that @xmath125 .
let @xmath126 be isotropic flags defining a real instance of the schubert problem @xmath21 such that is finite . since this instance is real , for each @xmath99 if @xmath25 , then @xmath26 . thus there is a permutation @xmath127 such that @xmath128 for @xmath99 , and
so the image of @xmath126 in @xmath110 is a real point @xmath129 .
we complete the proof by exhibiting a point @xmath130 for which @xmath124 consists of @xmath51 real points , none of which are fixed by @xmath63 . for distinct @xmath131 ,
the intersection @xmath132 is transverse and consists of @xmath51 real points , by the mukhin - tarasov - varchenko theorem @xcite .
the osculating flags @xmath133 are real and isotropic , and we would be done if there were no @xmath63-fixed points in .
equivalently , if the intersection of the corresponding lagrangian schubert varieties were empty .
this is unknown , but expected , as it follows from conjecture 5.1 in @xcite which is supported by significant evidence .
since the intersection is transverse , if @xmath134 are real isotropic flags that are sufficiently close to the osculating flags in , then the intersection @xmath135 is transverse and consists of @xmath51 real points . by kleiman s theorem
@xcite we may also assume that @xmath136 are general in that the intersection @xmath137 of lagrangian schubert varieties is either empty or has dimension @xmath138 we conclude that is empty and therefore contains no lagrangian subspaces .
consider the quotient of @xmath94 by the group @xmath105 , which is the product @xmath140 where is the quotient @xmath141 and @xmath21 consists of @xmath101 distinct partitions @xmath102 with @xmath103 occurring @xmath104 times in @xmath21 . for @xmath142 ,
let @xmath143 be those flags in linear general position with respect to @xmath13 .
this dense subset of @xmath144 is a schubert variety isomorphic to @xmath145 . as @xmath13 varies in @xmath144 , these form an affine cover of @xmath144 .
given a finite set @xmath146 of isotropic flags , there is an isotropic flag @xmath13 that is simultaneously in linear general position with each @xmath147 , so that @xmath148 .
thus @xmath149 is covered by the @xmath150-invariant affine varieties @xmath151 , each isomorphic to @xmath152 . by descent , this implies that the quotient @xmath153 is well - defined and covered by affine varieties @xmath154 , each isomorphic to @xmath155 , as @xmath156 .
it follows that @xmath157 is a smooth irreducible variety whose real points are connected which implies the same for @xmath110 .
the map @xmath158 is proper as it comes from a projection along a grassmannian factor .
its fibers are preserved by the lagrangian involution and are equal over points in an @xmath105-orbit .
both properties hold for @xmath159 ( as @xmath160 is @xmath105-stable ) and therefore for @xmath161 .
we conclude that @xmath89 descends to the quotient @xmath109 , where it is a proper dominant map and the lagrangian involution preserves its fibers .
we study the @xmath63-fixed points of @xmath87 which form the universal family , @xmath163 consider the projection @xmath164 .
let @xmath165 .
then @xmath166 for @xmath14 symmetric and @xmath165 , define @xmath167 this is a schubert subvariety of @xmath144 of codimension @xmath168 . thus @xmath169 which has codimension @xmath170 in @xmath94 and is irreducible as each @xmath171 is a schubert variety and is therefore irreducible .
thus @xmath164 exhibits @xmath172 as a fiber bundle .
we compute its dimension , @xmath173 thus @xmath174 .
as @xmath175 is the set of @xmath63-fixed points of @xmath90 , @xmath176 , and @xmath112 is the quotient of @xmath90 by the finite group @xmath105 , the @xmath63-fixed points in @xmath112 have codimension at least @xmath177 .
[ r : last ] if @xmath21 is a symmetric schubert problem , the quantity @xmath178 is an integer , so @xmath53 has the same parity as @xmath7 .
for generic flags @xmath136 , the intersection of lagrangian schubert varieties is a subset of the intersection of schubert varieties . by kleiman s theorem
, this gives the inequality @xmath179 which implies that @xmath180 .
thus the only possibilities for @xmath53 for which theorem [ th : main ] does not imply a congruence modulo four are @xmath7 or @xmath54 . when @xmath181 , we have @xmath182 so that @xmath21 is a schubert problem for @xmath68 with @xmath183 solutions . that is , for general isotropic flags @xmath184 the intersection is transverse and consists of @xmath183 points .
when @xmath185 the family @xmath186 is reducible : @xmath172 is one component and @xmath187 is the other .
for example , the problem @xmath188 on @xmath189 has @xmath190 .
the corresponding problem in @xmath191 has four solutions .
thus four of the eight solutions on @xmath189 will be isotropic and the other four will not be isotropic . in our experimentation ( number 490 on @xcite ) this problem exhibits a congruence modulo four . in case ( 1 ) , @xmath196 has codimension one as does the image of the set of @xmath63-fixed points of @xmath112 , so proposition [ p : key ] does not necessarily imply a congruence modulo four . in cases ( 2 ) and ( 3 ) , @xmath196 has codimension two , and so there will be a congeuence modulo four .
we have observed some symmetric schubert problems with @xmath192 that have a congruence modulo four and some that do not ( so that ( 1 ) holds ) .
for example , @xmath197 on @xmath189 has @xmath198 ( number 495 on @xcite ) and its numbers of real solutions do not exhibit a congruence modulo four , but @xmath199 on @xmath189 has @xmath198 ( number 497 on @xcite ) and its possible numbers of real solutions appear to be congruent modulo four .
we believe that ( 3 ) is unlikely and that ( 2 ) holds if and only if there is a congruence modulo four . | we previously obtained a congruence modulo four for the number of real solutions to many schubert problems on a square grassmannian given by osculating flags . here , we consider schubert problems given by more general isotropic flags , and prove this congruence modulo four for the largest class of schubert problems that could be expected to exhibit this congruence . |
the study of packing unit squares into a square goes back to erds and graham @xcite , who examined the asymptotic packing efficiency as the side length of the containing square increased towards infinity .
gbel @xcite was the first to show that particular packings are optimal for a given non - square number of unit squares .
the search for good packings for given number of unit squares was addressed in the popular science literature in various articles by gardner @xcite .
let @xmath0 be the side length of the smallest square into which @xmath1 non - overlapping unit squares can be packed .
non - trivial cases for which @xmath0 is known are @xmath8 for @xmath9 ( nagamochi @xcite , single values previously shown by gbel @xcite , el moumni @xcite , and friedman @xcite ) , @xmath10 ( gbel @xcite ) , @xmath4 ( kearney and shiu @xcite ) , @xmath11 ( stromquist @xcite ) , @xmath2 , and @xmath3 ( bentz @xcite ) .
there are moreover non - trivial best packings and lower bounds known for various values of @xmath1 .
examples on many of these results and the underlying techniques used are given in the survey article by friedman @xcite .
in @xcite and @xcite , it was shown that @xmath5 for @xmath12 .
these results suggested that the holds for the intermediate values @xmath7 . we will show this result in this article , by adopting the proof for @xmath13 from @xcite .
previously , the best lower bounds in these cases are @xmath14 and @xmath15 , and follow from a general result in nagamochi @xcite . as the trivial ( or `` chess board '' ) packings show that @xmath16 , and @xmath17 , it it suffices to establish the opposite inequality . let a box be the interior of any square with side length @xmath18 satisfying @xmath19 . following stromberg , we will establish that @xmath20 squares can not be packed in a square with side length smaller then @xmath21 by proving the equivalent statement that it is impossible to pack @xmath20 boxes in a square of side length @xmath21 . in order to do so
, we will adopt the previously used method of unavoidable points to continuously varying sets of such points .
we will introduce this modification in section [ nonav ] , in addition to given several technical lemmas .
the optimality proofs for @xmath22 ( @xmath23 ) are then given in sections [ sec33 ] , [ sec22 ] , respectively .
optimality proofs for square packing utilize arguments based on resource starvation .
subsets of a containing square are associated with numerical resources in such a way that each packed box uses up a certain amount of resources ( by intersecting the subset corresponding with the resource ) .
the overall amount of resource available limits the number of boxes that can be packed . the proofs in @xcite and many later publications are based on finite number of points , each of which has resource value @xmath24 . in @xcite ,
resources were associated with line segments , such that the length of intersection between a box and the line segment determined the amount of resource allocated to the box .
a more complex configuration in @xcite uses a combined system of ( weighted ) points , line segments , and a rectangular area .
our arguments will use a two - tier approach .
we will first start out with systems of points containing too many resources for a direct proof .
a new technical result ( theorem [ t : cont ] ) will allow us to use the flexibility in our initial systems to show that any potential packing must contain a local abundance of boxes .
we will use this local over - concentration " of boxes to obtain a contradicting in combination with a second resource system based on a line segment .
we will start by stating several `` non - avoidance '' lemmas , which guarantee that a box will intersect particular type of subsets in its vicinity .
[ nonav ] [ triangle ] let @xmath25 be a triangle with sides of length at most 1 .
then any box whose centre is in @xmath25 must contain one of the vertices of @xmath25 .
[ rimevendistance ] let @xmath26 , @xmath27 , and @xmath28 , then any box whose centre is in the rectangle @xmath29\times[0,b]$ ] must intersect the @xmath30-axis , the point @xmath31 or the point @xmath32
. we will use lemma [ rimevendistance ] in the cases of @xmath33 , @xmath34 and @xmath35 , @xmath36 .
[ rimunevenpoints ] let @xmath37 , and @xmath32 within a distance of @xmath24 from @xmath38 .
moreover , let @xmath39 be the infimum of @xmath40 for @xmath41 $ ] .
if @xmath42 , then any box whose centre is in the quadrilateral with vertices @xmath43 and @xmath32 must intersect the @xmath30-axis , the point @xmath38 or the point @xmath32 . moreover , the infimum of ( [ fa ] ) is a minimum and is obtained at a value of @xmath44 satisfying @xmath45 we will be using lemma [ rimunevenpoints ] in the case @xmath46 . if @xmath47 is a line that lies within a distance of @xmath48 of the centre of a box @xmath49 , then @xmath47 will intersect @xmath49 with a length of more than @xmath24 .
[ closeline ] [ l : paralines ] let @xmath50 and @xmath51 be two parallel lines of distance @xmath52 , and @xmath49 a box with its centre between them .
then @xmath49 must intersect the two lines with a common length of intersection of at least @xmath53 .
the following lemma extends lemma [ rimunevenpoints ] to values of @xmath54 smaller than @xmath55 .
let @xmath56 , and @xmath32 within a distance of @xmath24 from @xmath38 .
then any box whose centre is in the quadrilateral @xmath57 with vertices @xmath43 and @xmath32 must intersect the @xmath30-axis , the point @xmath38 or the point @xmath32 .
* proof : * if @xmath58 then the distance from @xmath59 to @xmath32 is less than @xmath24 and so the line segment between these points divides @xmath57 into two triangles , all of whose sides have length at least one .
the result now follows from lemma [ triangle ] .
so assume that @xmath60 and that the box @xmath49 does not intersect @xmath38 or the @xmath30-axis .
now the two line segments from @xmath59 to @xmath61 , and from @xmath38 to @xmath61 divide @xmath57 into three triangles , such that lemma [ triangle ] is applicable to each of them . by our assumption
, the box @xmath49 must contain either @xmath32 or @xmath61 . in the first case , the lemma holds ,
so assume that @xmath49 contains @xmath61 .
as the centre of @xmath49 is contained in @xmath57 , it is also contained in the larger rectangle @xmath62 with corners @xmath59 , @xmath63 , @xmath64 , and @xmath38 .
applying lemma [ rimevendistance ] to @xmath62 yields that @xmath49 contains @xmath64 .
as @xmath32 lies on the line segment from @xmath61 to @xmath64 , it is contained in @xmath49 .
we will use the lemma for @xmath65 , @xmath66 .
[ l : pbl ] let @xmath47 be a line and @xmath67 a point with a distance of more than @xmath68 from @xmath47 . if a box @xmath49 covers @xmath67 such that @xmath67 and the center of @xmath49 lie on opposite sides of @xmath47 , than @xmath49 intersect @xmath47 with a length of intersection that exceeds @xmath24 .
* proof : * the midpoint of b must lie within @xmath69 of @xmath67 , and hence within a distance of @xmath70 from the line @xmath47 . as this value is less than @xmath71 , the results follows from lemma 5 . consider a square @xmath72 of side length @xmath47 in the euclidean plane ( which in our cases we will take to be @xmath73\times [ 0,m]$ ] for @xmath74 ) .
a set of points @xmath75 is called _ unavoidable _ if every box @xmath76 contains one of the points in @xmath67 . in practice , we show unavoidability of @xmath67 by dividing @xmath72 into several regions @xmath77 , so that by one of our unavoidability lemmas , any box with midpoint in @xmath77 must either intersect a point in @xmath67 or the boundary of @xmath72 . if @xmath72 contains an unavoidable set of @xmath78 points , it follows that no more than @xmath78 boxes can be packed into @xmath72 , and hence @xmath79 .
figure [ 6pat ] depicts an unavoidable set of points for the square @xmath80\times [ 0,6]$ ] .
the points in the lowest row are @xmath81 and the remaining ones are arranged so that all shown triangles are equilateral of side length 1 .
lemma [ triangle ] is applicable to the triangles , lemma [ rimevendistance ] to the rectangles , and lemma [ rimunevenpoints ] to the remaining quadrilateral regions ( with @xmath82 , @xmath83 ) .
figure [ 6pat ] is a variant of configurations used to show that @xmath3 in @xcite and to derive a lower bound on @xmath84 in @xcite .
as the unavoidable set consists of @xmath85 points , it demonstrates the ( known ) result that @xmath86 .
note that the configuration in figure [ 6pat ] contains a degree of flexibility .
for example , we can obtain a different unavoidable configuration by deleting one of the points closest to the left hand side of the square and instead adding a different point a small amount further to the right .
if there exists a packing of @xmath85 boxes , then each of them must contain exactly one point in each configuration , and hence one box must contain both the deleted and added point ( and the line segment between them ) , an argument that has appeared in several previous proofs .
the next theorem shows that this approach can be generalized to situations in which more than one point is moved at one time .
[ t : cont ] let @xmath72 be a square with a packing @xmath87 of boxes , @xmath88 $ ] , @xmath89 , and @xmath90 a collection of continuous mappings , for @xmath91 .
suppose further that 1 . for each @xmath92 , @xmath93 is an unavoidable set of points ; 2 . if for some @xmath94 , @xmath95 is not contained in a box of @xmath87 , then @xmath96 for all @xmath92 .
if for some @xmath97 , @xmath98 , @xmath95 and @xmath99 lie in the same box of @xmath87 , then @xmath96 for all @xmath92 .
then for all @xmath94 , the image @xmath100 will either lie entirely within one box , or completely outside any box .
* proof : * if @xmath95 is not contained in any box , then @xmath101 is constant .
hence for the theorem to be wrong , there must be @xmath91 , @xmath102 , such that @xmath95 lies in some box @xmath103 while @xmath104 .
assume that this is indeed the case .
as boxes are open and @xmath101 is continuous , it follows that there is a smallest such @xmath105 for which @xmath106 lies outside @xmath103 . minimizing over all indices
, we may assume w.l.o.g . that @xmath107 is the smallest value of @xmath108 for which any @xmath109 lies outside the box containing @xmath110 .
now , as @xmath111 is an unavoidable set of points , there exist a @xmath112 , necessarily with @xmath113 , such that @xmath114 .
boxes are open , therefore there exist an @xmath115 such that @xmath116 . by the minimality of @xmath107 , it follows that @xmath117 .
however , now @xmath118 , and so @xmath101 is constant by condition 3 . , contradicting that @xmath119 .
the result follows .
[ s33 ] 33 non - overlapping unit squares can not be packed in a square of side length less than 6 .
* proof : * let @xmath72 be the square @xmath80 ^ 2 $ ] and assume by way of contradiction that there is a packing @xmath87 of 33 boxes into @xmath72 . consider the collection of 33 red points and 33 blue points depicted in figure [ fig : redblue ] .
the red points are the points from figure [ fig : redblue ] , while the blue points are obtained from the red ones by mirroring along the line @xmath120 .
it follows that both red and blue points form unavoidable sets of points .
hence each box in @xmath87 will contain exactly one red and blue point
. we will apply theorem [ t : cont ] twice . in the first instance we choose our values
@xmath121 so that @xmath122 is the set of red points from figure [ fig : redblue ] , while in the second case the start configuration will be the set of blue points . we will give an informal description of the other values of @xmath123 by describing the movement " of the points configuration .
many of our movement will move an entire row of equally - colored points from figure [ fig : redblue ] .
we will denote the red rows and blue rows by @xmath124 , and @xmath125 , respectively , where the rows are numbered from bottom to top .
for any two such rows @xmath126 , we denote by @xmath127 their vertical distance .
we first note that as every red point and every blue points lies in exactly one box of @xmath87 , the second and third condition of theorem [ t : cont ] are automatically full - filled . now consider a simultaneous vertical movement of a row @xmath128 .
such a move will preserve the unavoidability of the points configuration as long as , whenever they are defined , @xmath129 , @xmath130 and in addition , the vertical distance from @xmath131 and @xmath132 to the top or bottom edge of @xmath72 , respectively , is at most @xmath133 . with regard to a given configuration , for @xmath134 , let @xmath135 , @xmath136 , for @xmath134 , and @xmath137 .
clearly , for @xmath138 , there exists a unique configuration @xmath139 that minimizes @xmath140 and that is reachable from @xmath122 by vertical movement of rows , such that unavoidability is preserved throughout .
let @xmath141 be the second coordinate values of the points in the @xmath108-th row in @xmath139 ( the exact values of @xmath142 can be easily calculated , but are not needed ) . as @xmath143
we note that in @xmath139 , we have @xmath144 , wherever defined . now consider @xmath139 for @xmath145 . here
the the @xmath108-th row contains @xmath146 points , and we may move the @xmath108-th row horizontally to to the left and right , provided the distances to all critical " points in the adjacent rows stays within @xmath24 . as the adjacent rows have a vertical distance of less than @xmath147 , it is easy to check that this allows for a movement of at least @xmath148 to either side .
in addition , in this situation , we may move the left - most point of any such row horizontally to the right , until it reaches the point @xmath149 .
this maximal reflection is possible , as the distance to the adjacent row is less than @xmath150 .
we now proceed as follows .
we move the point configuration to one of the @xmath139 .
if the @xmath108-th row contains @xmath151 points , we note that one of the points occupies the point @xmath149 . if the @xmath108-th row contains @xmath146 points , we move the row @xmath148 to the left and back to the right .
finally , move the left - most point of row @xmath108 from @xmath152 to @xmath149 and back , and note that this point , combining both movements , has moved over the line segment from @xmath153 to @xmath149 .
we repeat this procedure for all @xmath139 . by theorem [ t : cont ] , the line segments @xmath154\times \{y_i\}$ ] for @xmath155 lie within the same box of @xmath87 . moreover , as these sets are the result of movement from different points of the base configuration , they , as well as the point sets @xmath156 from the movement of the @xmath151-point rows lie in different boxes .
we now repeat this procedure for the blue points , noting that we obtain the same values @xmath142 as in the case of the red points .
the blue points will have a @xmath146-point row where the red points have a @xmath151-point row and vice versa .
we can conclude that for @xmath157 , the line segment @xmath154\times \{y_i\}$ ] lies within one box of @xmath87 .
hence , taking both configurations together , the segments @xmath154\times \{y_i\}$ ] lie each in one box of @xmath87 for @xmath158 .
moreover , these segments all lie in different boxes as the points @xmath149 are all within the movement of different points from the red base configuration .
it follows that in @xmath87 , there are @xmath146 distinct boxes @xmath159 such that @xmath160 covers @xmath153 .
now let @xmath47 be the line segment from @xmath161 to @xmath162 .
we are interested in the length of the intersection of @xmath160 and @xmath47 .
if the midpoint of @xmath160 lies on the same side of @xmath47 as @xmath163 , then this intersection exceeds @xmath24 by lemma [ l : paralines ] ( with @xmath164 ) .
if the midpoint of @xmath160 lies on @xmath47 or on the side of @xmath47 opposite from @xmath153 , then the intersection has length larger than @xmath24 by lemma [ l : pbl ] . hence all six boxes intersect @xmath47 with a length larger than @xmath24
however , the length of @xmath47 is @xmath146 , for a contradiction .
hence @xmath85 boxes can not be packed in a square of side length @xmath146 and so @xmath165 .
the trivial packing is optimal for packing @xmath85 unit squares in a square , and we have that @xmath166 .
[ th22 ] @xmath167 non - overlapping unit squares can not be packed in a square of side length less than @xmath151 .
* proof : * let @xmath72 be the square @xmath168 ^ 2 $ ] , and assume , by way of contradiction , that there exist a packing @xmath87 of @xmath167 boxes into @xmath72 . consider the configuration of @xmath167 red and @xmath169 blue points shown in figure [ fig : bluered22 ] .
together , the points of both colours are exactly the elements the set @xmath170 .
we will denote the rows of red and blue points by @xmath171 , and @xmath172 , with numbering from bottom to top , and we let @xmath173 be the value of their second coordinate . by the same basic argument
as applied to figure [ fig : redblue ] we can show that both the set of red points and the set of blue points form unavoidable sets .
it follows that each red points lies in exactly one box from @xmath87 . in the case of the blue points , either exactly one blue point
is not contained in a box of @xmath87 or exactly one box of @xmath87 contain two blue points . in the latter case
, the two blue points that lie in the same box must be within a distance of each other that is smaller than the diagonal of a maximal size box , which is @xmath174 .
thus , by the symmetry in our blue point configuration , we may assume that all blue points with first coordinate value lower than @xmath175 lie in a box of @xmath87 that does not contain any other blue points .
as in the proof of theorem [ s33 ] , we will apply theorem [ t : cont ] twice , with the red and blue points as the respective starting configurations .
once again we describe the function @xmath101 informally in terms of the movement of points . in the case of the red points , we will first move @xmath176 and @xmath177 ( i.e. the red rows containing @xmath151 points ) horizontally a distance of @xmath148 to the left and back , followed by a movement of the left - most point in each such row horizontally to the right until it reaches a first coordinate value of @xmath24 .
these movement preserve unavoidablility , as the vertical distance between adjacent rows is @xmath147 in all cases .
as in theorem [ s33 ] we record that different boxes of @xmath87 contain the line segments @xmath154\times \{1.7\}$ ] and @xmath154\times \{3.3\}$ ] , and that these boxes must also be different from the boxes containing the ( stationary ) points @xmath178 , @xmath179 , and @xmath180 .
for the blue points , we want repeat these movement with rows @xmath181 and @xmath182
. however , we need to modify our procedure , as theorem [ t : cont ] requires the a blue point remains stationary if it does not lie in a box of @xmath87 or is in a box of @xmath87 that contains a different blue point . if row @xmath128 , @xmath183 does not contain such point , we move it horizontally a distance of @xmath148 and back to its original position .
we then move its leftmost point horizontally from @xmath152 to @xmath149 . in case that @xmath128 does contain such point
, we just move its leftmost point from @xmath152 to @xmath149 .
we note that the last case can only happen for one of the rows @xmath184 , because if there are two such exceptional blue points , they must lie in the same box , and hence ( due to the size of the boxes ) in either the same or adjacent rows . by theorem [ t : cont ]
the trajectory of each of the leftmost points of @xmath185 lies completely within a box of @xmath87 .
moreover each of these trajectories intersects a different red point from the initial configuration , and hence the trajectories lie within different boxes .
we can conclude that there are @xmath151 boxes @xmath186 in @xmath87 such that the line segments @xmath154 \times \{y_i\}$ ] lie completely within @xmath160 , except that at most one of @xmath187 might only cover the line segment @xmath188 \times \{y_i\}$ ] .
let @xmath47 be the line segment @xmath189 $ ] .
as in theorem [ s33 ] we can check that if @xmath160 contains @xmath154\times \{y_i\}$ ] , it intersect @xmath47 with a length of intersection that exceeds @xmath24 .
as @xmath47 has length @xmath151 , one of @xmath187 does not cover the entire line segment from @xmath154\times\{y_i\}$ ] .
we consider @xmath175 cases : 1 .
first assume that @xmath190 does not completely cover @xmath191\times\{y_3\}$ ] .
as there was at most one exceptional row , @xmath192 and @xmath193 all intersect @xmath47 with a length of intersection exceeding @xmath24 .
it follows that @xmath194 , and so @xmath190 does not cover the points @xmath195 and @xmath196 . in figure [
fig : midline ] , these two points are depicted in green .
+ let @xmath21 be the midpoint of @xmath190 .
the location of @xmath21 is constraint as follows : @xmath21 must lie on the right side of @xmath47 and separated from it by a distance of at least @xmath197 , for otherwise the length of intersection of @xmath190 and @xmath47 would exceed @xmath24 by either lemma [ l : paralines ] or lemma [ closeline ] . as @xmath190 does not cover @xmath195 or @xmath198 ,
@xmath21 can not be within a distance of @xmath199 from either of these points . finally , as @xmath190 covers @xmath188 \times \{y_3\}$
] , the distance from @xmath21 to @xmath200 must be smaller than half the diagonal of a maximal box , i.e. smaller than @xmath201 .
figure [ fig : midline ] shows the remaining possible locations of @xmath21 as a shaded area .
the area is bounded by line and circle segments that intersect in @xmath202 points with approximate coordinates @xmath203 .
+ an easy calculation shows that the entire area is within a distance of @xmath199 from the point @xmath204 . in figure
[ fig : midline ] , this distance is indicated by a circle .
it follows that @xmath190 also covers the point @xmath205 .
hence in our initial configuration of points , the box @xmath190 covers two blue points , namely those at @xmath206 and @xmath207 .
however , this contradict our assumption that all blue points with a second coordinate value smaller than @xmath175 do not share a box with another blue point .
2 . assume that for one @xmath208 , @xmath160 does not cover the entire line segment @xmath154 \times \{y_i\}$ ] .
by symmetry , we may assume that this is the case for @xmath209 .
then @xmath160 covers @xmath154\times \{y_i\}$ ] for @xmath210 , and , as in the previous case , this implies that each such @xmath160 intersects the line segment @xmath47 with a length of intersection larger than @xmath24 .
it follows that the point @xmath211 is denied to @xmath212 .
this point is depicted green in figure [ fig : lowline ] .
+ let @xmath21 be the midpoint of @xmath212 .
as before we can conclude that @xmath21 lies on the opposite side of @xmath47 from the point @xmath213 , with a distance of at least @xmath197 from @xmath47 , but within a distance of @xmath201 of @xmath214 .
these constraints intersect at approximately @xmath215 and @xmath216 .
+ the resulting area is depicted in figure [ fig : lowline ] and lies completely within a distance of @xmath217 from the point @xmath211 .
it follows that @xmath211 lies in @xmath212 .
however , the point is denied to @xmath212 , for a contradiction .
99 w. bentz , optimal packings of 13 and 46 unit squares in a square _ electronic journal of combinatorics _ * 17 * ( 2 ) : # r126 , 2010 .
s. el moumni , optimal packings of unit squares in a square , _ studia sci .
_ * 35 * ( 1999 ) , no . 3 - 4 , 281 - 290 .
p. erds and r. l. graham , on packing squares with equal squares , _
j. combin .
theory ser .
_ a * 19 * ( 1975 ) 119 - 123 .
e. friedman , packing unit squares in squares : a survey and new results _ _ , the electronic journal of combinatorics _ _ , dynamic surveys ( # ds7 ) ( version of aug 14 , 2009 ) m. gardner , mathematical games " , _ scientific american _
( oct 1979 , nov 1979 , mar 1980 , nov 1980 ) .
f. gbel , geometrical packing and covering problems , in _ packing and covering in combinatorics _ , a. schrijver ( ed . ) , _ math centrum tracts _ * 106 * ( 1979 ) 179 - 199 .
m. kearney and p. shiu , efficient packing of unit squares in a square , _ elect .
_ * 9 * # r14 ( 2002 ) .
h. nagamochi , packing unit squares in a rectangle , _ elect .
_ * 12 * # r37 ( 2005 ) .
w. stromquist , packing unit squares inside squares i , ii , iii " , unpublished manuscripts , 1984 - 5 ; http://www.walterstromquist.com/publications.html
. w. stromquist , packing 10 or 11 unit squares in a square , _ elect .
j. comb . _
* 10 * # r8 ( 2003 ) . | let @xmath0 be the side length of the smallest square into which @xmath1 non - overlapping unit squares can be packed . in 2010 ,
the author showed that @xmath2 and @xmath3 .
together with the result @xmath4 by keaney and shiu , these results strongly suggest that @xmath5 for @xmath6 , in particular for the values @xmath7 , which correspond to cases that lie in between the previous results . in this article
we show that indeed @xmath5 for @xmath7 , implying that the most efficient packings of 22 and 33 squares are the trivial ones . to achieve our results , we modify the well - known method of sets of unavoidable points by replacing them with continuously varying families of such sets . |
the transit method ( tm ) has recently gained much attention among the various methods for the detection of extrasolar planets due to its potential for discovering earth - like planets .
this method also offers the possibility for studying the existence of planetary satellites , rings , and atmospheric features ( see , for instance , deeg 2002 ) .
all these possibilities have been clearly demonstrated with hd209458 , the first planetary transit discovered ( charbonneau et al .
2000 ; henry et al .
2000 ; mazeh et al .
the light curve of this object has allowed an unprecedented characterization of the stellar / planet system ( e.g. planetary and stellar radii , stellar limb darkening , orbital inclination , period , etc ) , imposing limits to the presence of planetary satellites and rings ( brown et al .
2001 ) , and detecting features associated with the planet atmosphere ( charbonneau et al . 2002
; vidal - madjar et al . 2003 , 2004 ) .
the success of the tm largely relies on the ( spectro ) photometric accuracy for which the light curve can be obtained .
for some science applications ( e.g. detection of planetary atmospheric features ) only accurate _ relative _ spectrophotometry ( i.e. band1/(nearby)band2 and in / out transit ) is required , reducing drastically systematic correlated errors which may otherwise limit the s / n .
experiments with the hst ( gilliland , goudfrooij , kimble 1999 ; brown et al . 2001 ; gilliland and arribas 2003 ) have demonstrated that accuracies of @xmath010@xmath3 or higher can be reached from space .
such a high accuracy is not always required .
for instance in the wavelength domain 11901710 , lyman @xmath4 , oi , and cii lines suffer during the transit a drop of @xmath015 , 13 and 7.5 percent , respectively ( vidal - madjar et al . 2003 , 2004 ) . in optical wavelengths ,
the variations are indeed smaller .
charbonneau et al .
( 2002 ) have found a drop of 2.5 parts in 10000 ( i.e. 0.025 percent ) during the transit in the nad resonant lines .
this drop is smaller than expected from ( poorly constrained ) models , which also predict that other spectral characteristics in the optical ( e.g. rayleigh continuum ) should be detected with lower accuracies ( e.g. brown 2001 ) .
the smaller than expected nad absorption detected by charbonneau et al .
( 2002 ) may have resulted from a high cloud deck , a low atomic na abundance , or a combination of both . in this context
, it is relevant to understand which are the accuracy limits that can be achieved from the ground .
this should allow us to know if part of future studies can be performed taking advantage of the more abundant and flexible observing time available from ground - based observatories , or if we need to rely entirely on space data .
note that , with the advent of large aperture ground telescopes , the poisson limit may be boosted allowing access to more distant systems .
this is particularly relevant in connection with follow up observations of the planetary transits to be discovered by missions such as kepler ( borucki et al .
pioneering attempts to detect spectral features due to the planetary atmosphere from the ground in 51 peg and hd209458 include those of coustenis et al .
( 1998 ) , rauer et al .
( 2000 ) , bundy & marcy ( 2000 ) , moutou et al .
( 2001 ) , and brown et al .
more recently deming et al .
( 2005 ) using nirspec on the keck ii telescope have reported upper limits to the co absorption at a sensitivity level which suggests that a general masking mechanism is present in the planetary atmosphere of hd209458b . also recently narita et al .
( 2005 ) have studied the transmission spectrum of this object on the basis of high resolution spectroscopy with _
subaru_. although the accuracy of these observations were unable to confirm or contradict the result of charbonneau et al . , they reported upper limits on absorption due to several optical transitions .
all these ground based observations have gained importance since stis
hst is no longer operational . in this paper
we propose to take advantage of the comprehensive collecting nature of ifs to perform this type of observation .
the observational method , whose advantages are described in section 2 , is tested with ground based data .
the observations are described in section 3 .
details about the reduction process for this particular application are described in section 4 .
the results are shown and discussed in section 5 , while section 6 summarizes the main conclusions .
gilliland , goudfrooij , and kimble ( 1999 ) proposed to increase the photometric accuracy of very high s / n observations by using the spectroscopic mode of stis . by distributing the star light along the dispersion direction it is possible to increase the total number of photons collected before the detector reaches the saturation limit .
this simple idea enhances the duty cycle of the instrument ( i.e. total number of photons by unit of available time , including overheads ) , which translates to an increase in the s / n when the photon noise is the major source of noise .
the larger detector footprint of the stellar light has a negligible consequence in terms of adding noise in a very high s / n regime .
actually it has the beneficial effect of averaging out flat field and sensitivity residuals .
another practical advantage is the possibility to define _ a posteriori _ ad hoc filters well suited for the science applications .
here we propose to extend this idea by means of integral field spectroscopy ( ifs ) .
thanks to the 2d @xmath5 1d conversion done by a fiber bundle ( or by an image slicer ) , the star light is not only spread along the dispersion direction but also across dispersion .
this technique allows us to use a larger area of the detector , boosting the limit imposed by the photon - noise per exposure .
note that to optimize the use of the detector , the image of the star at the focal plane should be well oversampled by the integral field unit ( ifu ) . since for these transit observations we are interested in the total number of photons recorded in a limited period of time ( i.e. duration of the transit )
, the image at the input of ifu could be defocused .
note that defocussing the star image on the ifu does not modify the spectral resolution of the observations , which is another practical advantage of the ifs method with respect to previously used spectroscopic methods .
the ability to distribute the light over the whole detector is particularly relevant with the advent ( and prospects ) of very large telescopes .
this can be easily understood with the following example .
typical imaging observations of hd209458b ( v = 7.64 ) in a 100 band with an 8 m .
telescope will saturate in @xmath00.1 sec . considering a typical readout time of 60 sec , and taking into account that the transit lasts @xmath03 hours , an effective integration time of @xmath020 seconds on the source during the transit could be obtained .
if the observations are now done with a standard slit spectrograph working at , say , 1 / pixel , saturation is roughly reached in @xmath010 sec , and the effective integration time becomes @xmath030 minutes . if now the observations are done with an ifs at the same spectral resolution and sampling the ( defocused ) star image with , say , 100 spatial elements on the ifu ( _ spaxels _ , according to ifs community ) , the effective exposure time may be increased up to 2.8 hours ( out of 3 available ) .
of course , real observations can be optimized by defining faster readouts , modifying the spectral resolution , etc .
detectors like that used by ultracam ( dhillon & marsch 2002 ) may also reduce overheads .
the numbers above are just illustrative of current standard detectors .
apart from the ability to gather photons , stability is crucial for tm observations . of course , beyond some limits it is not possible to control the stability of the environment and/or the observational set - up .
while this is true for both space and ground observations , the unstable nature of the earth atmosphere makes very accurate ground observations particularly challenging . in any case , when uncontrolled instabilities may affect the photometric accuracy of the system , a good record of them may give the possibility to decorrelate the photometric signal .
this provides another important advantage of the ifs .
since both , the spectra from which the photometric information is obtained _ and _ the images of the object ( psfs ) are extracted from the same data cube , we can remove photometric variations induced by psfs instabilities .
furthermore , in fiber - based ifs ( and to a lesser degree a slice - based ifu ) a shift of the image on the focal plane may be well tracked independently of a shift between the pseudo - slit and the detector .
this gives full control of the psf variations , as well as the optical and mechanical shifts during time series observations .
fiber systems also provide a series of generic advantages such as , for instance , the azimuthal scrambling of the light ( which reduces errors in the barycenter of the recorded fiber image ) , the possibility to use static spectrographs ( which reduces flexure induced errors ) , etc ( see more details in , for instance , arribas & mediavilla , 2000 ) . on the other hand ,
potential problematic aspects of fiber - based ifs systems are the modal noise due to small changes in fiber position or illumination ( e.g. baudrand and walker 2001 ) , and a possible temporal variation of the fiber throughput with wavelength .
however , we will see later that , at least for the present instrumental configuration , our data do not suggest they are significant sources of noise . another advantage of ifs for these type of observations is a consequence of the large number of spectra collected simultaneously , which can be used to _ autocalibrate _ the data themselves from detector and background signatures .
this will be shown in section 5 .
summarizing , ifs provides three main advantages for transiting planets observations : i ) improves the ability to collect photons during the transit , enhancing the s / n , ii ) it allows to autocalibrate the data , and iii ) by tracking in an independent manner the instabilities produced at the focal plane and those due to the spectrograph , it is possible to remove noise correlated to psf characteristics .
we performed the observations with the 4.2 m william herschel telescope during 2004 august 17/18 using the integral system ( arribas et al .
1998 ) connected to wyffos ( wide field fiber optic spectrograph ; bingham et al . 1994 ) .
we used the recently installed wyffos long camera , equipped with two eev-42 - 80 thinned and coated ccds butted along their long axis to provide a 4k @xmath6 4k pixel mosaic .
the high speed read out mode ( 45 sec ) was used , with gains of 2.3 e- ( chip1-red ) and 1.94 e- ( chip2-blue ) yielding a readout noise of 6 e- .
these detectors are considered linear for standard observations until @xmath052 k adu ( see below ) .
note that wyffos is mounted in one of the nasmyth platforms of the wht , which provides good mechanical stability ( see below ) .
the bundle of fibers sb1 ( standard bundle 1 ) , consisting of 205 fibers , each 0.45@xmath7 in diameter , was used .
this bundle is arranged such that 175 fibers cover a rectangular area of 7.8@xmath8 while 30 additional fibers , forming a ring of 90@xmath7 in diameter concentric with the rectangle , measure simultaneously the sky background .
the actual distribution of the fibers at the focal plane can be found in figure 6 of arribas et al .
( 1998 ) and also in figure 1 of del burgo et al . ( 2000 ) .
the spectra were taken using the r316r line / mm grating with an effective resolution of @xmath03 and covering the @xmath0400010000 spectral range .
integral does not have an independent focus mechanism for the fiber bundles , so the only possibility to spread the light over the input of the fiber bundle in a controlled way was to defocus the telescope .
however we decided not to apply any major defocus to the telescope , because it would have affected the guiding quality .
the observations on target consisted of 8 time series of 20 individual exposures of 120 sec each ( the 8th series had to be suspended after 10 individual exposures were taken due to the detectable flux increase due to the twilight ) . after each time series we took a short flatfield exposure .
therefore , the total object data set consisted of 150 individual exposures ( with a potential of @xmath030000 individual spectra though , of course , many had low s / n ) , extending for a period of more than 7 hours , from which 5 hours were integrating on the target .
the weather conditions were good , with a stable seeing of @xmath00.8 arcsec .
the values of atmospheric extinction in the visible ( @xmath9-band ) obtained by the carlsberg meridian telescope ( http://www.ast.cam.ac.uk/@xmath0dwe/srf/[email protected] ) also suggest stable and photometric conditions for the night , with a mean value for the extinction coefficient of @xmath11 magnitudes per airmass using data expanding a period of 4 hours .
non - photometric data were not detected this night .
we located the object slightly off - center ( i.e. around fiber number 111 in the standard code for this bundle ) in order to minimize the effects of one broken fiber located in the opposite direction
. we will refer to it as _ central fiber _ or _ fiber 111 _ , and it is the closest to the peak of the image in most of the exposures .
the 37 fibers around it covered most of the target flux . apart from lamp - flats taken between sets of 20 exposures , bias , wavelength calibration , and sky flat images for calibration purposes were taken at the beginning and at the end during the twilight . the internal calibration illumination does not mimic the exit pupil of the telescope , producing slightly larger fiber images .
this should be irrelevant provided the use of the these calibration images for the present application ( see next section ) .
the reductions of these ifs data have the goal of generating both spectra ( from which the relative spectrophotometric index is derived ) and images ( from which psf characteristics are obtained ) for each of the 150 individual exposures . here
we will briefly describe the steps followed during this process , commenting on the relative importance of each step for this particular application .
the present study will be focussed on the analysis of the nad lines . as mentioned in section 1 these lines
have been studied with great accuracy with hst ( charbonneau et al .
2002 ) , which provide us with a good reference to compare the present observations .
for this reason , and with the aim of simplifying and making the reduction process more manageable , we trimmed our original images into a spectral region of @xmath0200 including the above mentioned resonant lines .
the bias value was obtained from the overscan strips signal . for the red chip used here ,
the mean value in that region increased linearly over the night from 1018 to 1028 adus .
the rms of this relation is about @xmath12 1 adus , a very small value compared with the typical signal of the brightest spectrum ( @xmath040000 adu ) .
this indicates that the expected errors associated with an incorrect bias subtraction should have a minimal effect on the relative photometry of these very high signal to noise observations . by construction , in many ifs systems is not possible to illuminate directly the detector with a uniform source ( e.g. blank sky ) .
this is due to the presence of the ifu ( i.e. fiber bundle , image slicer , microlens array ) in the optical path .
this prevents obtaining directly flats fields for correcting pixel to pixel sensitivity variations .
instead _ flats _ usually refers to images obtained when illuminating uniformly the input of the ifu , and sometimes are called _ flat - spectra_. these are generally used to correct both , differences in throughput among the different sampling elements ( i.e. fibers ) , and sensitivity variations of _ extracted pixels_. note that an extracted pixel is generally obtained by integrating several individual pixels across the dispersion direction , so they represent an averaged ( weighted ) behavior .
although this approach is generally enough for most applications , in some cases it may represent a limitation .
we will show below that our spectra drift about 0.6 detector pixels during the 7.5 hours of the observations , so different pixel sensitivities could emulate object flux variations .
although it is not possible to obtain direct detector - flats with integral / wyffos , its detector is also used at the prime focus of the wht for direct imaging .
fortunately this detector was used for several imaging programs during september 2004 .
so we collected the flats images recorded during this period ( kindly provided to us by d. lennon ) to create a very high s / n _
superflat_. for the spectral range considered here we combined about 40 images ( mostly in the v filter ) and corrected them from low spatial frequency variations using the iraf tasks _ imcombine _ and _ flat1d _ , respectively .
note that although relatively large fringing amplitudes have been measured at longer wavelengths ( @xmath13 ) , the wavelength range analyzed in this paper is free of that effect .
this superflat has about @xmath14 adu ( @xmath15 ) and , therefore , a nominal s / n per pixel of @xmath01500 .
the photometric index at a given time is obtained after averaging the individual values obtained from the spectra recorded at that particular time .
the fact that the relative intensity of these individual spectra may change during the night ( due not only to guiding and seeing instabilities , but also to the unavoidable differential atmospheric diffraction ) , makes the spectral index sensitive to detector linearity deviations .
the linearity of the present detectors was studied by tulloch ( 2001 ) .
we have fit a polynomial function to his experimental data with signals @xmath1238 k adu ( rms of 0.14 percent ) . for signals
@xmath1638 k adu , the slope of the linearity curve changes sharply from its relatively smooth behavior at lower signals .
this slope change is difficult to model in detail , so we decided not to correct for this linearity slope change ( i.e. we apply a constant correction for values @xmath1638 k adu ) .
linearity corrections at the lower signal end ( i.e. @xmath171000 adu ) are also difficult to model with high accuracy , but effects on the relative photometry are small . in section 5 , we will analyze and correct for the residuals of this linearity correction .
the detector readout noise has a minimal effect in these very high signal to noise observations , we therefore decided to use relatively wide extraction windows of 14 pixels .
this width guarantees the collection of most of the flux from a fiber / spectrum , but still avoids overlapping information from different spectra .
in addition , taking into account the relatively small drift of the spectra in cross dispersion over the night ( @xmath00.6 pixels in 7.5 hours ) , we decided to use a fixed extraction window for all the spectra / exposures .
( however , we also reduced the data allowing for global shifts in the location of the extraction windows , but no appreciable changes were detected . ) the functions defining the centers of the extraction windows were obtained by adjusting polynomials to the spectra of the image obtained by adding all the flat - lamp exposures ( after previous removal of the cosmic rays ) .
it was checked that these spectra were actually located near the mean position during the night .
these functions were 5th degree polynomials obtained after fitting 130 data points along the selected spectral range .
typical rms were of 0.015 pixels ( no significant improvement was found when using higher degree polynomials ) .
the extraction was done as a simple summation ( i.e. not with psf shape and inverse variance weighting ) in order to avoid undesired renormalization factors which could affect the flux .
although these factors should have a very small effect in standard observations , one could think that they may have some relevance for the current ultra - precise photometry .
therefore , we decide not to use this approach which , in addition , has very little advantage in a high s / n regime . in any case
, we checked that no relevant differences were found when an optimal extraction was performed .
for the present analysis the wavelength calibration is important to define accurately the bands used for creating the spectral index .
although the absolute conversion from pixels to wavelength coordinates does not need to be particularly accurate , it is important to correct any drift or change of the wavelength solution among fibers and , especially , during the night . for obtaining the absolute wavelength calibration for each fiber we used the wavelength calibration exposures obtained for such purpose .
that was done using as reference the brightest spectrum with the iraf task _
calibrate_. then , the individual wavelength solutions for each fiber were obtained with _
recalibrate_. to detect any drift along the dispersion direction we cross - correlated the three brightest spectra in the images obtained along the night .
these spectra are located in the region of the detector where most of the signal is extracted .
we find a similar behavior in the three cases with a linear shift of @xmath00.3 pixels over the night at a constant rate of @xmath00.04 pix / h , with a typical rms of 0.05 pixels . therefore , to improve the accuracy of the drift associated with each exposure we combined all the spectra obtained in a given exposure , and cross - correlated these integrated spectra .
once the wavelength solution ( and spectral drift ) was known , we decided not to resample the individual spectra to a common spectral range and dispersion .
this step would have implied an extra interpolation of the data .
we found it preferable to use this information to define homogeneously the spectral bands of all the spectra ( see section 5.2 ) .
the wavelength solution was applied with the task _ dispcor _ of iraf .
the cosmic ray rejection is an important step in the reductions , since a relatively small number of badly removed cosmic rays could degrade substantially the s / n of the data . after trying several automatic rejection methods we decided that none of these methods was sufficiently reliable for our particular case .
the reason for this is that , because the spectra are well separated on the detector , slight shifts of the image at the focal plane may produce relatively important changes in the flux at the detector , which may be interpreted as cr when different exposures are compared .
therefore we decided to follow the secure method of removing the cr in an interactive manner . in order to identify more easily cosmic rays ,
each individual spectrum was divided by a high s / n clean template , so even if a cr hits in the middle of the spectral feature ( i.e. the nad lines ) it could be removed reliably .
after the cosmic rays were identified and removed the spectra were multiplied by the template .
a total of @xmath080 cosmic rays were identified and removed in the relevant spectral range but , of course , the full set of over 5000 spectra were inspected .
flat - spectra were applied to correct for fiber sensitivity .
this was done using the blank sky observations during the twilight at the beginning and at the end of the night .
although the fiber - to - fiber sensitivity variations are large ( rms @xmath020 percent ) these are stable during the night , with typical differences of smaller than @xmath01 percent ( i.e. the difference of two flats taken at the beginning and at the end of the night gave a rms of 1.1 percent for the 37 fibers used , which probably imposes an upper limit to the actual fiber throughput stability ) .
note that global fiber sensitivity variations have no impact on the determination of spectral indices , which always probe relative flux variations ( i.e. band1/band2 ) .
a temporal variation in the fiber throughput with wavelength could be a limitation , but we do nt have evidence for such a behavior .
( as we will see in section 5 , changes in the throughput - wavelength function seem to be dominated by atmospheric more than for instrumental effects ) .
although the photometric index is not dependent on the correction of fiber throughputs , this affects the star image reconstruction .
in fact errors in the fiber throughput corrections may produce incorrect psf reconstruction , introducing further uncertainties during the decorrelation process .
the throughput correction may also affect the background subtraction .
however , in these two cases the effects on the final results are very small ( see below ) .
as described in section 3 , 30 fibers located in a ring of 90@xmath7 in diameter around the star allowed us to obtain contemporaneous background information .
the spectra of some of these fibers are located relatively close to fibers with very high signal , so they may be affected by cross - talk ( or scattered light ) .
therefore , the averaged value of only 18 of these fibers was actually used to obtain the background information . in order to obtain the star images , we created files with the x and y positions of the fibers in the focal plane of the telescope and the integrated ( background - subtracted ) flux in the range 5829.2005957.321 used for the definition of the photometric index in section 5 ( i.e. @xmath18 ) . with the help of the nag routine e01saf , an interpolating 2-dimensional surface f(x , y ) is generated .
this routine guarantees that the constructed surface is continuous and has continuous first derivatives .
the interpolant f(x , y ) was then evaluated regularly each 0.05@xmath7 with the routine e01sbf on a grid of @xmath19 pixels to create the maps presented here .
only 37 fibers were considered to have enough signal for the definition of the photometric index ( see section 5.2 ) , and these were also the ones used for the generation of the images .
( they cover a hexagon centered on the fiber closest to the image peak ) .
figure 2 shows an example of such a reconstruction .
the visual inspection of these images immediately allowed us to identify the loss of lock in the guiding system for four exposures , something confirmed later by the keywords in the headers of these images .
the images were adjusted to 2d gaussians with the help of the task n2gaussfit " of iraf .
it is well known that the seeing function does not fit a gaussian in the outer parts ( wings ) .
therefore , we restrict these fits to a relatively small box of 0.8 arcsec ( the estimated seeing ) . as a result of these fits we obtained the evolution of the following parameters : position ( x , y , position with respect to the central fiber ) , fwhm ( semi major axis ) , amplitude , ellipticity , and pa of the major axis .
note that for this particular application we are not interested in the absolute value of these variables ( which depend on the assumptions about the function assumed ) , but only on their relative evolution along the night .
we studied the variation along the night of a set of auxiliary variables , which will be compared with the photometric index defined in the next section .
these variables were obtained from the ing meteorological data base , from the reduced spectra , and from the reduced images . as commented above
, an important advantage of the ifs ( especially the optical fiber approach ) is the possibility to track independently the variations produced at the focal plane of the spectrograph ( due , for instance , to instabilities of the mechanical structures , changes in the spectrograph focus , etc . ) and those produced at the telescope focal plane ( seeing changes , guiding instabilities , sky level variations , etc . ) .
figures 35 show the changes of these variables along the night . in figure
3 we show the change in air temperature , humidity , atmospheric pressure , sky level , and airmass . in this figure
we can see that the object was well located on the sky for these observations , and we could observe it for more than 7 hours at an airmass lower than 2 .
the sky background level was somewhat unstable at the beginning of the night . in four images
there was a relatively sudden drop ( @xmath0 + 100 minutes , in the time scale of the figure ) , but it was later discovered that for these images we lost the guiding system , so it was a spurious result ( deviant results from these images were also found for other variables see below ) .
the background has a general trend to decrease over the night .
the last two points in the panel indicate the increase due to the twilight .
the values for the air t , humidity , atmospheric pressure and wind intensity ( not shown ) indicate that the night had good stable conditions .
measurements for the seeing outside the dome were below 0.8 arcsec most of the night .
figure 4 shows a set of variables inferred from the spectra .
the two lower panels illustrate the drifts of the spectra during the night along and across the dispersion directions ( due to the motion of the image of the slit with respect to the detector ) .
we can see that the spectra move at a relatively constant rate of 0.08 pix / h ( across ) and 0.03 pix / h ( along ) .
although these are very small changes , drifts of the spectra over the night make the observations sensitive to flat - field corrections , and extraction aperture location .
we did not find appreciable relative motion among the spectra , nor global rotation .
figure 4 also shows the evolution of the fwhm of the spectra integrated over the central band of the index ( i.e. @xmath20 , see below ) .
this also shows a monotonic small change during the first part of the night , to reach a stable value for the second part of the night . the two deviant points ( t @xmath0 + 100 m. ) correspond to the images for which the guiding system was lost ( the relative change in the spectra intensity produces a small but appreciable change in the fwhm ) .
the variable denoted @xmath21fwhm represents the relative change in the fwhm between the line ( @xmath20 ) and the continuum ( @xmath22 ) bands used for the definition of the index as it will be explained below .
although very small , we can appreciate a small trend .
the variation of the total flux integrated over the 37 fibers and over the total spectral range used for the definition of the photometric index ( i.e. @xmath23 ) is also represented , together with the individual @xmath24 and @xmath25 fluxes .
as we will see , the variation of these fluxes is clearly correlated with the star position in the bundle , seeing , and airmass .
finally , the top panel represents the ratio of the two continua used for the definition of the index ( i.e. @xmath26 ) .
figure 5 shows how the variables obtained after fitting the reconstructed images with a simple 2d gaussian model change over the night .
this figure shows a slight drift of the star image over the fiber bundle ( @xmath00.4 arcsec ) during the first three hours until the point in which the auto - guiding was lost , to remain relatively constant the rest of the night .
also these panels show a clear correlation between the ellipticity and the fwhm , and an anticorrelation of these two variables with the amplitude of the gaussian fit and total flux .
this behavior is expected due to the fact that , under these good seeing conditions , the fiber bundle undersamples the psf .
we define a photometric index probing the nad lines as : @xmath27 where @xmath28 , and @xmath25 represent the sky subtracted flux ( on a scale of photo - electrons ) in the following spectral ranges : 582.9200588.5643 , 588.5643590.0887 , and 590.0887595.7321 nm , respectively .
therefore , the nad feature is well centered within @xmath20 , while @xmath29 and @xmath25 represent symmetrical blue and red continua , respectively .
note that this index is insensitive to global ( i.e. grey ) variations in the flux .
also note that by increasing the width of bands @xmath29 and @xmath25 ( and assuming stable continuum values ) one would reduce the standard deviation of a distribution of index values .
this apparent gain in accuracy is obviously compensated by the fact that the index is less sensitive to changes in @xmath20 .
so , once the width of @xmath29 and @xmath25 are large enough compared with @xmath20 ( so the statistical noise fixing the continuum is small compare to that due to @xmath20 ) , no real gain is obtained by increasing the width of those bands ( and further systematic errors can be introduced ) . as defined in eq .
( 1 ) , the index @xmath30 will probe the nad lines under the assumption that @xmath29 and @xmath25 are stable continua .
charbonneau et al .
( 2002 ) have reported very stable continuum around the nad lines , so one could only expect temporal variation of @xmath29 and @xmath25 due to instrumental or atmospheric conditions . some authors ( e.g. winn et al . ,
2004 and references therein ) have found relatively large time - dependent instrumental ( blaze function ) variations , which were interpreted as caused by flexure of the spectrograph .
however , in the present case , our fiber - fed spectrograph is mounted on a stable platform and , therefore , it is not affected by motions of the rotator . in any case , and in order to check instrumental stability , we have displayed in figure 4 the fluxes in the bands @xmath29 , @xmath20 , and @xmath25 as a function of time .
no apparent differences among these three plots can be directly observed , ruling out instrumental variations in flux of the order of 510 percent , as reported by other authors . in the top panel of figure 4
we have also represented the ratio @xmath26 , which has the expected much lower ( @xmath171 percent ) variation with time ( i.e. airmass ) due the differential atmospheric extinction between the two bands . from eq .
( 1 ) we can determined how relative changes in @xmath20 translate into relative changes in @xmath30 .
specifically , @xmath31 and , taking into account that @xmath32 , @xmath33 thus , a relative change of the flux in the nad band translates into a relative change @xmath0 4 times smaller for the index , with the opposite sign .
so , taking into account that in a typical exposure we collect about @xmath34 in @xmath20 , the standard deviation due to photon noise for @xmath20 would be @xmath35 and , therefore , for the index @xmath30 should be @xmath0 @xmath36 .
in addition to the index as defined above , we also evaluated a _
control _ index using the same definition as in eq .
( 1 ) , but a relatively line - free band for @xmath20 .
specifically the bands selected for @xmath29 , @xmath20 , and @xmath25 were : 590.300595.146 , 595.146596.454 ,
596.454601.300 nm , respectively .
the index as defined above was calculated for the 37 brightest spectra of each exposure .
we excluded the four exposures for which we lost the the auto - guiding lock , and the four exposures taken at airmass larger than 2 .
the resulting 5254 ( 37 spectra @xmath6 142 exposures ) values are represented in figure 6 ( top ) versus the total flux in the band ( i.e. @xmath18 ) .
clearly there are some systematic deviations from the behavior expected in the case that noise were only due to photon noise .
this is particularly obvious for the largest @xmath37 values , which are clearly non - linear . at the lowest flux levels the deviations
can not be explained by the expected systematic trend in a ratio due to statistical errors .
they are likely due to a small constant background due to scattered light and , perhaps , some non - linearity residuals too . in any case
, these effects can be well calibrated using the data themselves ( i.e. _ autocalibration _ ) . to this aim
, we fit all the data points to a smooth function ( splines3 of order 16 ) of the total flux , @xmath38 , and we redefine the index as @xmath39 . the corrected photometric index , @xmath40 , for the 5254 spectra are represented as a function of the total flux in figure 6 ( bottom ) .
these new @xmath40 values should be free of the non - linearity residuals commented above , since the dependence with @xmath37 has been removed .
the distribution of values in figure 6 was also used to check and select the proper methodology during the reduction process .
for instance , we studied the effects of resampling the spectra to a common pixel - wavelength solution by analyzing the scatter in this figure .
it was found that the scatter was appreciably larger when the spectra were resampled .
this plot was also used to check the importance of the flat - field and linearity corrections .
it was found that such corrections do indeed improve the quality of the results ( i.e. the scatter in figure 6 is reduced ) , but by relatively modest amounts .
the next step was to obtain the photometric index value for each exposure ( time),@xmath41 .
this was done averaging the 37 contemporaneous individual values obtained in each exposure .
we use an inverse variance weighting of the index ( which is proportional to the total flux , @xmath37 ) in order to take into account the statistical errors of each individual @xmath42 value .
therefore , @xmath43 where the subindex i runs over the 37 spectra obtained simultaneously in each exposure , and @xmath44 is the corresponding total flux ( in photo - electrons ) .
note that apart from the possibility to perform the linearity calibration mentioned above , this procedure for generating the index is more advantageous than adding all the 37 spectra and generating the index from the combined spectrum , in terms of the influence of backgrounds ( i.e. sky , scattered light , etc . ) since these are actually weighted by the importance of each spectrum for the generation of the index at a given time .
also this method is also less sensitive to the actual spatial extend used ( i.e number of fibers involved ) .
thus if , for instance , a number larger than the 37 fibers considered were used , the extra spectra would have had very little weight in @xmath41 , as a consequence of their low flux .
figures 7 and 8 show the dependence of the index with some of the auxiliary variables . in table 1
the degree of correlation is indicated in 4th column .
figure 7 and table 1 show that the strongest correlation ( 0.76 ) is found against the airmass ( which contrasts with the weak anticorrelation [ @xmath450.19 ] , shown by the ` _ control _ index ' , when @xmath20 is defined outside the nad lines , as indicated in section 5.2 ) .
initially we considered the possibility that this strong correlation between the index and the airmass could be due to geographical and temporal variations of the emission spectrum of the nad lines produced in the earth s atmosphere .
however , the emission spectrum has a very small contribution compared with the signal .
in fact , considering a typical sky brightness of v = 21.8 mag arcsec@xmath46 for la palma and an ew for the telluric emission of nad of @xmath0150 ( benn and ellison 1998 ; pedani 2004 ) , and the parameters used in our experiment ( e.g. width of @xmath47 , fiber collecting area @xmath00.16
arcsec@xmath46 , etc . )
, one should collect more that @xmath48 photons from hd209458 for each photon coming from the nad emission lines . on the other hand , charbonneau et al .
( 2002 ) have shown that the expected drop in the nad band during the transit is @xmath02.3@xmath49 .
so the amount we expect to detect during the transit is much larger ( at least by one order of magnitude ) than that due to the sky emission ( the difficulty , of course , is in recording that signal on top of the photon noise generated by the star ) .
therefore , the strong correlation between the index and the airmass is likely to be dominated by the absorption produced by the earth atmosphere .
the fact that the index has a relatively strong correlation with airmass ( which obviously change with time ) , may induce correlations with other variables which show a temporal variation , for which there is not a causal connection . by _ decorrelation
_ we understand the process of removing trends in the index correlated with the auxiliary variables .
this was based on a multiple - regression ( least - square ) fit to the data with a function which is linear in the coefficients of independent variables , following the prescriptions given by bevington ( 1969 ) .
column 4th of table 1 shows the correlation parameter for the 15 selected auxiliary variables . in order to have an idea of the effects of the decorrelation procedure reducing noise
, we can compare the global standard deviation of the 142 data values before ( @xmath50 ) and after ( @xmath51 ) this procedure was applied .
note that the standard deviation of the decorrelated values is remarkably close to the one expected in a purely photon - noise dominated regime .
in fact , considering a typical level of counts per exposure of @xmath52 adu ( counts ) for @xmath20 , a gain of @xmath53 , and eq .
( 3 ) , a standard deviation of the index of @xmath0@xmath54 is expected in a photon - noise dominated regime .
( it is interesting to note that the fact that the decorrelated signal shows a standard deviation close to that due to the photon noise suggests that the modal noise is very small or it can be well tracked by the selected variables ) .
as we can see in table 1 , the airmass is the variable which shows the highest correlation with the index .
if the data are only decorrelated linearly against the airmass ( i.e. if only the linear trend observed in the bottom panel of fig . 7
is removed ) the standard deviation is @xmath55 . if we decorrelate against all other variables , but we exclude the airmass the standard deviation is @xmath56 . in figure 9 ( left )
we show the original data ( upper panel ) , the fit to a function linear to the airmass ( center ) , and the decorrelated data ( lower ) , as a funtion of time .
figures 9 ( center ) and ( right ) are similar but fitting the data to a function linear to 14 ( all but the airmass ) , and to all the 15 selected auxiliary variables , respectively . it is interesting to note how the rest of the variables also carry information about the effects of the airmass . if we select for the decorrelation the six auxiliary variables derived from the spectra ( background , shift x , shift y , fwhm , @xmath21fwhm , total flux ) the standard deviation is @xmath57 .
if we select the nine variables from the images ( background , peak , shift x , shift y , distance to central fiber ( 111 ) , fwmh , ellipticity , p.a . ,
total flux ) the standard deviation drops to @xmath58 .
this illustrates one of the advantages of using ifs : having control of the image characteristics at the focal plane offers the possibility to improve the photometry by removing correlated noise .
although this decorrelation method can be applied safely if the functional dependence of the auxiliary variables are different to that of the signal to be detected , it may destroy part of it when that is not the case .
this is more likely when many different variables are involved in the decorrelation process .
also the fact that our observations include only one transit , and this represents a relatively large fraction ( @xmath030 percent ) of the whole observed period makes this possibility more likely .
ideally , observations during long out - of - transit periods will characterize more accurately the dependence on the variables . obviously having data for multiple transits will minimize the possible artificial attenuation of the signal during the transit . in any case , to study the degree of destruction of the signal during the transit due to the decorrelation process we proceed as follows .
similarly to brown et al .
( 2002 ) we injected an artificial signal to the data ( during the period that the transit took place ) before the decorrelation was applied , and we analyzed how much of this signal was recovered after decorrelation . in a first step the effects of each auxiliary variable were analyzed individually .
that was done comparing the input and output signals after decorrelating against each of the 15 variables .
column 5th in table 1 indicates the fraction of signal recovered , when the decorrelation is done using only this variable . in an attempt to estimate the error for these conservation factors we used different levels of injected artificial signal ( i.e. 0.5 , 1 . and 2 times the expected change according to the stis
hst result ) , but the conservation factors were similar for these three cases .
table 1 shows that in several cases more than 80 percent of the signal is recovered , while in other cases this is less than 30 percent .
this is the case of three variables inferred from the spectra ( shift x , shift y , fwhm ) .
this is likely due to their monotonic behavior along the night , something that can mimic a transit occurred during the first part of the night as it was in the present case .
however , a variable with a different type of dependence such as , for instance , the airmass ( which decreases and increases along the night , with the inflexion point offset with respect to the transit ) better preserves the signal . during the 132 minutes considered as in - transit observations ( i.e. excluding the ingress and egress ) the airmass changed from 1.242 to 1.019 .
however , after the pass through the meridian , the change in airmass from 1.019 to 1.242 ( and on ) was done out transit .
then we studied the combined effects of several variables .
we found that when all the 15 variables are used to remove noise ( i.e. decorrelation process ) most of the signal was actually destroyed : only 17 percent of the signal artificially injected is recovered after decorrelation .
the next step in the analysis was to select a subset of auxiliary variables , which lead to a significant reduction of noise while preserving a substantial fraction of the signal .
after several trials based on the individual destruction factors , we found that three variables ( airmass , fwhm of the image , distance to 111 ) give a good compromise .
specifically , the standard deviation of the 142 points was @xmath59 , while 79 percent of the signal is preserved after decorrelation .
although the addition of fwhm and distance to 111 implies a relatively modest improvement in the standard deviation with respect to the decorrelation against the airmass only , these variables were included because they take into account the position and shape of the image on the fiber bundle , which show a relatively strong correlation with the total flux ( see fig . 5 ) . in any case , we checked that this selection did not affect our results , and values after decorrelating against the airmass only will also be given hereafter .
at this point we evaluated the effects of the autocalibration ( see section 5.3 ) on the final standard deviation . if the decorrelation procedure is applied to the data without applying the autocalibration step a standard deviation of @xmath60 is obtained , which compares with the value of @xmath51 mentioned above .
( values for which the total flux @xmath37 was larger than @xmath61 , were excluded since they show saturation see figure 6 ) .
this demonstrates that , the autocorrelation step allows to improve the photometric accuracy as well as to recover some of the saturated values ( improving the temporal coverage , and the error of the mean value ) .
figure 10 ( upper panel ) shows the time series data values .
vertical lines indicate the contacts .
following charbonneau et al .
( 2002 ) we define in - transit observations as those that occurred between second and third contact ( i.e. @xmath62 m. ) , and out - of - transit those that occurred before first or after fourth contact ( i.e. @xmath63 m. ) . there are 45 in - transit observations and 84 out - of - transit observations , ignoring those observations that occurred during ingress and egress . as mentioned above the standard deviation of all the 142 values represented in figure 10
is @xmath59 .
however , the standard deviation in - transit ( @xmath64 ) is significantly larger than out - of - transit ( @xmath65 ) .
we think this is attributable to the more unstable conditions of the atmosphere / background at the beginning of the night .
although our meteorological data do not show strong evidence for that ( see figure 3 ) , variations of the telluric vapor absorption ( e.g. lundstrom et al .
1991 ) and/or of the mesospheric sodium column density ( e.g. ge et al .
1998 ) could produce these instabilities .
in any case it is remarkable the fact that out - of transit , and considering data that expand for a period of @xmath0 4 hours , the standard deviation is very similar to that expected from poison - noise .
the lower panel in figure 10 shows the mean in - transit and out - of - transit values of the index : @xmath66 and @xmath67 , respectively .
the quoted 1-@xmath1 errors correspond to the error in the mean of the statistical distribution ( i.e. standard deviation/@xmath68 ) .
( if the index is decorrelated against the airmass only , the corresponding values are @xmath69\cdot10^{-5}$ ] and @xmath70\cdot10^{-5}$ ] , respectively . )
these results are compatible with no variation of the index during the transit ( i.e. the relative variation is @xmath71\cdot10^{-5}$ ] ) and , taking into account equation ( 3 ) , no variation of the relative the flux in the nad lines ( i.e. @xmath20 ) with respect to the continuum during the transit .
we also show in this figure the expected change in the index due to a change in @xmath20 as the one detected by stis
hst ( charbonneau et al . 2002 ) .
provided that these authors found a relative change in @xmath20 of @xmath72 during the transit , according to equation ( 3 ) one should have expected a relative change in the index of @xmath73 with a conservation factor of 1 , and @xmath74 with the current conservation factor of 0.79 ( see section 5.4 ) .
( this neglects the slightly different width of the 12 band used by charbonneau et al . with respect to the 15 band used by us .
this is justified by comparing their results for 12 and 38 bands see table 2 : a change of 3 in the band width would imply a change @xmath75 in the index , about a factor of 5 smaller than the error quoted by them ) .
therefore , if a signal with an amplitude as the one reported by charbonneau et al . were real , and after repeating the current experiment many times , the ( gaussian ) distribution of results should peak at @xmath74 with a sigma of @xmath76 .
however , we ( and charbonneau et al . ) have performed the experiment only once , and the obtained result is about 2 @xmath1 away from the expected value ( for a signal as the one detected by charbonneau et al .
if we take into account the quoted errors , the present data and the hst / stis results disagree by more than 1 @xmath1 , but they are compatible within 1.5 @xmath1 .
this limited statistical significance makes us cautious to over interpret the present results , which are about a factor 2 more uncertain than the hst data . since atmospheric variables , and especially the airmass ,
dominate the decorrelation procedure , a better knowledge of the typical time scale for variation of the differential absorption in the nad band by the earth s atmosphere would be useful to further analysis .
however , note that we could have obtained significantly better accuracy if the atmospheric conditions during the transit would have been similar to out - of - transit . in that case
, the current @xmath77 1-@xmath1 ) error would have dropped to @xmath78 .
in addition , obviously , multiple transit observations would lead also to a significant reduction of this error .
the recent ground based observations by narita et al .
( 2005 ) could not confirm nor contradict the charbonneau et al . result .
although these subaru observations have much higher spectral resolution , its lower s / n and temporal resolution lead to an error in the predicted relative flux in nad lines ( for a 12 band ) larger ( by a factor of about 4 times ) than that of the current observations .
the main conclusions of the present work can be summarized as follows : 1- we have shown that integral field spectroscopy is a powerful technique for obtaining accurate relative time series photometry and , therefore , of great potential for studying transits of extrasolar planets .
the three main advantages with respect to previously used methods ( based on imaging or standard slit spectroscopy ) are : i ) it allows an increase in the duty cycle of the observations ( and , therefore , the s / n ) by distributing the light into the two dimensions of the detector ; ii ) the data can be _ autocalibrated _ from detector non - linearities and from background effects ; and iii ) since the photometric index and the image of the star are extracted from the same data - cube , the noise correlated with psf characteristics can be removed .
furthermore , the proposed ifs method should not have major demerits with respect to standard slit spectroscopy .
although the present study is based on ground based observations , the full potential of ifs for these type of observations will be reached from space ( e.g. jwst , tpf ) .
2- these observations have also allowed us to explore the accuracy limits that can be achieved from the ground for these type of observations .
in particular , using a photometric index that probes the strength of the nad lines , a standard deviation of @xmath50 was found ( during a @xmath07 hour period ) , which is about a factor 2 larger than that expected due to photon noise . however , after removing correlated noise by fitting the data to a function linear in three selected auxiliary variables the standard deviation drops to @xmath59 , only @xmath025 percent larger than that due to photon noise .
simulations indicate that 79 percent of the relative flux change during the transit should be preserved after decorrelating against those three variables .
our nominal 1-@xmath1 error for the mean in - transit ( with respect to the mean out - transit value ) for the index , is equivalent to @xmath79 for the relative flux in the nad band ( @xmath20 ) with respect to the continuum .
this is substantially better than previous ground observations and only a factor @xmath02 times worse than the 4-transit observations with stis hst .
3- the present mean values for the photometric index in transit and out - of - transit are compatible with no extra depression of the nad during the transit . although the present results are in apparent contradiction with the hst stis results , we stress the limited statistical significance of this disagreement : they disagree at 1-@xmath1 , but they are consistent within 1.5-@xmath1 .
we acknowledge that our results have larger nominal errors , and are more susceptible to systematically underestimate the signal through decorrelation .
4- the present study indicates that a good level of accuracy can be obtained from the ground with this type of ifs observations .
the 5254 spectra analyzed here were collected during only one night / transit and , therefore , are insufficient to reduce the nominal errors to the level of the four transit stis hst data .
multiple transit observations will also allow to analyze the typical time scale variations for relevant atmospheric variables and , therefore , better understand the effects of the decorrelation procedure on the photometric index .
5- from an instrument design perspective , this application of ifs can be fostered if : i ) the fiber bundle ( image slicer , array of micro - lenses ) have an independent focussing mechanism , ii ) direct illumination of the detector is possible to generate flat fields , and iii ) the sky is simultaneously recorded .
fiber systems seem to have some generic advantages with respect to another type of ifu ( e.g. azimuthal scrambling of the light within the fibers , the possibility to use static spectrographs , etc . ) .
furthermore , our data indicate that modal noise or the temporal variation of the fiber throughput with wavelength are not significant sources of noise .
further multi - transit observations , with long out - of - transit data ( ideally from space ) should lead to the development of the full potential of the proposed method . for ground based observations , the use of the recently developed new integral field spectrographs for 8 m .
class telescopes ( e.g. gmos : allington - smith et al .
2002 , sinfoni / spiffi : eisenhauer et al .
2003 , vimos : le fevre et al .
2003 ) also opens an opportunity to push this technique forward .
we are grateful to chris benn , ivo busko , katrina exter , phill hodge , danny lennon , jess maz - apellniz , and simon tulloch , who have provided us with useful information .
we also thank the anonymous referee for useful comments .
this paper is based on observations made with the wht , operated on the island of la palma by the ing in the spanish observatorio del roque de los muchachos of the instituto de astrofsica de canarias .
we thank all the staff at the observatory for their kind support .
support for this work was also provided by the spanish ministry of education and science through grant aya2002 - 01055 .
allington - smith , j.r .
2002 , , 114 , 892 arribas , s. et al .
1998 , spie , 3355 , 821 arribas , s , and mediavilla , e. , 2000 , in ` imaging the universe in three dimensions , ' ed .
w. van breugel and j. bland - hawthorn , asp , conference series , 195 , 295 baudrand , j. , and walker , g.a.h .
2001 , , 113 , 858 bevington , p.r .
1969 , ` data reduction and errror analysis for the physical sciences , ' mcgraw - hill book company benn , c.r . , ellison , s.l .
1998 , new astr .
reviews , 42 , 503 bingham , r.g . , gellatly , d.w . , jenkins , c.r .
, worswick , s.p .
1994 , spie , 2198 , 56 borucki et al .
2004 , in ` second eddington workshop : stellar structure and habitable planet finding , ' ed .
f. favata , s. aigrain and a. wilson .
esa sp-538 brown , t.m . ,
charbonneau , d. , gilliland , r.l .
, noyes , r.w . , and burrows , a. 2001 , , 552 , 699 brown , t.m .
2001 , , 553 , 1006 brown , t.m . , libbrecht , k.g . , and charbonneau , d. 2002 , , 114 , 826 bundy , k.a . , and marcy , g.w .
2000 , , 112 , 1421 charbonneau , d. et al .
2000 , , 529 , l45 charbonneau , d. , brown , t.m . ,
noyes , r.w . , and gilliland , r.l .
2002 , , 568 , 377 coustenis , a. , schneider , j. , wittemberg , r. , chazsefiere , e. , guillot , t. , penny , a. , greene , t. , rauer , h. , bockelee - morvan , d. 1998 , in ` brown dwarfs and extrasolar planets , ' ed .
r. rebolo , e. l. martin , m. r. zapatero osorio , asp conference series , 134 , 296 deeg , h. 2002 , 36th eslab symposium , eds .
b. foing , b. battrick .
esa sp-514 , 237 del burgo , c. , mediavilla , e. , and arribas , s. 2000 , , 540 , 741 deming , d. , brown , t.m . , charbonneau , d. , harrington , j. , richardson , l.j .
2005 , , 622 , 1149 dhillon , v. and marsch , t. 2002 , the newsletter of the isaac newton group , 6 , 25 eisenhauer , f. et al .
2003 , spie , 4844 , 1548 ge , j. , jacobsen , b.p . ,
angel , j.r.p . ,
mcguire , p.c . ,
roberts , t. , mcleod , b.a . , and lloyd - hart , m. 1998 , spie , 3353 , 242 gilliland , l.r . , goudfrooij , p. , and kimble , r.a .
1999 , , 111 , 1009 gilliland , l.r . , and arribas , s. 2003 , isr - nicmos-03 - 01 , stsci henry , g.w . ,
marcy , g.w . , butler , r.p . , and vogt , s.s .
2000 , , 529 , l41 le fevre , o. et al .
2003 , spie , 4841 , 1670 lundstrom , i. , ardeberg , a. , maurice , e. , and lindgren , h. 1991 , a&a , 91 , 199 mazeh et al .
2000 , , 532 , l55 moutou , c. , coustenis , a. , schneider , j. , st gilles , r. , mayor , m. , queloz , d. , kaufer , a. 2001 , a&a , 371 , 260 narita , n. , suto , y. , winn , j.n . , turner , e.l . ,
aoki , w. , leigh , c.j . , sato , b. , tamura , m. , yamada , t. , 2005 , pasj , 57 , 471 pedani , m. 2004 , new astr .
, 9 , 641 rauer , h. , bockele - morvan , d. , coustenis , a. , guillot , t. , schneider , j. 2000 , a&a , 355 , 573 tulloch , s. 2001 , isaac newton group technical note , nro . 119 vidal - madjar , a. et al .
2003 , nature 422 , 143 vidal - madjar , a. et al .
2004 , , 604 , l69 winn , j.n . ,
suto , y. , turner , e.l . ,
narita , n. , frye , b.l . , aoki , w. , sato , b. , yamada , t. 2004 , pasj , 56 , 655 1 & airmass & 1.074 & 5.41e-04 & 0.76 & 0.88 + 2 & total flux & 8.69e-04 & 1.33e-10 & 0.14 & 0.88 + 3 & background & 7.74e-02 & @xmath456.61e-07 & 0.48 & 0.75 + 4 & spectra shift ( along disp . ) & 2.11 & 4.64e-04 & 0.38 & 0.33 + 5 & spectra shift ( across disp . ) & 1.60 & 3.60e-04 & 0.29 & 0.32 + 6 & spectra fwhm & 3.16 & @xmath451.87e-05 & 0.11 & 0.38 + 7 & image total flux & 5.82e-03 & @xmath451.01e-08 & 0.16 & 0.88 + 8 & flux - ratio & 1.68 & @xmath455.12e-04 & 0.32 & 0.56 + 9 & gaussian peak & 4.55e-04 & @xmath456.08e-11 & 0.27 & 0.77 + 10 & gaussian x & 0.330 & @xmath452.81e-06 & 8.9e-02 & 0.40 + 11 & gaussian y & 0.480 & 3.20e-05 & 6.3e-02 & 0.46 + 12 & distance to 111 & 0.384 & 3.24e-05 & 0.17 & 0.99 + 13 & gaussian fwhm & 0.422 & @xmath455.55e-05 & 0.27 & 0.84 + 14 & gaussian ellipticity & 1.59 & 6.51e-04 & 2.3e-03 & 0.85 + 15 & gaussian p.a .
& 0.320 & @xmath453.57e-06 & 0.19 & 0.94 + & all & & & & 0.17 + & 1 + 12 + 13 & & & & 0.79 + | we explore the use of integral field spectroscopy ( ifs ) for observing extrasolar planet transits .
although this technique should find its full potential in space observations ( e.g. jwst , tpf ) , we have tested its basics with ground based time series observations of hd209458b obtained with wht + integral during a transit in august 17/18 , 2004 . for this analysis
we have used 5550 spectra ( from a potential of @xmath030000 ) , obtained in 150 exposures during a period of more than 7 hours .
we have found that ifs offers three fundamental advantages with respect to previously used methods ( based on imaging or standard slit spectroscopy ) .
first , it improves the effective s / n in photon limited observations by distributing the light coming from the star into the two dimensions of the detector .
second , this type of ifs data allows to _ autocalibrate _ instrumental and background effects .
third , since the star image characteristics ( i.e. seeing , spatial shifts , etc . ) as well as its photometric properties are extracted from the same data - cube , it is possible to decorrelate photometric instabilities induced by psf variations .
these data have also allowed us to explore the accuracy limits of ground based _ relative _ spectrophotometry .
this was done using a photometric index that probes the nad lines , for which we obtained a nominal 1-@xmath1 error of @xmath0@xmath2 .
this result , based on observations of only one transit , indicates that this type of ground observation can constrain the characterization of the transmission spectrum of extrasolar planets , especially if they cover multiple transits under good weather conditions .
the present observations are compatible with no extra nad depression during the transit .
although this result seems to be inconsistent with the recently reported hst
stis findings we point out its limited statistical meaning : the results disagree within 1-@xmath1 , but agree within 2-@xmath1 .
this method requires careful and specific reductions , and details about this process are given .
we also give some recommendations to instrument developers in order to enhance the efficiency of the method . |
high resolution observations show strong outflows and jets in black hole accreting systems , both in active galactic nuclei ( agns ) or quasars ( begelman et al . 1984 ; mirabel 2003 ) and microquasars ( ss433 , grs 1915 + 105 ) [ margon 1984 ; mirabel & rodriguez 1994 , 1998 ] .
extragalactic radio sources show evidence of strong jets in the vicinity of spinning black holes ( meier et al .
2001 ; meier 2002 ) .
outflows are also observed in neutron star low mass x - ray binaries ( lmxbs ) ( fender et al . 2004 ; migliari & fender , 2006 ) and also in young stellar objects ( mundt 1985 ) .
it has been argued ( ghosh & mukhopadhyay 2009 ; ghosh et al .
2010 ; hereinafter gm09 , g10 respectively , and references therein ) that outflows and jets are more prone to emanate from strong advective accretion flows ; the said paradigm is more susceptible for super - eddington and sub - eddington accretion flows . however , the exact ( global if any ) mechanism of formation of the jet , its collimation , acceleration , composition from the accretion powered systems still remain inconclusive .
extensive works have been pursued on the origin of outflow / jet , since the pioneering work of blandford & payne ( 1982 ) ( e.g. pudritz & norman 1986 ; contopoulos 1995 ; ostriker 1997 ) , where the authors used a self - similar approach to demonstrate that the poloidal component of the magnetic field can be seemingly used to launch outflowing matter from the disk . the formation of the jet is directly related to the efficacy of extraction of angular momentum and energy from the accretion disk .
physical understanding of the hydromagnetic outflows from disks has been developed from magnetohydrodynamic ( mhd ) simulations ( shibata & uchida 1986 ; ustyogova et al .
1999 ; de villiers et al . 2005 ; hawley & krolik 2006 ) both in non - relativistic as well as in relativistic regimes , mostly in the keplerian paradigm .
however , the large lorentz factors as well as faraday rotation measures suggest that the observed vlbi jets in quasars and active galaxies are in the poynting flux regime ( homan et al . 2001 ; lyutikov 2006 ) . on the other front
, strong radiation pressure can serve as a different mechanism to effuse outflows / jets .
this is likely to occur when the accretion rate is super - eddington or super - critical ( fabrika 2004 ; begelman et al .
2006 ; gm09,g10 and references therein ) and the accretion disk is precisely `` radiation trapped '' as in ultra - luminous x - ray ( ulx ) sources . it has been confirmed by radiation hydrodynamic simulation at super - critical accretion rate ( okuda et al . 2009 ) as well to explain luminosity and mass outflow rate of relativistic outflows from ss433 .
the under - luminous accreting sources , having high sub - critical accretion flow , were explained by an advection dominated accretion flow ( adaf ) model ( narayan & yi 1994 ) .
the promising outcome of this model lies in the large positive value of the bernoulli parameter because of the small radiative energy loss .
it leads to conceive that the gas in the inflowing disk is susceptible to escape , leading to strong unbounded flows in the form of outflows and jets .
this also signifies that even in the absence of magnetic field and radiation pressure , outflows are plausible from strongly advective accretion flow if the system is allowed to perturb . nevertheless , the definitive understanding of the origin of outflows / jets is sill unknown .
it remains one of the most compelling problems in high energy astrophysics .
whatever might be the reason for the origin of outflow and then jet from the disk , one aspect is however definite that strong outflows producing relativistic jets are powered by extreme gravity .
although seems paradoxical , the strength and the length scale of observed astrophysical jets vary directly with the strength of the central gravitating potential .
observationally , it is evident that strong outflows and relativistic jets are more powerful in observed agns and quasars , harboring supermassive black holes , compared to that in black hole x - ray binaries ( xrbs ) .
in addition , the jets observed in agns and quasars have greater length scale compared to that seen in stellar mass black hole systems .
one of the most important signatures of relativistic gravitation is the spin of the central object .
it is presumably believed that the spin ( practically specific angular momentum ) of the neutron star is less than that of a black hole , and thus the observed jet from black holes is much stronger and powerful than that of neutron star sources . in
early , blandford & znajek ( 1977 ) demonstrated that if there is a magnetic field associated with the black hole due to threading of magnetic field lines from the disk and the angular momentum of the kerr black hole is large enough , then the energy and the angular momentum can be extracted from the underlying black hole by a purely electromagnetic mechanism , which can thus be expected to power the jet in an agn .
these imply that the spin might play a significant role in powering jets , both in microquasars and in agns , rather , it can act as a fundamental parameter in an accretion powered system .
most of the studies of accretion disk and studies of related outflow / jet have been evolved separately , assuming these two apparently to be dissimilar objects . however ,
several observational inferences ( for details see gm09 , g10 and references therein ) and improved understanding of accretion flow and outflow reveal that accretion and outflow / jet are strongly correlated .
the unifying scheme of disk and outflow is essentially governed by conservation laws ; conservations of matter , energy and momentum .
hence , in modeling the accretion and outflow simultaneously in any accretion powered system , following aspects should be taken into account : ( 1 ) the effect of relativistic central gravitational potential including its spin , ( 2 ) proper mechanism of origin of outflow / jet from the disk , ( 3 ) appropriate hydrodynamic equations ( considering that the accreting gas be treated as a continuum fluid ) , capturing the information about the intrinsic coupling between inflow and outflow which are governed by the conservation laws in a strongly advective paradigm .
recently , gm09 , g10 made an endeavor to explore the dynamics of the accretion - induced outflow around black holes / compact objects in a 2.5-dimensional paradigm .
the authors formulated the disk - outflow coupled model in a more self - consistent way by solving a complete set of coupled partial differential hydrodynamic equations in a general advective regime through a self - similar approach in an axisymmetric , cylindrical coordinate system .
they explicitly incorporated the information of the vertical flux in their model .
however , they restricted their study to newtonian regime , thus negating the requisite effect of general - relativity , especially the effect of spin of the central object . based on the model , the authors computed the mass outflow rate and the power extracted by the outflow from the disk self - consistently , without proposing any prior relation between the inflow and outflow . in the present paper ,
we propose a new model for the accretion - induced outflow by extending the work of gm09 , g10 , by incorporating the general relativistic effect of the central potential without limiting ourselves to a self - similar regime . as the definitive mechanism of launching of outflows / jets
is still evasive , we do not embrace any specific mechanism of outflow like inclusion of the magnetic field into our model equations .
nevertheless , the importance of the magnetic field can not be , in principle , discarded to explain the launching and collimation of jets off the accretion disk . here ,
owing to our inability to solve coupled partial differential mhd equations in a 2.5-dimensional advective regime , we neglect the influence of the magnetic field in our model equations .
however , the implicit coupling between the inflow and outflow , dictated by the conservation equations , have been taken into account appropriately .
we arrange our paper as follows . in the next section
, we present the formulation of our model .
the section 3 describes the computational procedure to solve the model equations of the accretion - induced outflow . in sections 4 and 5 ,
we study the dynamics and the energetics of the flow respectively .
finally , we end up in section 6 with a summary and discussion .
we formulate the disk - outflow coupled model by considering a geometrically thick accretion disk , which is strongly advective as strong outflows / jets are more likely to eject from a thick / puffed up region of the accretion flow ( gm09,g10 ) . the vertical flow is explicitly included in the system .
the basic features of the model are similar to that in gm09 .
we adopt the cylindrical coordinate system to describe a steady , axisymmetric accretion flow .
the dynamical flow parameters , namely , radial velocity ( @xmath0 ) , specific angular momentum ( @xmath1 ) , vertical velocity or outflow velocity ( @xmath2 ) , adiabatic sound speed ( @xmath3 ) , mass density ( @xmath4 ) and pressure ( @xmath5 ) depend both on radial and vertical coordinates . we have already highlighted the importance of the spin of black hole to power the outflow / jet and its presumed relative effect on the observation of various agn classes .
thus we have included the effect of the spin in our model . as the spin of the black hole is a signature of pure general relativity , i.e. , einstein s gravitation , its effect on the accretion flow , especially in the inner region of the disk ,
is mimicked approximately with the use of pseudo - general - relativistic or pseudo - newtonian potential ( pnp ) .
because of the disk - outflow system to be geometrically thick and the flow to be 2.5-dimensional ( not confined to the equatorial plane ) , we use the pnp of ghosh & mukhopadhyay ( 2007 ) , which is a pseudo - newtonian vector potential , to capture the inner disk properties of the accretion flow around a kerr black hole approximately . at the first instant we neglect the effect of viscosity in our system .
one of the reasons behind it is the unavailability of the effective computational technique to solve coupled partial differential viscous hydrodynamic conservation equations for the compressible flow . to make the inviscid assumption more arguable ,
it can also be noted that the angular momentum transport in the accretion flow , for which the necessity of turbulent viscosity is invoked , can alternatively take place purely by outflow . at this extreme end
, the outflow extracts angular momentum from the disk allowing the matter to get accreted towards the black hole and hence the inviscid assumption can be adopted .
this is in essence similar to the blandford & payne ( 1982 ) mechanism to extract energy and angular momentum from the magnetized disk , where the extraction of the angular momentum and energy is essentially done by the outflow , and not due to the viscous dissipation .
the outflow originates from just above the equatorial plane of the disk and this lower boundary is maintained at @xmath6 , when @xmath7 , unlike other work where the outflow is hypothesized to effuse out from the disk surface ( e.g. xie & yuan 2008 ) .
our model is effectively valid only in the region where disk and outflow are coupled , i.e. the region from where essentially the outflow is emanated from the accretion flow .
hence our study will remain confined within this predefined region .
further , we neglect the contribution of the magnetic field as argued in 1 .
we circumvent the idea of vertical integration of the flow equations .
the validity and the reliability of the height integrated equations is normally gratifying in the geometrically thin limit . in that circumstances ,
the flow velocities are likely to be more or less independent of the disk scale - height , which is not the case of the present paradigm of interest .
we further consider that the disk to be non - self - gravitating , assuming that the mass of the disk to be much less than that of the black hole .
the radial and vertical coordinates are expressed in units of @xmath8 , flow velocities in @xmath9 , time in @xmath10 and the specific angular momentum in the unit of @xmath11 . here
@xmath12 , @xmath13 and @xmath9 are gravitational constant , mass of the black hole and speed of light respectively .
the steady state , axisymmetric disk - outflow coupled equations in cylindrical geometry in the inviscid limit are then as follows .
\(a ) mass transfer : @xmath14 where the first term is the signature of accretion and the second term attributes to outflow . as the outflow starts from just above @xmath6 surface , within the inflow region itself , we make a reasonable hypothesis that within the prescribed disk region the variation of the dynamical flow parameters with @xmath15 is much less than that with @xmath16 , allowing us to choose @xmath17 ; for any parameter @xmath18 .
strictly speaking , the weak variation with @xmath15 ensures that the outflow originating from the mid - plane of the disk does not disrupt the disk structure , and thus allowing a smooth accretion flow towards the black hole
. thus eqn .
( [ 1a ] ) then reduces to @xmath19 \(b ) radial momentum balance : @xmath20 where @xmath21 is radial component of the gravitational force . as discussed earlier , here @xmath21 is the radial force corresponding to the pnp in cylindrical coordinate system given by ghosh & mukhopadhyay ( 2007 ) containing the information of spin of the black hole . with the similar argument as above the equation then reduces to @xmath22 where @xmath23 is the radial velocity at the mid - plane of the disk .
\(c ) azimuthal momentum balance : @xmath24 where @xmath25 is the azimuthal velocity of the flow .
the first term of this equation signifies the radial transport of the angular momentum , while the second term describes the extraction of angular momentum due to mass loss in the outflow .
if we consider that the net angular momentum extracted by the outflow be @xmath26 , and the remaining angular momentum retained by the disk @xmath27 , then total angular momentum @xmath28 can be assumed to remain constant throughout the flow within our predefined disk - outflow region by the virtue of an inviscid flow .
therefore , @xmath29
\(d ) vertical momentum balance : @xmath30 where @xmath31 is the vertical component of the gravitational force , described by ghosh & mukhopadhyay ( 2007 ) .
following previous arguments this reduces to @xmath32 where @xmath33 is the pressure of the flow at the mid - plane of the disk . if there is no outflow , then @xmath34 , and eqn .
( [ 4b ] ) reduces to the well known hydrostatic equilibrium condition in the disk , and the hydrostatic disk scale - height can be obtained .
similarly , one can customarily extend the vertical momentum equation to compute the disk scale - height when there is an outflow coupled with the disk .
thus we will use eqn .
( [ 4b ] ) to obtain the scale - height of the disk - outflow coupled system .
we can reasonably assume that at height @xmath35 ( i.e. at @xmath15=@xmath35 ) , the pressure of the disk is much less compared to that at the equatorial plane to prevent any disruption of the disk , permitting a steady structure of accretion flow .
the variation of density along @xmath15 direction is approximately kept constant ( @xmath36 ) which in turn means that the disk - outflow coupled region is weakly stratified . considering the above facts , eqn .
( [ 4b ] ) is simplified to obtain the scale - height as @xmath37 where @xmath35 is thus the the root of the above transcendental equation
. thus eqns .
( [ 1b ] ) , ( [ 2b ] ) , ( [ 3b ] ) & ( [ 4c ] ) will simultaneously have to be solved to obtain the dynamics and the energetics of the accretion - induced outflow .
we assume that the accretion flow follows an adiabatic equation of state @xmath38 where @xmath39 , @xmath40 and @xmath41 are the polytropic and the adiabatic indices respectively .
note that constant @xmath42 carries the information of entropy ( e.g. mukhopadhyay 2003 ) of the flow .
thus for an adiabatic flow , @xmath43 . in an earlier work ( gm09 ) , a reasonable assumption was made that @xmath44 , which can be approximately accepted for a geometrically thick , 2.5-dimensional disk structure . presuming that the outflow velocity is not likely to exceed the sound speed at the disk - outflow surface ( the outer boundary of the accretion flow in the @xmath15 direction )
, we propose a simplified relation between @xmath3 and @xmath45 as @xmath46 hence , at @xmath47 , we obtain @xmath48 this implies that the outflow can ideally come out off the disk surface which can appropriately be termed as the sonic surface in the vertical direction . with this notion in mind
let us generalize this particular scaling between @xmath45 and @xmath3 , pertaining to our formalism as @xmath49 where @xmath50 and @xmath51 are the constants which will be determined by substituting eqn .
( [ 5c ] ) in our model conservation equations described in 2 .
the index @xmath51 measures the degree of subsonic nature of the vertical flow within the disk - outflow coupled region ( i.e. in between the mid - plane and the surface of the accretion flow ) . generalizing the procedure adopted in earlier works for 1.5-dimensional disks ( chakrabarti 1996 ; mukhopadhyay 2003 ; mukhopadhyay & ghosh 2003 ) , we solve eqns .
( [ 1b ] ) , ( [ 2b ] ) , ( [ 3b ] ) & ( [ 4c ] ) . using eqn .
( [ 5c ] ) along with eqn .
( [ 3b ] ) and combining eqns .
( [ 1b ] ) & ( [ 2b ] ) , we obtain @xmath52 equation ( [ 5d ] ) shows that to guarantee a smooth solution at the critical point " , @xmath53 . from the critical point condition we obtain @xmath54 , at @xmath55 . here
subscript @xmath9 is referred to critical point .
the radius @xmath56 is also called the `` sonic radius '' since no disturbance created within this radius can cross the radius ( also known as sound horizon ) and escape to infinity .
conditions at critical / sonic radius give @xmath57^{1/2 } , \label{5e}\end{aligned}\ ] ] where @xmath58 , is the sound speed or the radial velocity of the flow at the sonic radius in the disk mid - plane ( @xmath6 ) .
@xmath59 is the radial gravitational force at the sonic radius and @xmath60 is the corresponding value in the disk mid - plane . now at sonic location @xmath61 . hence by applying lhospital s rule , eqn .
( [ 5d ] ) reduces to @xmath62 , \label{5f}\end{aligned}\ ] ] where @xmath63 , @xmath64 and @xmath65 , are complicated functions of @xmath16 and @xmath15 at sonic location ; @xmath66 , @xmath67 & @xmath68 have been explicitly given in the appendix . equations ( [ 5d ] ) and ( [ 5f ] ) are then solved with an appropriate boundary condition to obtain @xmath0 and @xmath3 as functions of @xmath16 .
the value of specific angular momentum in our flow always remains constant and is same as @xmath69 , the value at sonic radius , by virtue of eqn .
( [ 3b ] ) . until now ,
we have emphasized on velocity profiles of the accretion - induced outflow at any arbitrary @xmath15 .
we have , however , mentioned before that the intrinsic coupling of inflow and outflow is confined within a specified region , from where the outflow emanates .
the disk - outflow inter - correlated region is bounded vertically from the mid - plane ( @xmath6 ) to an upper surface above which any inflow of matter ceased to exist .
we aim at precisely investigating the nature and dynamics of the flow at different layers within this disk - outflow coupled region .
therefore , we first need to calculate the scale - height of the accretion flow , based on our proposed model , and obtain the disk - outflow surface ( upper boundary ) .
it is to be noted that the transcendental eqn .
( [ 4c ] ) can not be independently used to calculate the scale - height @xmath35 or its general variance with @xmath16 , for unknown variables @xmath0 and @xmath3 .
nevertheless , we can easily obtain the scale - height @xmath35 at @xmath56 , say @xmath70 , from eqn .
( [ 4c ] ) as radial velocity and sound speed at the sonic location are known .
thus at @xmath56 , eqn .
( [ 4c ] ) simplifies to @xmath71\\ & & + f_{gzc}|_h -\frac{c_{s0c}^2}{\gamma h_c } = 0 , \label{6a}\end{aligned}\ ] ] where @xmath72 is @xmath73 at the sonic location in the plane with scale - height @xmath35
. then @xmath70 obtained from eqn .
( [ 6a ] ) can be inserted into eqn .
( [ 5f ] ) to obtain @xmath74 at scale - height @xmath35 .
the disk scale - height @xmath35 without any outflow is seen to be linearly increasing with @xmath16 ( approximately by dimensional analysis ) . for a 2.5-dimensional disk - outflow system
, it seems that @xmath35 and @xmath16 can also be linearly connected , based on the order of magnitude analysis ( gm09 ) .
however , unlike the thin disk , strong gas pressure gradient in the disk - outflow coupled system leads to a geometrically thick 2.5-dimensional disk .
therefore , for the present purpose we make a generalized scaling of @xmath35 with @xmath16 as @xmath75 where @xmath76 be a dimensionless arbitrary constant or any numerical variable ( discussed in detail later ) .
the impressionable choice about the factor @xmath76 would be that it should contain precisely the information of the nature of the flow .
the only physical information we can extract related to the scale - height from our model equations is @xmath70 .
thus we rationally demand an approximate expression for dimensionless parameter @xmath76 as @xmath77 it is found that the value of @xmath76 circumvents around @xmath78 , for the entire range of spin parameter of the black hole , which here acts as a normalization constant .
thus eventually , eqns .
( [ 5d ] ) , ( [ 5f ] ) , ( [ 6b ] ) & ( [ 6c ] ) are combined together to obtain @xmath0 and @xmath3 along the scale - height @xmath35 for an accretion - induced outflow .
it also appears that under all these circumstances , and for a physically acceptable flow , the constant @xmath79 and the index @xmath51 of eqn .
( [ 5c ] ) connecting @xmath45 and @xmath3 yield to be @xmath80 and @xmath81 respectively . after establishing the scale - height @xmath35
, we in principle can obtain the profiles of the dynamical parameters at different layers of the disk , i.e. at @xmath82 , where @xmath83 , within our prescribed disk - outflow region .
we find that the radial velocity profile @xmath0 along all layers of disk for any arbitrary spin parameter exhibits some unusual , yet very interesting behavior . for any specific @xmath15 and spin parameter @xmath84
, @xmath0 attains a negative value at a radius greater than a certain distance @xmath85 .
the magnitude of @xmath86 decreases with the increase of @xmath87 , which indicates that the positive trait of @xmath0 is greater for lower @xmath15 .
the negative value of @xmath0 does not necessarily reflect any unphysical behavior .
if the outflow originates from any particular layer @xmath88 at a radius @xmath89 @xmath86 , it then appears that the coupling between the disk and the outflow ceased to exist .
it infers that along the layer @xmath88 , both accretion and outflow simultaneously occur up to the radial distance @xmath86 , and beyond which the solution ( with negative @xmath0 ) reflects the truncation of the disk ( along the specified layer ) .
let us consider different layers of @xmath15 . as we ascend to layers from a lower to a higher @xmath15
, the truncation of the disk - outflow region along these layers occurs at a radius smaller than that of the lower @xmath15 .
as accretion is the source of the outflow , we can ostensibly conclude that the region from where the outflowing matter originates in the disk shrinks as we go to higher latitudes .
note that @xmath86 corresponds to zero @xmath0 .
therefore , we restrict our study up to the boundary @xmath86 . with this insight , we simply compute @xmath86 along various layers of the disk and construct a surface in the @xmath90 plane connecting all @xmath86 for different layers .
the enclosed region bounded by this surface is defined by the positivity of @xmath0 , where both the accretion and outflow simultaneously persist and are intrinsically coupled to each other .
we attribute the outer surface of the region as disk - outflow surface ( @xmath91 ) and above this layer no inflow takes place .
the arrows in the diagram shown in fig . 1
reveal the direction of flow .
the arrow exactly at @xmath86 points vertically upwards , which indicates that just at @xmath86 the flow pattern exclusively corresponds to vertical motion and any accretion flow ceased to exist .
however , more realistic disk - outflow surface could be visualized with a thick - solid line shown in fig
. 1 . as in our system
we have neglected the viscosity and any radiative loss , and thus the flow is considered to be strongly advective , presumably it is gas pressure dominated .
notwithstanding , various scattering processes will indeed produce radiation in the system , whose contribution could be insignificant .
hence , @xmath92 is reasonably an appropriate choice in our model . and @xmath93 , the ratio of gas pressure to the total pressure , is given by @xmath94 ( gm09 ) .
] note that this is in essence similar to the choice of the earlier authors who modeled gas dominated low mass advection dominated accretion flows ( e.g. narayan & yi 1994 ) .
we would also like to clarify that @xmath95 corresponds only to the case of zero angular momentum .
figure 2 shows the profiles of disk - outflow surface for various spin parameters @xmath84 of the black hole .
all the input parameters corresponding to each value of @xmath84 are listed in table 1 .
it is seen that the disk - outflow region and the peak of the surface namely , @xmath96 , shifts to the vicinity of the black hole for a higher spin .
table 1 input parameters for various @xmath84 [ cols="^,^,^",options="header " , ] respectively .
] the marginally stable orbit in a keplerian accretion flow defines the inner edge of the disk ( with zero torque condition ) .
however the orbit is more or less an artifact of the flow . for a transonic advective disk ( as is the present case ) ,
an apriori definition of inner edge is somewhat fuzzy , due to which we can pretend to choose that the disk extends to the black hole horizon . in presence of an outflow
intrinsically coupled to the disk , there is supposed to be an inner boundary beyond which any outflow and then jet will ceased to exist .
we define this inner boundary of the disk - outflow coupled region in explaining the energetics of the flow .
also , with the increase of the spin of the black hole , it is seen that the disk - outflow coupled region shrinks considerably , attaining a steeper nature . literally speaking
, the spin of the black hole directly influences the nature and region of the outflow .
the tendency of the outflow region to get contracted to the inner radius suggests that the outflow is more likely to exuberantly emanate from inner region of the disk for rapidly spinning black holes which have more gravitating power .
therefore , with the increase in spin , the disk - outflow region becomes more dense and more susceptible to eject the matter with a greater power .
as the system is gas pressure dominated and strongly advective , it is more receptive to strong outflows . figures 3 and 4 describe the variation of flow parameters as functions of radial coordinate @xmath16 along the disk - outflow surface for various spin of the black hole . with the increase of the spin of the black hole ,
@xmath0 increases at the inner region of the disk .
it is seen that along the disk - outflow surface , @xmath0 becomes zero beyond a certain distance @xmath96 , which is also termed as zero @xmath0 surface illustrated in 3.1 , which is explained it detail in the next section in order to understand the outflow power .
the zero @xmath0 surface extends upto more inner region of the disk with the increase of spin of the black hole , indicating the fact that outflows and then jets originate from more inner region of the disk for rapidly spinning black holes .
the increase of @xmath3 with spin ( fig .
4 ) indicates that the temperature of the disk - induced outflow is higher for rapidly spinning black holes .
it is found from fig .
4 that the maximum temperature of the accretion - induced outflow varies from @xmath98 to @xmath99 corresponding to zero to maximal spin of the black hole .
the truncation of the curves in the inner region symbolizes the inner boundary of the disk - outflow region as mentioned in 3.1 .
respectively . other parameter @xmath97 . ]
accretion by a black hole is the primary source of energy of the mass outflow from the inner disk region .
the energetics of the accretion - induced outflow are mainly attributed to the mass outflow rate and the power extracted by the outflow from the disk .
the derivation of the mass outflow rate had been elaborated in g10 .
we follow the same procedure here , however including the spin information of the black hole .
the mass outflow rate is given by @xmath100 where the constant @xmath101 is determined by an appropriate boundary condition . from the adiabatic conditions
, @xmath4 can be written in terms of @xmath3 which is already determined as a function of @xmath16 .
the corresponding proportionality constant is determined at a radius , outside which the contribution to the mass outflow rate is negligible ( @xmath7 ) .
however , the total mass accretion rate @xmath102 ( which is sum of the inflow rate and the outflow rate ) can be obtained by integrating the continuity equation along the radial and vertical directions .
hence , the constant is computed easily by supplying the values of @xmath0 and @xmath3 at that radius in the expression of @xmath102 . as we have discarded any possible small amount of the outflow at that radius ,
the actual value of the constant , and hence the outflow power , could be slightly over estimated .
@xmath103 in eqn .
( [ 5a ] ) refers to the rate at which the vertical mass flux ejects from the disk - outflow surface .
figure 5 shows the variation of @xmath104 profiles with the spin of the black hole .
it is seen that with the increase of spin of the black hole , @xmath104 increases .
all the profiles have been shown considering a supermassive black hole of mass @xmath105 with a mass accretion rate at infinity , @xmath106 , where @xmath107 is the eddington mass accretion rate @xmath108 .
low mass accretion rate ( sub eddington accretion flow ) is in conformity with our gas pressure dominated advective disk paradigm .
it is found that @xmath104 increases with the increase of mass of the black hole and mass accretion rate as well .
the truncation of the curves at an inner radius indicates the inner boundary of the disk - outflow coupled region , explained in detail in the next paragraph .
= 0 , ( b ) @xmath84 = 0.5 , ( c ) @xmath84 = 0.9 , ( d ) @xmath84 = 0.998 . other parameter @xmath97 . ] in computing the power extracted by the outflow from the disk , we follow the same procedure as in g10 .
thus the power of the outflow is given by @xmath109 \bigg |_{h_{surf } } \ , dr , \\
\label{5b}\end{aligned}\ ] ] which is the total power removed from the disk by the outflow along the disk - outflow surface . in measuring the net power of any astrophysical jet
, it appears that the computed power @xmath110 will then be the initial power of the jet .
figure 6 depicts the variation of the power @xmath110 with @xmath16 for various spin parameters of the black hole .
if we meticulously investigate the nature of the power profiles , we observe that when the radial distance is less than a certain value , @xmath110 begins to decrease .
the decreasing trend of @xmath110 infers that the characteristic flow is bounded ( integrand of eqn .
( [ 5b ] ) carrying the information of the vertical energy flux becomes negative ) .
= 0 , ( b ) @xmath84 = 0.5 , ( c ) @xmath84 = 0.9 , ( d ) @xmath84 = 0.998 . other parameter @xmath97 . ]
this occurs owing to the fact that in the extreme inner region of disk , due to its strong gravitating power the starved black hole sucks all of the matter in its sphere of influence , even if there is any outflow emanating from the disk .
we identify this inner transition radius as @xmath111 , beyond which no outflow occurs .
we attribute @xmath111 as the inner boundary of the disk - outflow region . in all of the previous profiles , the said truncation of the curves
is ascribed to @xmath111 .
we describe the dynamical variables in our study within the region between an outer boundary and @xmath111 , and within this prescribed region strong outflows are most plausible to originate , intrinsically coupled to the disk .
the power profile shows that with the increase in @xmath84 , @xmath110 increases as well as @xmath111 gradually shifts to the vicinity of the black hole , indicating the fact that outflow region moves further inward .
we show the variation of total @xmath110 extracted from the disk - outflow region with the spin of the black hole in fig .
considering a black hole of mass @xmath105 , as seen in agns and quasars , accreting with @xmath106 , it is seen that @xmath112 @xmath113 for @xmath114 . for a maximally spinning black hole ( @xmath115 ) with same parameters , the computed @xmath116 @xmath113 .
thus , there is an increase of two orders of magnitude of @xmath110 with the increase of spin of the black hole from @xmath117 to @xmath118 . in an earlier work ,
donea & biermann ( 1996 ) showed that the power extracted by the outflow / jet from the disk increases with @xmath84 .
however , they did not compute the power extracted from the disk explicitly . the numerical simulations by de villiers et al .
( 2005 ) , in a different accretion paradigm including magnetic field , also concluded that the jet efficiency is possible to increase with the increase of spin of the black hole from @xmath117 to @xmath118 .
figure 7b shows the variation of @xmath111 with @xmath84 . in describing the disk - outflow surface in 3.1 ,
we have articulated the impact of spin on the disk - outflow coupled region .
the geometry of the surface for a particular @xmath84 is identified with a parameter @xmath96 ( see fig .
we show exclusively the variation of @xmath96 with the spin of the black hole in fig .
7c . which too reveals that the fast rotating black hole retracts the outflow region towards it ; a sign of a pure relativistic gravitation .
erg s@xmath119 , ( b ) inner transition radius of the outflow , ( c ) peak of disk - outflow surface , as functions of spin of the black hole . ]
blandford - znajek process ( blandford & znajek 1977 ) is still one of the most promising mechanisms to drive powerful jets in agns and xrbs .
although the exact mechanism of formation of the jet in the vicinity of the black holes is still elusive , and whatever might be the reason for the origin of jet , the said work is significant mostly due to two underlying reasons : ( 1 ) extreme gravity is indispensable to effuse strong unbounded flows in the vertical direction from the inner region of the accretion disk , ( 2 ) the spin of the black holes , which is purely a relativistic effect , is essential to power strong outflows and jets . in the present study , we have neither laid importance to the nature of the outflow nor invoked any aspect for the origin of outflows or jets .
indeed , the distinctive or definitive understanding of the formation of strong outflows or jets is till unknown .
notwithstanding , the most distinguishable and obvious picture in this case is that outflows and jets observed in agns and xrbs can only originate in an accretion powered system , where the accretion of matter around relativistic gravitating objects like black holes acts as a source , and outflow and then jet takes the form of one of the possible sinks ( the other sink is the central nucleus ) .
the dynamics of the outflowing matter should then be intrinsically coupled to the accretion dynamics macroscopically through the fundamental laws of conservation ( of matter , momentum and energy ) .
the outflow is unbounded and the total energy just at the base of the outflow should be positive .
the accretion flow should be bounded in the vicinity of the central object as the central potential is attractive . however , the strength of the unbounded flows in the form of jets is distinctly proportional to the attractiveness of the central gravitational potential field .
this paradox is well manifested in the observed universe , as relativistic jets are more populated around extreme gravitating objects like black holes . also noticeably , length scale of jets increases from microquasars to quasars , which are supposedly harboring stellar mass and supermassive black holes respectively .
thus in any theoretical modeling of the accretion and outflow , it can be presumably argued that the mathematical equations governing the dynamics of the inflow and outflow should inherently be correlated and evolved self - consistently without any ad hoc proposition , as accretion and outflow should not be treated as dissimilar objects .
second , the relativistic gravitational effect of the black hole should be incorporated , as the nature of gravity is the cornerstone to both the accretion and the unbounded outflow . to capture this essential physics , in our present study to understand the connection between disk and outflow , we have incorporated the general relativistic effect of the spinning black hole through a pseudo - newtonian approach . although pseudo - newtonian formalism is an approximate method to mimic the space - time geometry of the kerr black hole , yet it captures the important salient features of the corresponding metric , and thus can be used to examine the nature of outflows from the inner region of the disk .
in 2 we have described the general disk - outflow coupled hydrodynamic equations in the inviscid limit following gm09 .
the necessity to simplify our model equations for an inviscid flow is discussed in 2 .
although gm09 explored the 2.5-dimensional accretion - induced outflow for a fully viscous system , they used a self - similar approach in order to solve the necessary partial coupled differential equations .
in addition , they neglected the most indispensable effect of relativistic gravitation or , precisely , the effect of spin of the black hole . in the present paper , we have solved the disk - outflow model equations in a more general 2.5-dimensional paradigm , while trying to limit our assumptions to the least possible extent theoretically .
one of the important premises we have made is the relationship between @xmath45 and @xmath3 , which we have established empirically .
we have not used the height integrated equations which are mostly valid in the circumstances where the dynamical fluid parameters are likely to be independent of @xmath15 . without presuming the fact that the outflow originates from the surface of the disk , as most of the authors do , we have logically constructed a disk - outflow surface with properly defined boundary conditions .
one of the most important computations we have done is to evaluate the mass outflow rate and the power of the outflow extracted from the disk self - consistently in the inviscid limit , unlike the previous works ( e.g. donea & biermann 1996 ; blandford & begelman 1999 ; xie & yuan 2008 ) .
we have found that the spin of the black hole plays a crucial role in determining the structure , dynamics and the energetics of the outflow coupled to the disk . with the increase of the spin
, the outflow region extends further inward and then the disk - outflow region shrinks and compresses ( see fig .
2 ) . as a result , the outflow and
then any plausible jet is likely to eject out from an extreme inner region of the flow around a rapidly spinning black hole with a greater efficiency .
note that the higher spin results in the system to get more compressed with a greater outflow power .
therefore , the efficiency of outflow and jet is directly related to the disk scale - height and hence the disk - outflow surface . previously , in a different context , mckinney & gammie ( 2004 ) , while examining the electromagnetic luminosity of a kerr black hole , assumed the ratio of the disk height to the radius ( @xmath120 ) constant , irrespective of the black hole spin . however , as seen in the present work , the constant @xmath120 for different spin of the black hole may not be an obvious choice .
the power extracted by the outflow from the disk not only depends directly on the mass of the black hole and the initial mass accretion rate of the flow , but also on the spin of the black hole . with our model ,
keeping the black hole mass and the accretion rate the same , the power of the outflow increases with the spin of the black hole and the computed power differs in two orders of magnitude between non - rotating and maximally rotating black holes .
we have restricted our study vertically up to the region where the inflow and the outflow are least coupled , i.e. the disk - outflow surface . above this surface accretion ceased to exist and probably the outflow gets decoupled from the disk , accelerates and eventually forms relativistic jet .
the modeling of the astrophysical jet is altogether a different issue and is beyond the scope of the present work .
nevertheless , it can be effectively argued that in modeling the dynamics of the jet , the computed outflow power may serve as an initial power fed to the jet .
thus , the dynamics and the energetics of the jet will eventually be related to the black hole spin . according to the unification scenario , it is possible to device a single astrophysical scenario , which can broadly explain the observed multitude of different types of agns ( see e.g. antonucci 1993 ; urry & padovani 1995 ) .
flat spectrum radio quasars ( fsrqs ) and bl lacs are probably the two most active types of agns which are collectively referred to blazar .
it is observed that bl lacs are relatively low luminous than fsrqs .
although both of these galaxies show high energy emissions , their spectral properties are different ( bhattacharya et al . 2009 ) which indicate that they are different source classes .
according to the unification scheme , for fsrqs the line - of - sight is almost parallel to the jet and , hence , strong relativistic doppler beaming of the jet emission produces highly variable and continuum dominated emission .
as one moves away from the jet axis , the central continuum emission falls and the nucleus looks like a fr - ii galaxy .
similarly , bl lacs are considered to be a sub class of fr - i galaxies whose line - of - sight is almost parallel to the jet axis .
the present work suggests that the total mechanical power of an outflow proportionately increases with the spin of the central supermassive black hole ; higher the spin , stronger is the outflow .
it is reasonable to consider that strong outflow can lead to a strong jet , and hence , one can expect to observe higher luminosity .
therefore , the work suggests that bl lacs are slow rotators than fsrqs . for the theoretical formulation of the disk - outflow coupling
even with approximations , one needs to be very thoughtful for the proper foundation of the model which essentially needs to solve hydrodynamic or magnetohydrodynamic conservation equations in presence of strong gravity .
the flow parameters vary in 2.5-dimension and are coupled to each other .
limited observational inputs put irremediable constraint on the boundary conditions as well as the fundamental scaling parameters , governing the coupled dynamics of the accretion and outflow . the inadequacy of an effective mathematical tool to handle partial coupled differential hydrodynamic equations for compressible flow motivates us to invoke approximations and assumptions . despite of this fact
, one needs to explore the possibilities to examine the accretion - induced outflow .
the question then arises : what are the valid assumptions and to which extent they can be considered ? in the present work , we have addressed this question in the following way . + ( 1 ) as the extreme gravity is the most important aspect to effuse jet from the disk , we have incorporated its effect through a pseudo - general - relativistic potential .
+ ( 2 ) the disk and outflow should not be treated as dissimilar objects , and hence their correlated - dynamics should be essentially governed by the conservation laws .
the energetics of the accretion - induced outflow would then be evaluated self - consistently as is shown in our work .
+ ( 3 ) any unbounded flow in the form of outflow is more plausible to emanate from a hot , puffed up region of the accretion flow ( may be low / hard state of the black hole ) .
hence we have formulated our model in a 2.5-dimensional , strongly advective paradigm , surpassing the simplicity of height integration .
+ ( 4 ) we have approximated our system to an inviscid limit , whose reason has been argued in 2 .
nevertheless , in any future work , viscosity should be incorporated into the flow to make it more realistic . +
( 5 ) we do not account for the mechanism of formation of strong jets , like due to magnetic field or strong radiation pressure , as the definitive mechanism of the formation of jet is still unknown .
indeed , it is beyond the scope to solve magnetohydrodynamic equations in the present scenario .
however , assuming that the outflow resides , and is coupled to the disk , the governing conservation equations of matter , momentum and energy should be treated accordingly .
+ ( 6 ) power of the outflow and jet is expected to directly depend on the spin of the black hole .
the spin , which is the signature of general relativity , can make an impact on the nature of the observed agn classes , and thus the spin of the black has been incorporated in our study .
as spin of the black hole has a direct impact on the flow parameters , we obtain different outflow power for different spin . what can be the most defined parameter for the accretion powered system mass accretion rate , mass of the central star / black hole or the spin of the central star / black hole ?
our analysis has shown that the power extracted by the outflow from the disk proportionately increases with the spin of the black hole .
it infers that if extreme gravity is essential to power the jet , then the strength and the length - scale of the observed astrophysical jets directly depend on the spin of the black hole .
if it is believed that the distant quasars harbor massive black holes of same mass scale and accrete matter with similar rate , perhaps , spin be the guiding parameter for the different observed agn classes .
we can end with the specific question : are the observed agn classes astrophysical laboratories to measure the spin of the supermassive black holes ? + this work is supported by a project , grant no .
sr / s2hep12/2007 , funded by dst , india .
the authors would like to thank the referee for making important comments which helped to prepare the final version of the paper .
+ * appendix * equation ( [ 5f ] ) consists of complicated terms , where @xmath121 and @xmath122 are given by following equations @xmath123 @xmath124 @xmath125 and @xmath126 @xmath127 + \frac{1}{2n } \biggl(\frac{\partial}{\partial r } f_{gr } \bigg |_{c } + 3 \frac{\lambda_{c}^{2}}{r_{c}^4}\biggr ) + \frac{v^{2}_{rc}}{2 n r_c^2 } + \mu \frac{\imath}{2
n z } \ , \frac{z^{\mu}}{r^{\mu + 1}_c } \ , v_{r0c } v_{rc } \,\ ] ] @xmath128 , \eqno(a3 ) \ ] ] where , + @xmath129 + @xmath130 + and + @xmath131 | the mechanism by which outflows and plausible jets are driven from black hole systems , still remains observationally elusive .
notwithstanding , several observational evidences and deeper theoretical insights reveal that accretion and outflow / jet are strongly correlated .
we model an advective disk - outflow coupled dynamics , incorporating explicitly the vertical flux .
inter - connecting dynamics of outflow and accretion essentially upholds the conservation laws .
we investigate the properties of the disk - outflow surface and its strong dependence on the rotation parameter of the black hole .
the energetics of the disk - outflow strongly depend on the mass , accretion rate and spin of the black holes .
the model clearly shows that the outflow power extracted from the disk increases strongly with the spin of the black hole , inferring that the power of the observed astrophysical jets has a proportional correspondence with the spin of the central object . in case of blazars ( bl lacs and flat spectrum radio quasars ) ,
most of their emission are believed to be originated from their jets .
it is observed that bl lacs are relatively low luminous than flat spectrum radio quasars ( fsrqs ) .
the luminosity might be linked to the power of the jet , which in turn reflects that the nuclear regions of the bl lac objects have a relatively low spinning black hole compared to that in the case of fsrq .
if the extreme gravity is the source to power strong outflows and jets , then the spin of the black hole , perhaps , might be the fundamental parameter to account for the observed astrophysical processes in an accretion powered system . |
optical interferometry provides physicists and astronomers with an exquisite set of probes of the micro and macrocosmos . from the laboratory to the observatory over the past few decades , there has been a surge of activity in developing new tools for ground - based optical astronomy , of which interferometry is one of the most powerful .
an optical interferometer is a device that combines two or more light waves emitted from the same source at the same time to produce interference fringes .
the implementation of interferometry in optical astronomy began more than a century ago with the work of fizeau ( 1868 ) .
michelson and pease ( 1921 ) measured successfully the angular diameter of ( @xmath0 ori ) , using an interferometer based on two flat mirrors , which allowed them to measure the fringe visibility in the interference pattern formed by starlight at the detector plane .
however , progress was hindered by the severe image degradation produced by atmospheric turbulence in the optical spectrum .
the field remained dormant until the development of intensity interferometry by hanbury brown and twiss ( 1958 ) , a technique that employs two adjacent sets of mirrors and photoelectric correlation .
turbulence and the concomitant development of thermal convection in the atmosphere distort the phase and amplitude of an incoming wavefront of starlight .
the longer the path , the greater the degradation that the image suffers .
light reaching the entrance pupil of an imaging system is coherent only within patches of diameters of order r@xmath1 , fried s parameter ( fried , 1966 ) .
this limited coherence causes blurring of the image , blurring that is modeled by a convolution with the point - spread function ( psf ) . both the sharpness of astronomical images and the signal - to - noise ( s / n ) ratio ( hence faintness of the objects that can be studied ) depend on angular resolution , the latter because noise comes from as much of the sky as is in the resolution element .
thus reducing the beam width from , say , 1 arcsecond ( @xmath2 ) to 0.5@xmath2 reduces sky noise by a factor of 4 .
two physical phenomena limit the minimum resolvable angle at optical and infrared ( ir ) wavelengths - diameter of the collecting area and turbulence in the atmosphere .
the crossover between domination by aperture size ( @xmath3/aperture diameter ) and domination by atmospheric turbulence ( ` seeing ' ) occurs when the aperture becomes larger than the size of a characteristic turbulent element .
the image of a star obtained through a large telescope looks ` speckled ' or grainy because different parts of the image are blurred by small areas of turbulence in the earth s atmosphere .
labeyrie ( 1970 ) proposed speckle interferometry ( si ) , a process that deciphers the diffraction - limited fourier spectrum and image features of stellar objects by taking a large number of very - short - exposure images of the same field .
computer assistance is then used to reconstruct from these many images a single image that is free of turbulent areas - in essence , an image of the object as it might appear from space .
the success of speckle interferometry in measuring the diameters of stars encouraged astronomers to develop further image - processing techniques .
these techniques are , for the most part , post - detection processes .
recent advances in technology have produced the hardware to compensate for wave - front distortion in real time .
adaptive optics ( ao ; babcock , 1953 ; rousset et al . ,
1990 ) is based on this hardware - oriented approach , which sharpens telescope images blurred by the earth s atmosphere .
it employs a combination of deformation of reflecting surfaces ( i.e. , flexible mirrors ) and post - detection image restoration ( roddier , 1999 ) .
one of its most successful applications has been in imaging of neptune s ring arcs .
adaptive optical imaging systems have been treated in depth by the review of roggemann et al .
( 1997 ) , which includes discussion of wavefront compensation devices , wavefront sensors , control systems , performance of ao systems , and representative experimental results .
it also deals with the characterization of atmospheric turbulence , the si technique , and deconvolution techniques for wavefront sensing .
long - baseline optical interferometry ( lboi ) uses two or more optical telescopes to synthesize the angular resolution of a much larger aperture ( aperture synthesis ) than would be possible with a single mirror .
labeyrie ( 1975 ) extended the concept of speckle interferometry to a pair of telescopes in a north - south baseline configuration , and subsequently astronomers have created larger ground - based arrays .
a few of these arrays , e.g. , the keck interferometer and the very large telescope interferometer ( vlti ) , employ large telescopes fitted with ao systems .
the combination of long - baseline interferometry , mimicking a wide aperture , and ao techniques to improve the images offers the best of both approaches and shows great promise for applications such as the search for extra - solar planets . at this point
it seems clear that interferometry and ao are complementary , and neither can reach its full potential without the other .
the present review addresses the aims , methods , scientific achievements , and future prospects of long - baseline interferometry ( lbi ) at optical and infrared wavelengths , carried out with two or more apertures separated by more than their own sizes . in order to embark on such a subject ,
we first review the basic principles of interferometry and its applications , the theoretical aspects of si as a statistical analysis of a speckle pattern , and the limitations imposed by the atmosphere and the detectors on the performance of single - aperture diffraction - limited imaging . other related concerns , such as the relationship between image - plane techniques and pupil - plane interferometry , phase - closure methods , and dark speckle imaging , will also be treated .
adaptive optics as a pre - detection compensation technique is described in brief , as are the strengths and weaknesses of pre and post detection . the final part of this review deals with the applications of multiple - telescope interferometry to imaging , astrometry , and nulling .
these applications entail specific problems having to do with delaylines , beam recombination , polarization , dispersion , fringe tracking , and the recovery of visibility functions .
various image restoration techniques are discussed as well , with emphasis on the deconvolution methods used in aperture - synthesis mapping .
astronomical sources emit incoherent light consisting of the random superposition of numerous successive short - lived waves sent out from many elementary emitters , and therefore , the optical coherence is related to the various forms of correlations of the random process . for a monochromatic wave field ,
the amplitude of vibration at any point is constant and the phase varies linearly with time .
conversely , the amplitude and phase in the case of quasi - monochromatic wave field , undergo irregular fluctuations ( born and wolf , 1984 ) .
the rapidity of fluctuations depends on the light crossing time of the emitting region .
interferometers based on ( i ) wavefront division ( young s experiment that is sensitive to the size and bandwidth of the source ) , ( ii ) amplitude division ( michelson s interferometer ) are generally used to measure spatial coherence and temporal coherence , respectively . in what follows ,
some of the fundamental mathematical steps pertinent to the interferometry are illustrated .
a monochromatic plane wave , @xmath4 at a point , @xmath5 , is expressed as , @xmath6}\right\}.\ ] ] here the symbol @xmath7 is the ` real part of ' , @xmath5 the position vector of a point @xmath8 , @xmath9 is the amplitude of the wave , @xmath10 the time , @xmath11 the frequency of the wave , @xmath12 the phase functions that are of the form @xmath13 , in which @xmath14 is the propagation vector , and @xmath15 the phase constants which specify the state of polarization , denoting the complex vector function of position by @xmath16 , the complex representation of the analytic signal , @xmath17 , associated with @xmath4 becomes , @xmath18}\\ & & = \psi({\bf r})e^{-i2\pi\nu_\circ t}. \end{aligned}\ ] ] this complex representation is preferred for linear time invariant systems , because the eigenfunctions of such systems are of the form @xmath19 , where @xmath20 is the angular frequency . from equations ( 1 ) and ( 2 )
, the relationship translates into , @xmath21 , \end{aligned}\ ] ] the intensity of light is defined as the time average of the amount of energy , therefore , taking the latter over an interval much greater than the period , @xmath22 , the intensity @xmath23 at the same point is calculated as , @xmath24 where @xmath25 stands for the ensemble average of the quantity within the bracket and @xmath26 represents for the complex conjugate of @xmath27 .
since the complex amplitude is a constant phasor in the monochromatic case , the fourier transform ( ft ) of the complex representation of the signal , @xmath17 , is given by , @xmath28 it is equal to twice the positive part of the instantaneous spectrum , @xmath29 . in the polychromatic case , the complex amplitude becomes , @xmath30 the disturbance produced by a real physical source is calculated by the integration of the monochromatic signals over an optical band pass . in the case of quasi - monochromatic approximation (
if the width of the spectrum , @xmath31 ) , the expression modifies as , @xmath32},\ ] ] where the field is characterized by the complex amplitude , @xmath33 , _
i.e. _ , @xmath34 this phasor is time dependent , although it varies slowly with respect to @xmath35 . the complex amplitude , @xmath36 diffracted at angle @xmath37 in the telescope focal - plane is given by , @xmath38 where @xmath39 is a two - dimensional ( 2-d ) space vector , @xmath40 the focal length , @xmath41 the complex amplitude at the telescope aperture , and @xmath42 the pupil transmission function of the telescope aperture .
for an ideal telescope , we have @xmath43 inside the aperture and @xmath44 outside the aperture . in the space - invariant case , @xmath45,\end{aligned}\ ] ] where @xmath46 represents for the complex ft and the dimensionless variable @xmath47 is equal to @xmath48 and hence , @xmath49 can be replaced by @xmath50 and so with @xmath51 by @xmath52 .
the irradiance diffracted in the direction @xmath37 is the psf of the telescope and the atmosphere and its ft , @xmath53 is the optical transfer function ( otf ) : @xmath54 } d{\pmb \alpha},\ ] ] where @xmath55 is the spatial frequency expressed in radian@xmath56 , and @xmath57 the modulation transfer function ( mtf ) .
the convolution of two functions simulates phenomena such as a blurring of a photograph that may be caused by various reasons , e.g. , ( a ) poor focus , ( b ) motion of a photographer during the exposure . in such a blurred picture each point of object
is replaced by a spread function .
for the 2-d incoherent source , the complex amplitude in the image - plane is the convolution of complex amplitude of the pupil plane and the pupil transmission function .
@xmath58 in the fourier plane , the effect of the convolution becomes a multiplication , point by point of the otf of the pupil , @xmath51 , with the transform of the object @xmath49 .
i.e. , @xmath59 the illumination at the focal - plane of the telescope observed as a function of image - plane is , @xmath60|^2.\end{aligned}\ ] ] the autocorrelation of this function , @xmath61 , is expressed as , @xmath62 = \widehat{\cal s}({\bf f } ) \widehat{\cal s}^\ast ( { \bf f } ) = |\widehat{\cal s}{\bf ( f)}|^2,\ ] ] where @xmath63 stands for correlation . in an ideal condition ,
the resolution that can be achieved in an imaging experiment , @xmath64 , is limited only by the imperfections in the optical system and according to strehl s criterion , the resolving power , @xmath64 , of any telescope of diameter @xmath65 is given by the integral of its transfer function , @xmath66 therefore , @xmath67 . the strehl ratio @xmath68 is defined as the ratio of the intensity at the centroid of the observed psf , @xmath69 , to the intensity of the peak of the diffraction - limited image or ` airy spike ' , @xmath70 , i.e. , @xmath71 where @xmath72 is the wave number and @xmath73 the rms optical path difference ( opd ) error .
typical ground - based observations with large telescopes in the visible wavelength range are made with a strehl ratio @xmath74 ( babcock , 1990 ) , while a diffraction - limited telescope would , by definition , have a strehl ratio of 1 .
when two light beams from a single source are superposed , the intensity at the place of superposition varies from point to point between maxima , which exceed the sum of the intensities in the beams , and minima , which may be zero .
this sum or difference is known as interference ; the correlated fluctuation can be partially or completely coherent ( born and wolf , 1984 ) .
let the two monochromatic waves @xmath75 and @xmath76 be superposed at the recombination point .
the correlator sums the instantaneous amplitudes of the fields .
the total field at the output is , @xmath77 then if @xmath78 and @xmath79 are the complex amplitudes of the two waves , with the corresponding phases @xmath80 and @xmath81 , these two waves are propagating in @xmath82 direction and linearly polarized with electric field vector in @xmath83 direction .
( a general radiation field is generally described by four stokes parameters @xmath84 , @xmath85 , @xmath86 , and @xmath87 , that specify intensity , the degree of polarization , the plane of polarization and the ellipticity of the radiation at each point and in any given direction ) .
therefore , the total intensity , ( see equation 4 ) , at the same point can be determined as , @xmath88 where @xmath89 , and @xmath90 , are the intensities of the two terms , and @xmath91 , is the interference term that depends on the amplitude components , as well as on the phase - difference between the two waves , @xmath92 , ( @xmath93 , is the opd between the two waves from the common source to the intersecting point and @xmath94 is the wavelength in vacuum ) . in general , two light beams are not correlated but the correlation term , @xmath95 , takes on significant values for a short period of time and @xmath96 = 0 .
time variations of @xmath97 are statistical in nature ( mendel and wolf , 1995 ) .
hence , one seeks a statistical description of the field ( correlations ) as the field is due to a partially coherent source . depending upon the correlations between the phasor amplitudes at different object points
, one would expect a definite correlation between the two points of the field emitted by the object .
the maximum and minimum intensity occur , when @xmath98 and @xmath99 , respectively . if @xmath100 , the intensity varies between 4@xmath23 , and 0 . in the case of quasi - monochromatic wave , the analytical signal , @xmath101 , obtained at the observation point
is expressed as , @xmath102 where @xmath103 s are constants , @xmath104 s the positions of two pinholes in the wave field , j = 1 , 2 , @xmath105 s the distances of a meeting point of the two beams from the two pinholes , @xmath106 , the time needed to travel from the respective pinholes to the meeting point , and @xmath107 the velocity of light .
if the pinholes are small and the diffracted fields are considered to be uniform , the value @xmath103 of the constants , @xmath108 turns out to be , @xmath109 and noting , @xmath110 , and therefore , the intensity at the output is found to be , @xmath111.\ ] ] the van cittert - zernike theorem states that the modulus of the complex degree of coherence ( describes the correlation of vibrations at a fixed point and a variable point ) in a plane illuminated by a incoherent quasi - monochromatic source is equal to the modulus of the normalized spatial ft of its brightness distribution ( born and wolf , 1984 , mendel and wolf , 1995 ) .
the observed image is the ft of the mutual coherence function or the correlation function
. the complex degree of ( mutual ) coherence , @xmath112 , of the observed source is defined as , @xmath113 where @xmath114 and @xmath115 .
the function , @xmath116 , is measured at two points . at a point where both the points coincide
, the self coherence , @xmath117 , reduces to ordinary intensity .
when @xmath118 , @xmath119 .
the ensemble average can be replaced by a time average due to the assumed ergodicity ( a random process that is strictly stationary ) of the fields .
if both the fields are directed on a quadratic detector , it yields the desired cross - term ( time average due to the finite time response ) .
the measured intensity at the detector would be , @xmath120.\ ] ] in order to keep the time correlation close to unity , the delay , @xmath121 , must be limited to a small fraction of the temporal width or coherence time , @xmath122 ; here @xmath123 , is the spectral width .
the relative coherence of the two beams diminishes with the increase of path length difference , culminating in a drop in the visibility ( a dimensionless number between zero and one that indicates the extent to which a source is resolved on the baseline being used ) of the fringes . for @xmath124 ,
the function , @xmath112 , can be approximated to , @xmath125 .
the exponential term is nearly constant and @xmath126 , measures the spatial coherence .
let @xmath127 , be the argument of @xmath112 , then , @xmath128.\ ] ] the measured intensity at a distance @xmath83 from the origin ( point at zero opd ) on a screen at a distance , @xmath82 , from the apertures is @xmath129 where @xmath130 , is the opd corresponding to @xmath83 , and @xmath131 the distance between the two apertures .
the modulus of the fringe visibility is estimated as the ratio of high frequency to low frequency energy in the average spectral density ; the visibility of fringes , @xmath132 , is estimated as , @xmath133 fizeau ( 1868 ) suggested that installing a screen with two holes in front of a telescope would allow measurements of stellar diameters with diffraction - limited resolution . in this set up , the beams are diffracted by the sub - apertures and the telescope acts as both collector and correlator .
therefore , temporal coherence is automatically obtained due to the built - in zero opd . the spatial modulation frequency , as well as the required sampling of the image change with the separation of sub - apertures .
the maximum resolution in this case depends on the separation between the sub - apertures ; the maximum spacings that can be explored are limited by the physical diameter of the telescope .
the number of stellar sources for measuring diameters is also limited .
one of the first significant results was the measurement of the diameter of the satellites of jupiter with a fizeau interferometer on the 40-inch yerkes refractor by michelson ( 1891 ) . with the 100-inch telescope on mt .
wilson anderson ( 1920 ) determined the angular separation ( @xmath134 ) of spectroscopic binary star capella .
results from the classical michelson interferometer were used to formulate special relativity .
they are also being used in gravity - wave detection . gravitational radiation produced by coalescing binaries , or exploding stars , for example , changes the metric of spacetime .
this effect causes a differential change of the path length of the arms of the interferometer , thereby introducing a phase - shift . today
, there are several ground - based long baseline laser - interferometric detectors based on this principle under construction , and within the next several years these detectors should be in operation ( robertson , 2000 ) .
the proposed laser interferometer space antenna , consisting of three satellites in formation about 50 million kilometers ( km ) above the earth in a heliocentric orbit , may detect gravitational waves by measuring fluctuations in the distances between test masses carried by the satellites .
the essence of the michelson s stellar interferometer is to determine the covariance @xmath135 of the complex amplitudes @xmath136 , at two different points of the wavefronts .
this interferometer was equipped with four flat mirrors that fold the beams by installing a 7-meter ( m ) steel beam on top of the mt .
wilson 100-inch telescope .
michelson and pease ( 1921 ) resolved the supergiant @xmath0 ori and a few other stars . in this case , the spatial modulation frequency in the focal - plane is independent of the distance between the collectors . in the fizeau mode ,
the ratio of aperture diameter / separation is constant from light collection to recombination in the image - plane ( homothetic pupil ) . in the michelson mode ,
this ratio is not constant since the collimated beams have the same diameter from the output of the telescope to the recombination lens .
the distance between pupils is equal to the baseline at the collection mirrors ( the resolution is limited by the baseline ) and to a much smaller value just before the recombination lens .
the disadvantage of the michelson mode is a very narrow field of view compared to the fizeau mode .
unfortunately the project was abandoned due to various difficulties , including ( i ) effect of atmospheric turbulence , ( ii ) variations of refractive index above the small sub - aperture , ( iii ) inadequate separation of outer mirrors , and ( iv ) mechanical instability .
intensity interferometry considers the quantum theory of photon detection and correlation .
it computes the fluctuations of the intensities @xmath137 , at two different points of the wavefronts .
the fluctuations of the electrical signals from the two detectors are compared by a multiplier .
the current output of each photoelectric detector is proportional to the instantaneous intensity @xmath23 of the incident light , which is the squared modulus of the amplitude @xmath27 .
the fluctuation of the current output is proportional to @xmath138 this expression indicates that the covariance of the intensity fluctuations is the squared modulus of the covariance of the complex amplitude .
having succeeded in completing the intensity interferometer at radio wavelengths ( hanbury brown et al .
1952 ) , hanbury brown and twiss ( 1958 ) demonstrated its potential at optical wavelengths by measuring the angular diameter of sirius .
subsequent development with a pair of 6.5 m light collector on a circular railway track spanning 188 m , provided the measurements of 32 southern binary stars ( hanbury brown , 1974 ) with angular resolution limit of 0.5 milli - arcseconds ( mas ) . in this arrangement ,
starlight collected by two concave mirrors is focused on to two photoelectric cells and the correlation of fluctuations in the photocurrents is measured as a function of mirror separation .
the advantages of such a system over michelson s interferometer are that it does not require high mechanical stability and remains unaffected by seeing .
another noted advantage is that the alignment tolerances are extremely relaxed since the pathlengths need to be maintained to a fraction of @xmath139 , where @xmath140 is the electrical bandwidth of the post - detection electronics .
the significant effect comes from scintillation induced by the atmosphere .
the sensitivity of this interferometer was found to be very low ; it was limited by the narrow band - width filters that are used to increase the speckle life time .
the correlated fluctuations can be obtained if the detectors are spaced by less than a speckle width .
theoretical calculations ( roddier , 1988 ) show that the limiting visual magnitude ( mag ) , @xmath141 that can be observed with such a system is of the order of 2 ( the faintest stars visible in the naked eye are 6th magnitude . the magnitude scale is defined as @xmath142 = -2.5 log @xmath143 , where @xmath144 and @xmath145 are the apparent magnitudes of two objects of fluxes @xmath146 and @xmath147 , respectively ) .
the density inhomogeneities appear to be created and maintained by the parameters , viz . ,
thermal gradients , humidity fluctuations , and wind shears , which produce atmospheric turbulence and therefore refractive index inhomogeneities .
the gradients caused by these environmental parameters warp the wavefront incident on the telescope pupil .
the image quality is directly related to the statistics of the perturbations of the incoming wavefront .
the theory of seeing combines the theory of atmospheric turbulence with the theory of optical physics to predict the modifications to the diffraction - limited image that the refractive index gradients produce ( young , 1974 , roddier , 1981 , coulman , 1985 ) .
atmospheric turbulence has a significant effect on the propagation of radio - waves , sound - waves and on the flight of aircraft as well .
this section is devoted to the descriptions of the atmospheric turbulence theory , metrology of seeing and its impact on stellar images .
the random fluctuations in the atmospheric motions occur predominantly due to ( i ) the friction encountered by the air flow at the earth s surface and consequent formation of a wind - velocity profile with large vertical gradients , ( ii ) differential heating of different portions of the earth s surface and the concomitant development of thermal convection , ( iii ) processes associated with formation of clouds involving release of heat of condensation and crystallization , and subsequent changes in the nature of temperature and wind velocity fields , ( iv ) convergence and interaction of air - masses with various atmospheric fronts , and ( v ) obstruction of air - flows by mountain barriers that generate wave - like disturbances and rotor motions on their lee - side .
the atmosphere is difficult to study due to the high reynolds number ( @xmath148 ) , a dimensionless quantity , that characterizes the turbulence . when the average velocity , @xmath149 , of a viscous fluid of characteristic size , @xmath150 , is gradually increased , two distinct states of fluid motion are observed ( tatarski , 1967 , ishimaru , 1978 ) , viz .
, ( i ) laminar ( regular and smooth in space and time ) , at very low @xmath149 , and ( ii ) unstable and random fluid motion at @xmath149 greater than some critical value . the reynolds number , obtained by equating the inertial and viscous forces , is given by , @xmath151 where @xmath152 is a function of the flow geometry , @xmath153 , and kinematic viscosity of the fluid , @xmath154 .
when @xmath152 exceeds critical value in a pipe ( depending on its geometry ) , a transition of the flow from laminar to turbulent or chaotic occurs . between these two extreme conditions ,
the flow passes through a series of unstable states .
high @xmath152 turbulence is chaotic in both space and time and exhibits considerable spatial structure .
the velocity fluctuations occur on a wide range of space and time scales . according to the atmospheric turbulence model ( taylor , 1921 , kolmogorov , 1941b , 1941c )
, the energy enters the flow at low frequencies at scale length , @xmath155 and spatial frequency , @xmath156 , as a direct result of the non - linearity of the navier - stokes equation governing fluid motion .
the large - scale fluctuations , referred to as large eddies , have a size of the geometrically imposed outer scale length @xmath155 .
these eddies are not universal with respect to flow geometry ; they vary according to the local conditions .
conan et al . ( 2000 ) derived a mean value @xmath157 24 m for a von krmn spectrum from the data obtained at cerro paranal , chile .
the energy is transported to smaller and smaller loss - less eddies until at a small enough reynolds number , the kinetic energy of the flow is converted into heat by viscous dissipation resulting in a rapid drop in power spectral density , @xmath158 for @xmath159 , where @xmath160 is critical wave number .
these changes are characterized by the inner scale length , @xmath161 , and spatial frequency , @xmath162 , where @xmath161 varies from a few millimeter near the ground to a centimeter ( cm ) high in the atmosphere .
the small - scale fluctuations with sizes @xmath163 , known as the inertial subrange , where @xmath164 is the magnitude of @xmath5 , have universal statistics ( scale - invariant behavior ) independent of the flow geometry .
the value of inertial subrange would be different at various locations on the site .
the statistical distribution of the size and number of these eddies is characterized by @xmath158 , of @xmath165 , where @xmath165 is a randomly fluctuating term , and @xmath10 the time .
the dependence of the refractive index of air @xmath166 , upon pressure , @xmath167 ( millibar ) and temperature , @xmath168 ( kelvin ) , at optical wavelengths is given by @xmath169 ( ishimaru , 1978 ) .
the optically important property of the kolmogorov law is that the refractive index fluctuations are largest for the largest turbulent elements up to the outer scale of the turbulence . at sizes below the outer scale ,
the one - dimensional ( 1-d ) power spectrum of the refractive index fluctuations falls off with -(5/3 ) power of frequency and is independent of the direction along the fluctuations are measured , i.e. , the small - scale fluctuations are isotropic ( young , 1974 ) . the three - dimensional ( 3-d ) power spectrum ,
@xmath170 , for the wave number , @xmath159 , in the case of inertial subrange , can be equated as , @xmath171 where @xmath172 is known as the structure constant of the refractive index fluctuations .
this kolmogorov - obukhov model of turbulence , describing the power - law spectrum for the inertial intervals of wave numbers , is valid within the inertial subrange and is widely used for astronomical purposes ( tatarski 1993 ) .
the refractive index structure function , @xmath173 , is defined as , @xmath174 which expresses its variance at two points @xmath175 , and @xmath176 .
the structure functions are related to the covariance function , @xmath177 , through @xmath178,\ ] ] where @xmath179 and the covariance is the 3-d ft of the spectrum , @xmath180 ( roddier , 1981 ) .
the structure function in the inertial range ( homogeneous and isotropic random field ) , according to kolmogorov ( 1941a ) depends on the magnitude of @xmath5 , as well as on the values of the rate of production or dissipation of turbulent energy @xmath181 and the rate of production or dissipation of temperature inhomogeneities @xmath182 .
the refractive index @xmath183 is a function of @xmath184 of the temperature , @xmath168 and humidity , @xmath185 . and therefore , the expectation value of the variance of the fluctuations about the average of the refractive index is given by , @xmath186 it has been argued that in optical propagation , the last term is negligible , and that the second term is negligible for most astronomical observations
. it could be significant , however , in high humidity situation , e.g. , a marine boundary layer ( roddier , 1981 ) .
most treatments ignore the contribution from humidity and express the refractive index structure function ( tatarski , 1967 ) as , @xmath187 similarly , the velocity structure function , @xmath188 , and temperature structure function , @xmath189 , can also be derived ; the same form holds for the humidity structure function .
the two structure coefficients @xmath190 and @xmath191 are related by , @xmath192 , and assuming pressure equilibrium , @xmath193 ( roddier , 1981 ) .
several experiments confirm this two - thirds power law in the atmosphere ( wyngaard et al .
1971 , coulman , 1974 , hartley et al . 1981 ,
lopez , 1991 ) .
robbe et al .
( 1997 ) reported from the observations using a long baseline optical interferometer ( lboi ) , interfromtre deux tlescopes ( i2 t ; labeyrie , 1975 ) that most of the measured temporal spectra of the angle of arrival exhibit a behavior compatible with the said power law .
davis and tango ( 1996 ) have measured the value of atmospheric coherence time that varied between @xmath1941 and @xmath1947 ms with the sydney university stellar interferometer ( susi ) .
the value of @xmath172 ( in equation 33 ) depends on local conditions , as well as on the planetary boundary layer .
the significant scale lengths , in the case of the former , depend on the local objects which primarily introduces changes in the inertial subrange and temperature differentials .
the latter can be attributed to ( i ) surface boundary layer due to the ground convection , extending up to a few km height of the atmosphere , ( @xmath195 ) , ( ii ) the free convection layer associated with orographic disturbances , where the scale lengths are height dependent , ( @xmath196 ) , and ( iii ) in the tropopause and above , where the turbulence is due to the wind shear as the temperature gradient vanishes slowly . in real turbulent flows ,
turbulence is usually generated at solid boundaries .
near the boundaries , shear is the dominant source ( nelkin , 2000 ) , where scale lengths are roughly constant . in an experiment , conducted by cadot et al .
( 1997 ) , it was found that kolmogorov scaling is a good approximation for the energy dissipation , as well as for the torque due to viscous stress .
they measured the energy dissipation and the torque for circular couette flow with and without small vanes attached to the cylinders to break up the boundary layer .
the theory of the turbulent flow in the neighborhood of a flat surface applies to the atmospheric surface layer .
masciadri et al .
( 1999 ) have noticed that the value of @xmath172 increases about 11 km over mt .
paranal , chile .
the turbulence concentrates into a thin layer of 100@xmath197200 m thickness , where the value of @xmath172 increases by more than an order of magnitude over its background level .
the spatial correlational properties of the turbulence - induced field perturbations are evaluated by combining the basic turbulence theory with the stratification and phase screen approximations .
the variance of the ray can be translated into a variance of the phase fluctuations . for calculating the same , roddier ( 1981 ) used the correlation properties for propagation through a single ( thin ) turbulence layer and then extended the procedure to account for many such layers .
several investigators ( goodman , 1985 , troxel et al .
1994 ) have argued that individual layers can be treated as independent provided the separation of the layer centers is chosen large enough so that the fluctuations of the log amplitude and phase introduced by different layers are uncorrelated .
let a monochromatic plane wave of wavelength @xmath3 from a distant star at the zenith , propagates towards the ground - based observer ; the complex amplitude at co - ordinate , @xmath198 , is given by , @xmath199 the average value of the phase , @xmath200 for height h , and the unperturbed complex amplitude outside the atmosphere is normalized to unity @xmath201 $ ] .
when this wave is allowed to pass through a thin layer of turbulent air of thickness @xmath202 , which is considered to be large compared to the scale of turbulent eddies but small enough for the phase screen approximation ( diffraction effects is negligible over the distance , @xmath202 ) , the complex amplitude of the plane wavefront after passing through the layer is expressed as , @xmath203 here the phase - shift @xmath204 introduced by the refractive index fluctuations , @xmath205 inside the layer can be written as , @xmath206 in this case , the rest of the atmosphere is thought to be calm and homogeneous . at the layer output
, the coherence function of the complex amplitude , @xmath207 , leads to , @xmath208}>.\ ] ] the quantity @xmath204 can be considered to be the sum of a large number of independent variables , and therefore , has gaussian statistics .
this equation is similar to fourier transform of the probability density function at unit frequency ; therefore , @xmath209 the term @xmath210 is the 2-d structure function of the phase , @xmath204 that can be read as ( fried , 1966 ) , @xmath211 let the covariance of the phase , @xmath212 be defined as , @xmath213 and by replacing @xmath204 , @xmath214 where @xmath215 and the 3-d refractive index covariance is , @xmath216 since the thickness of the layer , @xmath202 , is large compared to the correlation scale of the turbulence , the integration over @xmath217 from @xmath218 , leads to , @xmath219 the phase structure function is related to its covariance ( equation 35 ) ; therefore , @xmath220 d\zeta.\ ] ] the refractive index structure function defined in equation ( 37 ) is evaluated as , @xmath221 and using equation ( 35 ) , equation ( 48 ) can be integrated to yield , @xmath222 the covariance of the phase is deduced by substituting equation ( 50 ) , in equation ( 42 ) , @xmath223 using the fresnel approximation , the covariance of the phase at the ground level due to a thin layer of turbulence at some height off the ground is given by , @xmath224 for high altitude layers the complex field will fluctuate both in phase and in amplitude ( scintillation ) , and therefore , the wave structure function , @xmath210 , is not strictly true as at the ground level .
the turbulent layer acts like a diffracting screen ; however , correction in the case of astronomical observation remains small ( roddier , 1981 ) .
the wave structure function after passing through @xmath225 layers can be expressed as the sum of the @xmath225 wave structure functions associated with the individual layer . for each layer ,
the coherence function is multiplied by the term , @xmath226}$ ] ; therefore , the coherence function at ground level is given by , @xmath227 } \nonumber\\ & & = e^{-\frac{1}{2 } \left[2.91k^2\xi^{5/3 } \sum_{j=1}^n { \cal c}_n^2(h_j)\delta h_j\right]}. \end{aligned}\ ] ] this expression may be generalized for a star at an angular distance @xmath228 away from the zenith viewed through all over the turbulent atmosphere , @xmath229}.\ ] ] the term ` seeing ' is the total effect of distortion in the path of the star light via different contributing layers of the atmosphere to the detector placed at the focus of the telescope . let the mtf of the atmosphere and a telescope together be described as in figure 1 .
the long - exposure psf is defined by the ensemble average , @xmath230 , which is independent of the direction .
if the object emits incoherently , the average illumination , @xmath231 , of a resolved object , @xmath232 , obeys the convolution relationship , @xmath233 using 2-d ft , the above equation translates into , @xmath234 where @xmath235 denotes the transfer function for long - exposure images , @xmath47 , the spatial frequency vector with magnitude , @xmath236 and @xmath237 the object spectrum .
the argument of equation ( 56 ) is expressed as , @xmath238 where @xmath239 is the fourier phase at @xmath47 , @xmath240 stands for , ` the phase of ' , and @xmath241 the apertures , corresponding to the seeing cells .
the transfer function is the product of the atmosphere transfer function ( wave coherence function ) , @xmath242 , and the telescope transfer function , @xmath243 , @xmath244 for a long - exposure through the atmosphere , the resolving power , @xmath64 , of any optical telescope can be expressed as , @xmath245 it is limited either by the telescope or by the atmosphere , depending on the relative width of the two functions , @xmath242 and @xmath243 .
@xmath246 [ [ frieds - parameter ] ] fried s parameter + + + + + + + + + + + + + + + + + according to equation ( 54 ) , @xmath242 , can be expressed as , @xmath247}. \end{aligned}\ ] ] therefore , equation ( 61 ) is translated into , @xmath248^{-6/5 } \gamma(6/5).\ ] ] fried , ( 1966 ) had introduced the critical diameter @xmath249 , for a telescope ; therefore , placing @xmath250 in equation ( 60 ) , equation ( 62 ) takes the following form , @xmath251 the phase structure function ( equation 43 ) across the telescope aperture ( fried , 1966 ) becomes , @xmath252 by replacing the value of @xmath253 , in equation ( 54 ) , an expression for @xmath249 in terms of the distribution of the turbulence in the atmosphere is found to be .
@xmath254^{-3/5}.\ ] ] the fried s parameter may be thought of as the diameter of telescope that would produce the same diffraction - limited fwhm of a point source image as the atmospheric turbulence would with an infinite - sized mirror .
[ [ seeing - at - the - telescope - site ] ] seeing at the telescope site + + + + + + + + + + + + + + + + + + + + + + + + + + + + the major sources of image degradation predominantly comes from thermal and aero - dynamic disturbances in the atmosphere surrounding the telescope and its enclosure .
these sources include : ( i ) convection in and around the building and the dome , obstructed location near the ground , off the surface of the telescope structure , ( ii ) thermal distortion of the primary and secondary mirrors when they are warmer than the ambient air , ( iii ) dissipation of heat by the secondary mirror ( zago , 1995 ) , ( iv ) rise in temperature at the primary mirror cell , and ( v ) at the focal point causing temperature gradient close to the detector .
degradation in image quality can occur due to opto - mechanical aberrations as well as mechanical vibrations of the optical system .
various corrective measures have been proposed to improve the seeing .
these measures include : ( i ) insulating the surface of the floors and walls , ( ii ) introducing an active cooling system to eliminate the heat produced by electric equipment on the telescope and elsewhere in the dome , and ( iii ) installing a ventilator to generate a downward air flow through the slit to counteract the upward action of the bubbles ( racine , 1984 , ryan and wood , 1995 ) .
floor - chilling systems to dampen the natural convection have been implemented which keeps the temperature of the primary mirror closer to the air volume ( zago , 1995 ) .
saha and chinnappan ( 1999 ) have found that the average observed @xmath249 is higher during the later part of the night than the earlier part .
this change might indicate that the slowly cooling mirror creates thermal instabilities that decrease slowly during the night .
ever since the development of the si technique ( labeyrie , 1970 ) , it is widely employed both in the visible , as well as in the infrared ( ir ) bands at telescopes to decipher diffraction - limited informations . the following sub - sections deal with single aperture speckle imaging and related avenues , other techniques , ao imaging systems , dark speckle imaging , and high resolution sensors .
if a point source is imaged through the telescope by using the pupil function consisting of two sub - apertures ( @xmath255 ) , corresponding to the two seeing cells separated by a vector @xmath256 , a fringe pattern is produced with narrow spatial frequency bandwidth that moves within broad psf envelope ; with increasing distance between the sub - apertures , the fringes move with an increasingly larger amplitude .
the introduction of a third sub - aperture gives three pairs of sub - apertures and yields the appearance of three intersecting patterns of moving fringes .
covering the telescope aperture with @xmath249-sized sub - apertures synthesizes a filled aperture @xmath257 ( each pair of them , @xmath258 , is separated by a baseline ) interferometer .
the intensity at the focal - plane , @xmath23 , according to the diffraction theory ( born and wolf , 1984 ) , is determined by the expression , @xmath259 the term , @xmath260 , is multiplied by @xmath261 , where @xmath262 is the random instantaneous shift in the fringe pattern .
each sub - aperture is small enough for the field to be coherent over its extent .
atmospheric turbulence does not affect the contrast of the fringes produced but introduces phase delays .
if the integration time is shorter than the evolution time of the phase inhomogeneities , the interference fringes are preserved but their phases are randomly distorted , which produces ` speckles ' .
( the formation of speckles stems from the summation of coherent vibrations having random characteristics .
it can be modeled as a 2-dimensional random walk with fresnel s vector representation of vibrations ) .
each speckle covers an area of the same order of magnitude as the airy disc of the telescope .
the number of correlation cells is proportional to the square of @xmath263 and the number of photons , @xmath264 , per speckle is independent of its diameter .
the lifetime of speckles , @xmath265 , where @xmath123 is the velocity dispersion in the turbulent seeing layers across the line of sight .
the structure of the speckle pattern changes randomly over a short interval of time .
the sum of several such statistically uncorrelated patterns from a point source can result in a uniform patch of light a few arcseconds ( @xmath2 ) wide .
figures 2 and 3 depict the speckles of a binary star hr4689 and the results of summing 128 specklegrams , respectively .
averaging @xmath266 over many frames , the resultant for frequencies greater than @xmath267 , tends to zero because the phase - difference , @xmath268 , mod @xmath269 , between the two apertures is distributed uniformly between @xmath270 , with zero mean .
the fourier component performs a random walk in the complex plane and averages to zero : @xmath271 in general , a high quantum efficiency detector is needed to record magnified short - exposure images for such observation . to compensate for atmospherically induced dispersion at zenith angles larger than a few degrees , either a counter - rotating computer - controlled dispersion - correcting prism or a narrow - bandwidth filter
is used .
speckle interferometry estimates a ` power spectrum ' which is the ensemble average of the squared modulus of an ensemble of ft from a set of specklegrams , @xmath272 .
the intensity of the image , @xmath273 , in the case of quasi - monochromatic incoherent source can be expressed as , @xmath274 where @xmath275 , is an object at a point anywhere in the field of view .
the variability of the corrugated wavefront yields ` speckle boiling ' and is the source of speckle noise that arises from difference in registration between the evolving speckle pattern and the boundary of the psf area in the focal - plane .
these specklegrams have additive noise contamination , @xmath276 , which includes all additive measurement of uncertainties .
this may be in the form of ( i ) photon statistics noise , and ( ii ) all distortions from the idealized iso - planatic model represented by the convolution of @xmath232 with @xmath277 that includes non - linear geometrical distortions .
for each of the short - exposure instantaneous records , the imaging equation applies , @xmath278 denoting @xmath279 , for the noise spectrum .
the fourier space relationship between object and the image is @xmath280 taking the modulus square of the expression and averaging over many frames , the average image power spectrum is , @xmath281 since @xmath282 is a random function in which the detail is continuously changing , its ensemble average becomes smoother . by the wiener - khintchine theorem (
mendel and wolf , 1995 ) , the inverse ft of equation ( 73 ) gives the autocorrelation of the object , @xmath283 $ ] .
@xmath284 = { \cal f}^{- } [ |\widehat{\cal o}{\bf ( u)}|^2],\ ] ] in this technique , the atmospheric phase contribution is eliminated but the averaged signal is non - zero , i.e. , @xmath285 the argument of the equation ( 72 ) is given by the expression , @xmath286 the transfer function of @xmath277 , is generally estimated by calculating the power spectrum of the instantaneous intensity from an unresolved star .
saha and maitra ( 2001 ) developed an algorithm , where a wiener parameter , @xmath287 , is added to psf power spectrum .
the classic wiener filter that resulted from electronic information theory where diffraction - limits do not mean much , is meant to deal with signal dependent ` colored ' noise . in practice , this term is usually just a constant , a ` noise control parameter ' whose scale is estimated from the noise power spectrum . in this case , it assumes that the noise is white and that one can estimate its scale in regions of the power spectrum where the signal is zero ( outside the diffraction - limit for an imaging system ) .
@xmath288}.\ ] ] the si technique in the case of the components in a group of stars retrieves the separation , position angle with 180@xmath289 ambiguity , and the relative magnitude difference at low light levels .
figure 4 depicts the autocorrelation of a binary system , hr4689 .
another algorithm called directed vector autocorrelation ( dva ) method is found to be effective in eliminating the 180@xmath289 ambiguity ( bagnuolo et al . 1992 ) .
if a reference point source is available within the iso - planatic patch ( @xmath290 ) , it is used as a key to reconstruct the target in the same way as a reference coherent beam is employed in holographic reconstruction ( liu and lohmann , 1973 ) .
let the point source be represented by a dirac impulse , @xmath291 at the origin and @xmath292 be the nearby object to be reconstructed .
the intensity distribution in the field of view is @xmath293 the squared modulus of its ft is derived as , @xmath294 the inverse ft of this equation translates into , @xmath295 = a_\circ^2\delta{(\bf x ) } + a_\circ { \cal o}_1{(\bf x ) } + a_\circ{\cal o}_1{(\bf - x ) } + { \cal a}[{{\cal o}_1}{(\bf x)}],\ ] ] the first and the last terms in equation ( 80 ) are centered at the origin .
if the object is far enough from the reference source , @xmath232 , its mirror image , @xmath296 , is therefore recovered apart from a 180@xmath289 rotation ambiguity .
differential speckle interferometry ( dsi ) is a method to observe the objects in different spectral channels simultaneously and to compute the average cross - correlation of pairs of speckle images ( beckers , 1982 ) .
let @xmath297 and @xmath298 , be respectively the source brightness distributions at @xmath299 and @xmath300 , and @xmath301 and @xmath302 , their associated instantaneous image intensity distributions .
the relation between the object and the image in the fourier space becomes , @xmath303 where @xmath304 and @xmath305 are the related transfer functions .
the average cross - spectrum is given by , @xmath306 the transfer function , @xmath307 , can be calibrated on the reference point source for which , @xmath308 . if the two spectral windows are close enough ( @xmath309 ) , the instantaneous transfer function is assumed to be identical in both the channels [ @xmath310 .
therefore , equation ( 83 ) becomes , @xmath311 the noise contributions from two different detectors are uncorrelated , and thereby their contributions cancel out .
the dsi estimates the ratio , @xmath312 and the differential image , @xmath313 , is obtained by performing an inverse ft of this ratio , @xmath314,\ ] ] where @xmath313 represents an image of the object in the emission feature having the resolution of the object imaged in the continuum .
when any planetary body of a notable size passes in front of a star , the light coming from the latter is occulted .
the profiles of stellar occultations by the moon show diffraction patterns as the star is being occulted , provided the data is recorded at high time resolution .
the method remains useful because of the extraordinary geometric precision it provides .
the notable advantage of occultation of binary stars is that it can determine relative intensities and measure the separations comparable to those measured by long baseline interferometers .
speckle surveys have resolved known occulting binaries down to a separation of about @xmath315 0.025@xmath2 ( mason , 1996 ) .
further , this method provides a means of determining the limiting magnitude difference of si .
the shortcomings of this technique may be noted due to its singular nature : the object may not occult again until one saros cycle later ( 18.6 yr ) , and must be limited to a belt of the sky ( 10% of the celestial sphere ) .
the application of the si technique to speckle spectroscopic observations enables one to obtain spectral resolution with high spatial resolution of astronomical objects simultaneously .
the intensity distribution , @xmath316 , of an instantaneous objective prism speckle spectrogram is expressed as , @xmath317 where @xmath318 denotes the intensity of m@xmath319 object pixel and @xmath320 is the spectrum of the object pixel . in the narrow wavelength bands ( @xmath31530 nm ) ,
the psf , @xmath277 , is wavelength independent .
the objective prism spectrum , @xmath321 , can be reconstructed from the speckle spectrograms . in a speckle spectrograph , either a prism or a grism can be employed to disperse 1-d specklegrams ( grieger et al .
an imaging spectrometer uses two synchronized detectors to record the dispersed speckle pattern and the specklegrams of the object ( baba , kuwamura et al .
1994 ) ; a reflection grating acts as disperser .
in general , radiation is polarized and the measurement of polarization parameters is important in understanding of the emission mechanisms .
processes such as electric and magnetic fields , scattering , chemical interactions , molecular structure , and mechanical stress cause changes in the polarization state of an optical beam .
applications relying on the study of these changes cover a vast area , among them are astrophysics and molecular biology .
the importance of such observations in astronomy is to obtain information such as the size and shapes of dust envelopes around stars , the size and shape of the dust grains , and magnetic fields . among other astronomical objectives worth investigating
are : ( i ) the wavelength dependence of the degree of polarization and the rotation of the position angle in stars with extended atmospheres , ( ii ) the wavelength dependence of the degree of polarization and position angle of light emitted by stars present in very young ( @xmath322 years ) clusters and associations .
the modified incident polarization caused by the reflection of a mirror is characterized by two parameters : ( i ) the ratio between the reflection coefficients of the electric vector components which are perpendicular and parallel to the plane of incidence , known as @xmath323 and @xmath324 components respectively , ( ii ) the relative phase - shift between these electric vibrations .
the effect on the statistics of a speckle pattern is the degree of depolarization caused by the scattering at the surface .
if the light is depolarized , the resulting speckle field is considered to be the sum of two component speckle fields produced by scattered light polarized in two orthogonal directions .
the intensity at any point is the sum of the intensities of the component speckle patterns ( goodman , 1975 ) .
these patterns are partially correlated ; therefore , a polarizer that transmits one of the component speckle patterns is used in the speckle camera system ( falcke et al .
the advantage of using a speckle camera over a conventional imaging polarimeter is that it helps in monitoring the short - time variability of the atmospheric transmission .
image recovery is relatively simple where the target is a point source .
nevertheless , interferometric observations can reveal the fundamental processes on the sun that take place on sub - arcsecond scales concerning convection and magnetic fields .
the limitations come from ( i ) the rapid evolution of solar granulation that prevents the collection of long sequences of specklegrams for reconstruction , ( ii ) the lack of efficient detectors to record a large number of frames within the stipulated time before the structure changes .
another major problem of reconstructing images comes from difficulty in estimating the psf due to the lack of a reference point source .
the spectral ratio technique ( von der lhe , 1984 ) , which is based on a comparison between long and short - exposure images , has been employed ( wilken et al . 1997 ) to derive fried s parameter .
models of the speckle transfer function ( korff , 1973 ) and of average short - exposure mtf have also been applied to compare the observed spectral ratios with theoretical values .
high resolution solar images obtained during partial solar eclipse may help in estimating the seeing effect ( callados and vzquez , 1987 ) . the limb of the moon eclipsing the sun provides a sharp edge as a reference object .
the intensity profile falls off sharply at the limb .
the departure of this fall off gives an indirect estimate of the atmospheric psf .
several other methods , viz .
, pupil - plane techniques such as wavefront shearing interferometry , phase - closure methods , and phase - diversity techniques are also employed at single telescope in order to obtain diffraction - limited information .
shearing interferometers make use of the principle of self referencing , that is , they combine the wavefront with a shifted version of itself to form interferences .
fringes are produced by two partially or totally superimposed pupil images created by introducing a beam splitter . at each point
, interference occurs from the combination of only two points on the wavefronts at a given baseline , and therefore , behaves as an array of michelson - fizeau interferometers .
an important property of these interferometers is their ability to work with partially coherent light , which offers better s / n ratio on bright sources , and is insensitive to calibration errors due to seeing fluctuations and telescope aberrations .
a rotational shear interferometer was used at the telescope to map the visibility of fringes produced by @xmath0 ori ( roddier and roddier , 1988 ) . in this technique ,
the 2-d ft is obtained by rotating one pupil image about the optical axis by a small angle with respect to the other . if the rotation axis coincides with the center of the pupil , the two images overlap .
all the object fourier components within a telescope s diffraction cutoff frequency are measured simultaneously .
the phase of the visibility may be deduced from a closure - phase that is insensitive to the atmospherically induced random phase errors , as well as to the permanent phase errors introduced by the imaging systems ( jennison , 1958 ) using three telescopes .
the observed phases , @xmath325 , on the different baselines contain the phases of the source fourier components @xmath326 and also the error terms , @xmath327 , introduced by errors at the individual antennas and by the atmospheric variations at each antenna .
the observed fringes are represented by the following equations , @xmath328 where the subscripts refer to the antennae at each end of a particular baseline .
the closure phase , @xmath329 is the sum of phases of the source fourier components and is derived as , @xmath330 this equation implies cancellations of the antennae phase errors . using the measured closure phases and amplitudes as observables , the object phases are determined ( mostly by least square techniques , viz . , singular value decomposition , conjugate gradient method ) . from the estimated object phases and the calibrated amplitudes ,
the degraded image is reconstructed .
baldwin et al . ( 1986 ) reported the measurements of the closure - phases obtained at a high light level with a three hole aperture mask set in the pupil - plane of the telescope .
the non - redundant aperture masking method , in which the short - exposure images are taken through a multi - aperture screen , has several advantages .
these are : ( i ) an improvement of signal - to - noise ( s / n ) ratios for the individual visibility and closure - phase measurements , ( ii ) attainment of the maximum possible angular resolution by using the longest baselines , and ( iii ) built - in delay to observe objects at low declinations .
but the system is restricted to high light levels , because the instantaneous coverage of spatial frequencies is sparse and most of the available light is discarded .
phase - diversity ( gonsalves , 1982 , paxman et al .
1992 ) is a post - collection technique that uses a number of intensity distributions encoded by known aberrations for restoring high spatial resolution detail while imaging in the presence of atmospheric turbulence .
the phase aberrations are estimated from two simultaneously recorded images .
phase - diverse speckle is an extension of this technique , whereby a time sequence of short - exposure image pairs is detected at different positions in focus and out of focus near the focal - plane .
incident energy is split into two channels by a simple beam splitter : one is collected at a conventional focal - plane , the other is defocussed ( by a known amount ) and a second detector array permits the instantaneous collection of the latter .
it is less susceptible to the systematic errors caused by the optical hardware , and is found to be more appealing in astronomy ( baba , tomita et al .
1994 , seldin and paxman , 1994 ) .
significant technological innovations over the past several years have made it possible to correct perturbations in the wavefronts in real time by incorporating a controllable phase distortion in the light path , which is opposite to that introduced by the atmosphere ( babcock , 1953 , rousset et al .
this technique has advantages over post - detection image restoration techniques that are limited by noise .
adaptive - optics ( ao ) systems are employed in other branches of physics as well .
liang et al . (
1997 ) have constructed a camera equipped with adaptive - optics that allows one to image a microscopic size of single cell in the living human retina .
they have shown that a human eye with adaptive - optics correction can resolve fine gratings that are invisible to the unaided eye .
ao systems are useful for spectroscopic observations , as well as for low light level imaging with future very large telescopes , and ground - based lbois .
turbulence cells are blown by wind across the telescope aperture , hence , the wind velocity dictates the speed with which a corrective action must be taken .
greenwood ( 1977 ) derives the mean square residual wavefront error as a function of servoloop bandwidth for a first order controller , which is given by , @xmath331 where @xmath332 is the closed loop bandwidth of the wavefront compensator and @xmath333 the greenwood frequency that is defined by the relation , @xmath334 where @xmath335 is the wind velocity in the turbulent layer of air . for imaging in near - ir to ultraviolet ,
the ao system bandwidths need to have a response time of order several hundreds to 1 khz .
it is easier to achieve diffraction - limited information using ao systems at longer wavelengths since @xmath249 is proportional to the six - fifths power of the wavelength , @xmath336 the above equation ( 94 ) implies that the width of seeing limited images , @xmath337 varies with @xmath3 .
the number of degrees of freedom , i.e. , the number of actuators on the deformable mirror ( dm ) and the number of sub - aperture in the wavefront sensor , in an ao system should be determined by the following equation , @xmath338 the required components for implementing an ao system are wavefront sensing , wavefront phase error computation and a flexible mirror whose surface is electronically controlled in real time to create a conjugate surface enabling compensation of the wavefront distortion ( roggemann et al . 1997 and references therein ) . in order to remove the low frequency tilt error , generally the incoming collimated beam is fed by a tip - tilt mirror . after traveling further ,
it reflects off of a dm that eliminates high frequency wavefront errors . a beam - splitter devides the beam into two parts : one
is directed to the wavefront sensor to measure the residual error in the wavefront and to provide information to the actuator control computer to compute the dm actuator voltages and the other is focused to form an image .
implementation of the dynamically controlled active optical components consisting of a tip - tilt mirror system in conjunction with closed - loop control electronics has several advantages : ( i ) conceptually the system is simple , ( ii ) fainter guide stars increase the sky coverage , and ( iii ) field of view is wider ( glindemann , 1997 ) .
these systems are limited to two zernike modes ( x and y tilt ) , while a higher order system compensating many zernike mode ( zernike polynomials are an orthogonal expansion over unit circle ) is required to remove high frequency errors .
a variety of dms have been developed for the applications of ( i ) high energy laser focusing , ( ii ) laser cavity control , ( iii ) compensated imagery through atmospheric turbulence etc .
several wavefront sensors such as ( i ) lateral shearing interferometer , ( ii ) shack - hartman ( sh ) sensor , and ( iii ) the curvature sensor are in use as well .
the technical details of these dms and sensors are enumerated in the recent literatures ( roggemann et al .
1997 , roddier , 1999 , and saha , 1999 ) .
the phase reconstruction method finds the relationship between the measurements and the unknown values of the wavefronts and can be categorized as being either zonal or modal , depending on whether the estimate is either a phase value in a local zone or a coefficient of an aperture function ( rousset , 1999 ) . in the case of curvature sensing , the computed sensor signals are multiplied by a control matrix to convert wavefront slopes to actuator control signals , the output of which are the increments to be applied to the control voltages on the dm .
a conjugate shape is created using these data by controlling a dm .
the real time computation of the wavefront error , as well as correction of wavefront distortion involves digital manipulation of wavefront sensor data in the wavefront sensor processor , the re - constructor and the low - pass filter , and converting to analog drive signals for the dm actuators .
the functions are ( i ) to compute sub - aperture gradients , phases at the corners of each sub - aperture , low - pass filter phases , and ( ii ) to provide actuator offsets to compensate the fixed optical system errors and real time actuator commands for wavefront corrections .
an ao system requires a reference source to monitor the atmospheric aberrations for measuring the wavefront errors , as well as for mapping the phase on the entrance pupil .
it is generally not possible to find a sufficiently bright reference star close enough to a target star . in order to alleviate this limitation ,
many observatories are currently implementing the artificial laser guide star ( ragazzoni and bonnacini , 1996 , racine et al .
1996 , lloyd - hart et al .
however , the best results are still obtained with natural guide stars .
an artificial guide star can be obtained using either the resonance scattering by sodium in the mesosphere at 90 km ( foy and labeyrie , 1985 ) or rayleigh scattering between 10 and 20 km altitude ( fugate et al . 1994 ) .
a pulsed laser ( tuned to the sodium @xmath339 line to excite sodium atom ) is used to produce a bright compact glow in the upper atmosphere . concerning the flux backscattered by a laser shot ,
thompson and gardner ( 1988 ) stressed the importance of investigating two basic problems : ( i ) the cone effect which arises due to the parallax between the remote astronomical source and artificial source , and ( ii ) the angular anisoplanatic effects .
these effects can be restored by imaging the various turbulent layers of the atmosphere onto different adaptive mirrors ( tallon et al .
scattering of the upward propagating laser beam is due to rayleigh scattering , mostly by @xmath340 molecules .
mie scattering by aerosol or cirrus clouds may be important at lower altitudes but are usually variable and transient .
multi - conjugate ao system enables near - uniform compensation for the atmospheric turbulence over considerably larger field of view than can be corrected with normal ao system .
this method employs an ensemble of guide stars that allows for 3-d tomography of the atmospheric turbulence and a number of altitude - conjugate dms to extend the compensated field of view . however , its performances depends on the quality of the wavefront sensing of the individual layers .
ragazzoni et al .
( 2000 ) have demonstrated this type of tomography .
this new technique pushes the detection limit by @xmath1941.7 mag on unresolved objects with respect to seeing limited images ; it also minimizes the cone effect .
this technique will be useful for the extremely large telescopes of 100 m class , e.g. , the overwhelmingly large ( owl ) telescope ( dierickx and gilmozzi , 1999 ) .
however , the limitations are mainly related to the finite number of actuators in a dm , wavefront sensors , and guide stars .
the usage of a adaptive secondary mirror ( asm ) for corrections making relay optics obsolete that are required to conjugate a dm at a reimaged pupil , as well as to minimize thermal emission is a new key innovation ( bruns et al . 1997 ) .
the other notable advantages are ( i ) enhanced photon throughput that measures the proportion of light which is transmitted through an optical set - up , ( ii ) introduction of negligible extra ir emissivity , ( iii ) causes no extra polarization , and ( iv ) non - addition of reflective losses ( lee et al .
2000 ) . due to the interactuator spacing , the resonant frequency of such
a mirror may be lower than the ao bandwidth .
the asm system uses a sh sensor with an array of small lenslets , which adds two extra refractive surfaces to the wavefront sensor optical beam ( lloyd - hart , 2000 ) .
an @xmath341 ao secondary with 336 actuators is in the final stages of testing and will be installed on the 6.5 m telescope of mmt observatory , mt .
hopkins , arizona in 2002 ( wehinger , 2001 ) .
the relevance of using coronagraphy in imaging or spectroscopy of faint structure near a bright object can be noted in terms of reducing the light coming from the central star , and filtering out of the light at low spatial frequency ; the remaining light at the edge of the pupil corresponds to high frequencies .
a coronagraph reduces off - axis light from an on - axis source with an occulting stop in the image - plane as well as with a matched lyot stop in the next pupil plane . while using the former
stop the size of the latter pupil should be chosen carefully to find the best trade - off between the throughput and image suppression .
the limitations come from the light diffracted by the telescope and instrument optics .
coronagraphy with dynamic range can be a powerful tool for direct imaging of extra - solar planets .
nakajima ( 1994 ) estimates that imaging with such a method with a low order ao system in a 6.5 m telescope could detect jupiter - size extra - solar planets at separation @xmath1941.5@xmath2 with s / n of 3 in 10@xmath342 s. rouan et al .
( 2000 ) describes a four - quadrant phase - mask coronagraph where a detection at a contrast of more than 8 mag difference between a star and a planet is feasible .
the dark speckle method uses the randomly moving dark zones between speckles @xmath197 ` dark speckles ' .
it exploits the light cancellation effect in a random coherent field ; highly destructive interferences that depict near black spots in the speckle pattern ( labeyrie , 1995 ) may occur occasionally .
the aim of this method is to detect faint objects around a star when the difference of magnitude is significant .
if a dark speckle is at the location of the companion in the image , the companion emits enough light to reveal itself .
the required system consists of a telescope with an ao system , a coronagraph , a wynne corrector , and a fast photon - counting camera with a low dark noise .
if a pixel of the photon - counting camera is illuminated by the star only ( in the airy rings area ) , because of the ao system , the number of photons in each pixel , for a given interval ( frame ) , is statistically given by a bose - einstein distribution .
the number of photons per frame in the central peak of the image of a point source obeys a classical poisson distribution . for the pixels containing the image of the companion , the number of photons , resulting from both the star and the companion ,
is given by a different distribution ( computed by mixing bose - einstein and poisson distributions ) .
one noticeable property is that the probability to get zero photons in a frame is very low for the pixels containing the image of the companion , and much higher for the pixels containing only the contribution from the star .
therefore , if the ` no photon in the frame ' events for each pixel is counted , and for a very large number of frames , a ` dark map ' can be built that will show the pixels for which the distribution of the number of photons is not bose - einstein type , therefore revealing the location of a faint companion .
the role of the wynne corrector is to give residual speckles the same size regardless the wavelength .
otherwise , dark speckles at a given wavelength would be overlapped by bright speckles at other wavelengths . with
the current technology , by means of the dark speckle technique at a 3.6 m telescope should allow detection of a companion with @xmath3436 - 7 mag .
figures 5 and 6 depict the coronagraphic images of the binary stars , hd192876 and hd222493 , respectively ( boccaletti et al . 2001 ) ; the data were obtained with adonis in the k band ( 2.2 @xmath344 m ) on the european southern observatory s ( eso ) 3.6 m telescope . due to the lack of a perfect detector ( no read - out noise ) at near - ir band , every pixel under the defined threshold (
a few times the read - out noise ) is accounted as a dark speckle .
phase boiling , a relatively new technique that consists of adding a small amount of white noise to the actuators in order to get a fast temporal decorrelation of the speckles during long - exposure acquisition , may produce better results .
aime ( 2000 ) has computed the s / n ratio for two different cases : short - exposure and long - exposure . according to him , even with an electron - noise limited detector like a ccd or a near - ir camera multi - object spectrometer ( nicmos )
, the latter can provide better results if the halo has its residual speckles smoothed by fast residual ` seeing ' acting during the long - exposure than building a dark map from short - exposures in the photon - counting mode .
artificial very fast seeing can also be generated by applying fast random noise to the actuators , at amplitude levels comparable to the residual seeing left over by the ao system .
the question is , what is easiest : dark speckle analysis or a ` hyper - turbulated ' long - exposure ? labeyrie ( 2000 ) made simulations supporting the aime s ( 2000 ) results .
boccaletti ( 2001 ) has compared the dark speckle signal - to - noise ratio ( snr ) with the long - exposure snr ( angel , 1994 ) .
the speckle lifetime has to be of order 0.1 ms .
currently it is impossible to drive a dm at this frequency ( 10 khz ) . with the 5 m palomar telescope boccaletti ( 2001 )
tried to smooth the speckle pattern by adding a straightforward random noise on the actuators ( the dm is equipped with 241 actuators ) at maximum speed of 500 hz .
effectively , the halo is smoothed , but its intensity is also increased , so that the companion snr is actually decreased .
blurring the speckle pattern would probably require wavefront sensor telemetry ; implementation of a hyper - turbulated long - exposure at the palomar is still under study ( boccaletti , 2001 ) .
all the techniques that are described above require a high quality sensor so as to enable one to obtain snap shots with a very high time resolution of the order of ( i ) frame integration of 50 hz , or ( ii ) photon recording rates of a few mhz .
the performance relies on the characteristics of such sensors , e.g. , ( i ) the spectral bandwidth , ( ii ) the quantum efficiency , ( iii ) the detector noise that includes dark current , read - out and amplifier noise , ( iv ) the time lag due to the read - out of the detector , and ( v ) the array size and the spatial resolution .
the frame - transfer intensified ccd ( iccd ) camera employs micro - channel plate ( mcp ) as an intensifier .
the photoelectron is accelerated into a channel of the mcp releasing secondaries and producing an output charge cloud of about @xmath345 electrons with 5 - 10 kilovolt ( kv ) potential . with further applied potential of @xmath1945 - 7 kv ,
these electrons are accelerated to impact a phosphor , thus producing an output pulse of @xmath346 photons .
these photons are directed to the ccd by fiber optic coupling .
the main disadvantage of such system is the poor gain statistics results in the introduction of a noise factor between 2 and 3.5 .
recent development of a non - intensified ccd device which effectively reduces readout noise to less than one electron rms has enabled substantial internal gain within the ccd before the signal reaches the output amplifier ( mackay et al .
such a detector , although the photon counting performance appears moderate for the moment , shows promise for quantitative measurement of diffraction - limited stellar images .
the marked advantage of a photon - counting system is that of reading the signal a posteriori to optimize the correlation time of short - exposures in order to overcome the loss of fringe visibility due to the speckle lifetime ; the typical values for an object of m@xmath347 = 12 over a field of 2.5@xmath2 are less than 50 photons / ms through a narrow - band filter .
the other notable features are , ( i ) capability of determining the position of a detected photon , ( ii ) ability to register individual photons with equal statistical weight and produces signal pulse ( with dead time of a few ns ) , and ( iii ) low dark noise .
the major short comings of the photon - counting system that is based on frame integration ( blazit , 1986 ) arise from the ( i ) calculations of the coordinates which are hardware - limited , and ( ii ) limited dynamic range of the detector .
non - detectability of a pair of photons closer than a minimum separation by the detector yields a loss in high frequency information ; this , in turn , produces a hole in the center of the autocorrelation @xmath197 centreur hole , resulting in the degradation of the power spectra or bispectra ( ft of triple correlation ) of speckle images .
several 2-d photon - counting sensors that allow recording of the position and time of arrival of each detected photons have been developed such as , ( i ) precision analog photon address ( papa ; papaliolios et al .
1985 ) , ( ii ) resistive anode position detector ( clampin et al . 1988 ) , ( iii ) multi anode micro - channel array ( mama ; timothy , 1993 ) , ( iv ) wedge - and - strip anodes , ( v ) delay - line anodes , ( vi ) silicon anode detector etc .
baring papa which is based on a high gain image intensifier and a set of photomultiplier tubes , these sensors detect the charge cloud from a high gain mcp .
they provide spatial event information by means of the position sensitive readout set - up ; the encoding systems identify each event s location .
the short - comings of the mcps are notably due to its local dead - time which essentially restricts the conditions for use of these detectors for high spatial resolution applications .
these constraints are also related with the luminous intensity and the pixel size . in the infrared band ,
no photon - counting is possible with the current technology .
nevertheless , a near - ir focal - plane array , nicmos , has been developed .
it consists of 256@xmath348256 integrating detectors organized in four independent 128@xmath348128 quadrants and is fabricated in hgcdte grown on a saphire substrate that is very rugged and provides a good thermal contraction match to silicon multiplexer ( cooper et al .
the typical nicmos3 fpas have read noise less than 35 e- with less than 1 e-/sec detector dark current at 77 k and broadband quantum efficiency is better than 50% in the range of 0.8 to 2.5 @xmath344 m .
modern technology has solved many of the problems that were pioneered by michelson and pease ( 1921 ) .
the light collected by an array of separated telescopes could be coherently combined to measure the fourier components of the brightness distribution of a star .
while at an interesting stage of development , currently of more limited imaging capabilities , the following sub - sections elucidate the current state of the art of such arrays .
the potential of the aperture - synthesis interferometry in the optical domain is demonstrated by the spectacular images produced with aperture - masking of a single telescope ( tuthill et al .
the principle behind this method may be considered as the operation of a conventional filled - aperture telescope composed of @xmath225 elemental areas , in which there are @xmath349 independent baselines , with @xmath350 unknown phase errors .
this implies that by using many telescopes in an interferometric array , most of the phase information can be retrieved .
the signal in the @xmath183th area due to a source of emission is expressed as , @xmath351 where @xmath352 is the amplitude of the signal and @xmath353 the relative phase of the radiation .
if these signals are added together vectorily and time averaged , the intensity of the light @xmath354 is derived as , @xmath355 the first term is proportional to the sum of the power received by the elementary areas .
the resolving power is derived from the cross product .
each term can equally be measured with two elementary areas in positions @xmath356 and @xmath357 .
the term @xmath358 is expressed as , @xmath359 where @xmath360 is the separation of the two elemental areas , @xmath361 the unit vector defining to the source . generally three or more telescopes are required for aperture - synthesis , two can suffice if the object includes a point source usable as a phase reference .
different spectral channels are employed for differential visibility measurements ( continuum and spectral line ) and use data from one part of the spectrum such as the continuum emission to calibrate another part ( mourard et al .
the continuum channel is supposed to originate from the unresolved region of a star , e.g. , photosphere and the spectral line centered on a part of the spectrum created in an extended region , such as circumstellar medium .
the phase - referenced technique with a fringe - tracking channel was also used at mark iii interferometer to recover the phase information ( quirrenbach et al .
another convenient method is to measure the instantaneous phases of fringes from a bright point source that lies within the iso - planatic patch in order to correct the corrupted phases on the target .
shao and colavita ( 1992 ) used fringe - phase information to determine precise relative positions of nearby stars .
the direct measurements of the closure phase together with the measurements of visibility amplitude allow one to reconstruct an image of any object using three or more independent telescopes
. this technique has been successfully demonstrated by baldwin et al .
( 1998 ) in the visible band at the cambridge optical aperture - synthesis telescope ( coast ) .
each pair of telescopes in an array yields a measure of the amplitude of the spatial coherence function of the object at a spatial frequency @xmath362 . in order to make an image from an interferometer
, one needs estimates of the complex visibilities over a large portion of the @xmath363 plane , both the amplitudes and phases .
the @xmath363 coordinates corresponding to a snapshot projection of the baseline that are sampled from a star of declination @xmath364 , when its hour angle is @xmath365 , is given by ( fomalont and wright , 1974 ) : @xmath366 where @xmath367 and @xmath368 are orthogonal east - west and north - south components of the baseline vector at the ground of an interferometer located at the terrestrial latitude @xmath369 .
the accurate determination of relative - positions of stars will provide crucial data for astrophysics .
for example , precise parallax distances of cepheids will help to establish a period / absolute magnitude relationship in order to calibrate distances of galaxies , thus reducing the uncertainty on the value of @xmath370 .
an interferometer measures the angle projected onto baseline .
stellar fringes must be observed at two or more baseline orientations to determine two angular coordinates of an astronomical object .
hipparcos ( 1997 ) uses the phase - shift measurement of the temporal evolution of the photometric level of two stars seen drifting through a grid . for a two - aperture interferometer
( see figure 7 ) the external optical delay @xmath371 , while an object with an angle @xmath372 is observed in a broad spectral range ( i.e. white light ) , is : @xmath373 this delay can be determined from the position of the optical delay - line of the instrument set up such that the central fringe of the interference pattern appears in a narrow observation window .
the position , as well as @xmath374 , are measured by laser metrology .
hence , @xmath372 is deduced with a high precision . for ground - based interferometers ,
the baseline is fixed to the earth and will rotate with the earth , while in space the interferometer must reorient the baseline to measure both angular coordinates . for a space - borne interferometer ,
the issue is to find a reference for the angle measured .
usually , a grid of far objects like quasars are used as a reference frame .
there are two modes of observation possible : the ` wide - angle ' and the ` narrow - angle ' modes . in wide - angle mode
, the large angle difference between the reference and the studied object usually requires collector motions . in narrow - angle , the two objects are in the field of view of the instrument ,
therefore , no motions are required and the accuracy of the measurement is improved . however , it is difficult to have always a correct reference star within the field of view for any studied object .
narrow - angle astrometry is , therefore , more suitable for wobble characterization .
nulling interferometry was first proposed by bracewell ( 1978 ) for applications in radio astronomy .
such an interferometer could be employed at the upcoming large interferometers to observe faint structures close to non - obscured central sources .
this technique can also be used in space to search for extra - solar earth - like planets through their thermal emission and to determine the atmospheric signature of life with spectroscopic analysis ( angel et al .
1986 , hinz et al .
1998 ) . here
the light collected by two apertures is combined to generate a deep destructive interference fringe at the stellar position , thus selectively nulling the star by many orders of magnitude related to the surrounding off - axis environs , such as a planetary system .
basically , a @xmath375 phase - shift is introduced in one wavefront segment , so that when it interferes with another segment of the same wavefront and perfect cancellation is achieved .
therefore , the central fringe of the interference pattern is dark , allowing the fringe pattern from a faint object to appear . unlike a coronagraph , where useful imaging is possible beyond several airy radii from the on - axis stellar source
, this method is expected to be effective within the core of a telescope s psf , and so can be employed for stars at greater distances , where a larger sample is available ( serabyn , 2000 ) .
the quality of a nulling is defined by the ` null depth ' @xmath376 : @xmath377 where @xmath378 is the phase error between the two recombined beams , @xmath132 the fringe visibility modulus , and @xmath379 the standard deviation of the optical path difference ( opd ) between the two beams . to create the @xmath375 phase - shift , a few important techniques , such as
( i ) using roof reflectors to achieve a reversal of sign of the electric vector and ( ii ) introducing a precise thickness of glass whose index acts to retard all wavelengths by very nearly one - half wavelength may be used .
serabyn ( 2000 ) reported satisfactory nulling results obtained with fiber - coupled rotational shearing interferometer in visible wavelength .
hinz et al . (
1998 ) have demonstrated the viability of nulling interferometry using two 1.8 m mirrors of the original six - mirror mmt .
they have detected the thermal image of the surrounding , circumstellar dust nebula around @xmath0 ori .
hinz et al . (
2001 ) have measured the spatial extent of the mid - ir emission for a few herbig ae stars as well .
figure 8 depicts the schematic of nulling interferometer at the mmt ( hinz et al .
1998 ) .
atmospheric seeing affects the measurements of fringe visibility by introducing phase aberrations across the wavefronts incident on the interferometer , the relative phase of the wavefronts at the apertures changes with time , and also varies the optical paths through the arms .
fringes are to be obtained in a time as short as atmospheric fluctuations ( @xmath1940.01 s ) . the optical interferometers require accurate alignments , high stability , full control of any effect decreasing visibility .
these limitations come from the parameters , viz . ,
( i ) precise determination of visibility ,
( ii ) sensitivity in measuring weak sources , ( iii ) accurate measurement of fringe phases , and ( iv ) availability of range of baselines
. the s / n ratio with which the visibility can be measured is a function of @xmath380 .
the fringe s / n ratio is given by the expression ( lawson , 1995 ) , @xmath381 where @xmath382 is the number of photons detected per sub - aperture per integration time .
the dependence on @xmath380 implies that interferometry becomes increasingly difficult for faint sources , particularly for those with complex structures .
the @xmath380 limit can be addressed in various ways , viz . ,
( i ) using larger sub - apertures , ( ii ) slicing of the image at the entrance of the spectrograph , ( iii ) bootstrapping , and ( iv ) tracking fringes on a point source to increase integration time on the target .
the real art of developing interferometers is to combine the beams in phase with each other after they have traversed exactly the same optical path from the source through each telescopes down to the beam combination point .
the pathlengths of the two arms need to be equalized and maintained to a fraction of @xmath383 .
a correct determination of @xmath384 that is unstable over time is also necessary ; the situation becomes complicated in the presence of atmosphere .
the paths are made equal by adjusting the position of mirrors in the optical delay - line that corrects the drift induced by the diurnal rotation of the tracked star .
of course , the difficulty comes from avoiding various aberrations and vignetting , particularly when light is fed through long and narrow pipes .
the optical delay in terms of telescope and source parameters translates into ; @xmath385 until recently , the beam - recombining optical devices at both i2 t and grand interfromtre deux tlescopes ( gi2 t ; labeyrie et al .
1986 ) were kept on a computer controlled motor driven carriage parallel to the baseline in order to compensate the opd .
of late , a delay - line that is movable is inserted in one of the arms of the i2 t , by means of a cat s eye system ; the other arm was equipped with a fixed delay - line ( robbe et al .
the present recombiner , recombineur pour grand interfromtre ( regain ) at gi2 t also uses a delay - line featuring a cat s eye reflector with variable curvature mirror .
a few interferometers , viz . , mark iii stellar interferometer ( shao et al .
1988 ) , us navy prototype optical interferometer ( npoi ; armstrong et al .
1998 ) , infrared optical telescope array ( iota ; carleton et al .
1994 ) use the vacuum delay - lines . mark
iii interferometer delay lines used laser interferometers to measure the position of the delay line carts and nested servo loops for fine control of the opd .
delay - lines of iota , have a 28 m travel ` long ' delay - line which is moved each time a new object is observed , when the opd to compensate is very different .
it does not move during fringe acquisition .
iota has a second ` short ' delay - line of about 2 m travel which tracks the sidereal motion during fringe acquisition .
both the delay - lines use a dihedral ( two plane mirrors at 90@xmath386 mounted on a carriage . for the long delay - line ,
the carriage is moved by a pulley - and - cable system powered by a stepper motor , while for the short delay - line a linear motor system allows precise motion of the carriage ( 10 nm steps ) . for both delay - lines , measurement of the carriage position
is done by a laser metrology system using the doppler - fizeau effect of a laser beam sent to the carriage and bouncing back ( morel , 2000 ) .
another method known as group - delay tracking , based on the integration of the moduli of all the computed fts ( lawson , 1994 , lawson et al .
1998 ) , is used with the interferometers susi , and coast .
the group - delay is proportional to the rate of change of phase as a function of wavenumber , evaluated at the center of the band .
this delay can be measured if the combined beams from an interferometer are dispersed in a spectrometer .
the group - delay tracking yields a peak whose position is proportional to the opd .
fringes can be obtained either by utilizing the concept of merging speckles ( labeyrie , 1975 ) or by employing pupil - plane configuration ( tango and twiss , 1980 ) . at the initial stage , the beam recombining optics that were employed at both i2 t and gi2 t consisted of ( i ) a recombining element for reconfiguring the pupil and fixing the fringe spacing , ( ii ) an image slicer ( a series of 10 wedges of different angles cemented on a field lens slices the image ) , ( iii ) the compensating system of the atmospheric dispersion , ( iv ) gratings ( maximum spectral resolution 0.15 nm ) , and ( v ) detectors to record the fringes .
the advantages of the dispersion mode are the capabilities ( i ) of allowing continuous observation of fringes across the spectral bandwidth , ( ii ) of recording the fringes with longer integration time , and ( iii ) of selecting different spectral channels for differential visibility measurements . in the regain combiner
, each coud beam coming from the telescopes meets a pupil stabilizer , a field rotator , a wedge prism and the beam combiner ( rousselet - perraut et al . 1996 ) .
figure 9 depicts the process performed by an arm of the regain table prior to recombination .
interferometry with multiple apertures , such as coast uses multi - stage four - way combiner .
light from the telescopes is first combined in pairs .
these pairs are recombined with other pairs .
each detector sees light from all of the telescopes .
another method known as pair - wise recombination technique is employed at npoi , where a beam combiner uses a different detector for each baseline .
[ [ fiber - linked - recombination ] ] fiber - linked recombination + + + + + + + + + + + + + + + + + + + + + + + + + + fiber optics systems provide a perfect spatial filtering of the turbulence - induced corrugated wavefronts and find that the contribution variations are much reduced .
the advantages of this may be envisaged in the form of ( i ) selecting the plane - wave part of a wavefront , ( ii ) splitting a guided wave into any desired intensity ratio , and ( iii ) combining two guided waves interferometrically .
the significance of the fiber - linked unit for optical recombination ( fluor ) , iota is that visibilities can be calibrated with sub-1% precision , which is important for many astrophysical applications like stellar atmospheres studies , cepheid pulsation measurements for distance determinations , and detection of angular anisotropies arising in disks around young stars .
the fiber delay - lines , using spools of fiber that can be stretched or relaxed to increase or decrease the optical pathlength are used in fluor .
these fibers have been designed to propagate infrared light at near - ir ( mennesson et al .
( 1999 ) in tem mode , like a coaxial cable .
therefore , only plane - waves perpendicular to the axis of the fiber may propagate over long distance .
basically , this method results in a ` spatial filtering ' , thus smoothing the wavefront .
the advantage of such a method for interferometry is a reduction of the uncertainty on the measured visibility , and the drawbacks are the loss of optical coupling efficiency and larger photometric variations due to the turbulence .
figure 10 depicts the schematic of the fluor recombiner .
[ [ integrated - optics ] ] integrated optics + + + + + + + + + + + + + + + + + the integrated optics ( io ) , analogous to integrated chips in micro - electronics , potentially allows large tables of bulk optics to be replaced by miniature devices ( haguenauer et al .
such hardware provides easy access to spatial filtering and photometric calibration .
the astronomical validation of this approach to interferometrically combine beams from separated telescopes at iota has been reported recently by berger et al .
two different chips designed for two - telescope beam combination in h band were used by them .
one is manufactured using the ion exchange process : na@xmath387 ions from a glass substrate are exchanged with ag@xmath387 ions in a molten salt through a dedicated mask and the other is manufactured using silica etching technique .
[ [ polarization ] ] polarization + + + + + + + + + + + + for phased combination either in the image - plane or in the pupil - plane , the individual incoming beams from the arms of an interferometer should have identical pupil orientations , image orientations , and polarization characteristics ( traub , 2000 ) .
if the corresponding mirrors at each reflection ( in the case of two telescope interferometer ) are of same type , both the beams will experience the same phase - shifts ; the respective @xmath323 and @xmath324 components combine independently in the focal - plane and produce identical fringe packets . if the sequence of reflections is different , the visibility @xmath388 , of an interferogram is , @xmath389 where @xmath390 is the @xmath391 shift between the two beams .
the loss of coherence due to misalignment of optical train , aging of coatings , and accumulation of dust can also be analyzed ( elias , 2001 ) . at gi2 t
, the regain uses the field rotators consisting of four plane mirrors for each beam to compensate the polarization difference . in the coast ,
starlight passing through
the central siderostat undergoes an additional two reflections so that its @xmath323 and @xmath324 polarizations experience the same reflections as light from the other siderostats ( baldwin et al .
1998 ) .
fibers have a natural birefringence that introduce elliptic polarization at the output , when linearly - polarized light is injected which causes a loss of the measured visibility . a solution for compensating the fiber birefringence
consists in winding the fibers into one or two loops ( lefvre 1980 ) .
the supplementary birefringence introduced by this system depends on the radius of the loops ; it cancels the effects of the natural birefringence . with this system
, the polarization plane can be rotated by twisting the fiber ( by displacing the loops around the main fiber axis ) .
another solution to minimize birefringence effects consists in using a babinet compensator ( a birefringent quartz crystal consisting of two thin prisms cemented together to form a thin parallel plate ) at the input of each fiber .
[ [ dispersion ] ] dispersion + + + + + + + + + + a compensating system for correcting atmospherically induced dispersion is essential at the recombiner . in the regain , the different ` chromatic dispersions ' between the two beams
are compensated by using two prisms which can slide on their hypotenuse , forming therefore , a plate with adjustable thickness .
this thickness is modified every 4 minutes , following the variation of the altitude of the observed object ( rousselet - perraut et al .
1996 ) . in the case of fiber - linked recombiner
, the dispersion of a fiber optics coupler made by two fibers of 1 and 2 is expressed by the phase curvature : @xmath392 where @xmath262 is the phase of the spectrum of the interferogram .
it is demonstrated by coud du foresto et al .
( 1995 ) that the phase curvature can be given by : @xmath393 where @xmath394 is the dispersion coefficient of fiber 2 , @xmath395 the difference of length between the two fibers , and @xmath396 the difference of dispersion .
the dispersion coefficient depends on the refractive indexes of both the core and the cladding of the fiber .
one problem of fiber dispersion is the ` flattening ' of the interferogram , reducing the fringe contrast . to minimize the dispersion
, the length of each fiber must be calculated from the dispersion coefficients of each fiber . due to the atmospheric turbulence affecting the wavefronts before recombination , measurements of @xmath132 are biased by a random factor depending on the seeing quality .
instrumental flaws leading to optical aberrations and non - balanced flux between the two beams modify the measured visibility modulus as well .
it is , therefore , important to calibrate each measure on an object by measuring @xmath132 on an non - variable unresolved source ( e.g. , a farther star ) in the neighborhood , preferably within 1@xmath289 of the program star to minimize motions of telescopes and delay - lines , and at the same turbulence condition .
the calibrator and the studied object observations should be interleaved for recording the fringes on both back and forth a few times during the observing run .
hence , one can interpolate the transfer function for each object - observation period . to reproduce the instrumental conditions ,
the calibrator must roughly be as bright as the object to calibrate ; it should have ideally a spectral type and a magnitude similar to the studied object .
the calibration of the resulting visibility is given by , @xmath397 ( berio , mourard et al .
1999 ) .
fringes are searched by adjusting the delay - line position ; however , mechanical constraints on the instrument , errors on the pointing model , thermal drifts , various vibrations and atmospheric turbulence make the null - opd point changing .
the error on the opd must be less than the coherence length defined by : @xmath398 where @xmath399 is the mean wavelength observed and @xmath400 is the spectral interval .
this real - time control is called ` fringe - tracking ' .
there exist three possible set - ups for fringe acquisition for visible spectrum . in the first one , the opd is temporally modulated by a sawtooth signal , using a fast and short - travel delaying device .
the intensity of the recombined beams , describes therefore , over the time a fringe pattern that is recorded by single pixel detectors .
the second method consists of imaging the dispersed recombined beam on a linear detector . in the third one , beams are dispersed prior to the recombination ; recombination is done by focusing them with a common lens like in michelson stellar interferometer . in ir , it is essential to use as little pixel as possible in order to reduce global readout noise .
the ` white ' fringes set - up is preferably used for ir observations . [ [ coherencing - and - cophasing ] ] coherencing and cophasing + + + + + + + + + + + + + + + + + + + + + + + + + techniques to compensate the opd drift between the two beams of a standard optical interferometer may be classified into two categories : coherencing and cophasing .
the aim of the former is to keep the opd within the coherence area of the fringes , while the latter , the opd must remain much smaller than the wavelength : fast compensation of the opd variations due to the differential ` piston ' mode of the turbulence is , therefore , done in order to ` freeze ' the fringes .
the coherencing with non - white fringes is compared with the active optics . in white light
it is done , as for iota , by scanning the opd while acquiring signals to find the null - opd point in the fringe pattern ( morel et al .
this coherencing yields the opd correction to apply to the delay - line , at a few hz servo - loop rate . with a channeled spectrum ,
the opd is proportional to the fringe frequency .
another method , viz .
, the real time active fringe - tracking ( raft ) system has been applied to dispersed fringes on gi2 t using 2-d ft ( koechlin et al . 1996 ) .
both group - delay tracking and raft allow a slow servo - loop period ( up to a few seconds ) by multiplying the coherence length by the number of spectral channels used .
cophasing that may be compared with the ao system , is performed with white fringes , using the synchronous detection method ; the opd is quickly scanned over a wavelength range .
the signal acquired from the detector is then processed in order to yield the phase - shift and the visibility , which can be done by integrating signal ( shao and staelin , 1977 ) over four @xmath401 bins , named @xmath402 , @xmath403 , @xmath404 and @xmath65 .
phase - shift and visibility modulus are then given by , @xmath405 fringe - tracking methods may be enhanced by introduction of _ a priori _ information , in order to allow observations at fainter @xmath132 or fainter magnitudes . gorham ( 1998 )
has proposed to improve white light cophasing by filtering data with a function computed to reduce the photon noise .
the gain for the tracking limit magnitude , at constant @xmath132 , is between 0.5 and 0.7 .
methods introducing _ a priori _ information for gdt or raft have been imagined as well ( padilla et al . 1998 , morel and koechlin , 1998 ) . [ [ bootstrapping ] ] bootstrapping + + + + + + + + + + + + + it is possible to stabilize the fringes using bootstrapping methods at the interferometers .
these methods can be categorized into two systems , namely , wavelength bootstrapping and baseline bootstrapping .
it may be optimal to employ the former that requires two recombiners , one for visibility measurement , the other one for fringe - tracking .
for example , at long baselines , when the expected fringe visibility is too low for tracking , it is possible to use a longer wavelength where the fringe contrast , for a white object , is higher .
meanwhile , fringes for computing the visibility are acquired at shorter wavelength than for tracking .
this method was applied at mark iii interferometer ( quirrenbach et al .
1996 ) . on the other hand , the latter is performed by tracking fringes over a connected series of short baselines to allow low visibility fringes to be measured on the longest baseline . in this method
the photons are shared and it works well because of the closure - phase relationships . here
the baseline is divided into sub - baselines by adding apertures along the baseline .
fringe - tracking is performed on each sub - baseline , where the visibility is higher than with the entire baseline .
hence , fringes are tracked on the whole baseline as well .
figure 11 depicts the principle of baseline bootstrapping .
npoi employs such a system that enables one to reach spatial frequencies beyond the first visibility null ( pauls et al . 1998 ) .
a related idea has been developed for the keck interferometer , as well as for the eso s very large telescope interferometer ( vlti ) .
[ [ role - of - ao - systems ] ] role of ao systems + + + + + + + + + + + + + + + + + + an interferometer works well if the wavefronts from the individual telescopes are coherent .
the maximum useful aperture area is proportional to @xmath406 . in order to improve the sensitivity of an interferometer where it is used to target faint sources , each telescope will have to be reasonably large .
a large aperture produces more than 100 speckles in the image and the fringe pattern within each speckle is randomly phased .
enough photons are required to phase the telescopes into a coherent aperture , therefore , to enhance the instrumental visibility , an ao system should be incorporated .
a typical image - width reduction of roughly a factor of 10 and a central intensity enhancement of a factor of 10@xmath407 can be achieved on large telescopes ( traub , 2000 ) . though complete ao systems have not been implemented at any of the interferometers to date , baring keck - i which uses full - ao corrections , a few of them , viz . ,
gi2 t , iota are using tip - tilt control system .
this correction is sensed in the visible , using ccds , and fringe detection is done in the near - ir .
iota tip - tilt correction system uses a 32@xmath34832 pixel ccd for each beam .
the maximum rate is 200 hz ( morel , 2000 ) .
a computer reads each frame and computes the centroid .
the value of the centroid position is sent to a piezo mirror placed downstream the secondary mirror of each telescope .
the optimal integration time required for measuring a visibility point is a trade - off between the number of photons to collect and the earth rotation shifting the sampled point in the @xmath363 plane .
though with the large binocular telescope ( lbt ) , the two 8.4 m mirrors will be installed on a common alt - az mounting , thus information in the @xmath408 plane can be continuously combined or coadded , most of the interferometers use two apertures and are unable to recover the complex visibility .
therefore , the information to extract from a batch of fringes is the modulus of the visibility .
theoretically , using merely the fourier transform would give an optimal estimate of the visibility modulus , as demonstrated by walkup and goodman ( 1973 ) .
however , white - light fringes obtained from coherencing are flawed by the differential piston that modulates their frequency .
techniques used in radio - interferometry ( where wavelengths are much longer ) , like fitting a sinewave through the fringe data , are therefore , not suitable .
perrin ( 1997 ) has proposed a method to remove the piston from fringes . however
, this method requires a high fringe snr ratio and may only be applied when fringe snr is important . the model visibility amplitude @xmath409 for a uniform source of diameter @xmath410 is given by , @xmath411 d\theta,\ ] ] where @xmath412 and @xmath413 are the respective total flux and the brightness distribution of the source , @xmath372 the position angle of the source and @xmath414 the baseline length in wavelengths .
this equation reduces to , @xmath415 } { \pi s_\lambda \phi_{ud}}\ ] ] for the objects with circular symmetry , the visibility function is expressed as , @xmath416 where @xmath417 is a bessel function of the first kind .
however , a problem arises due to the limb - darkening of the stars .
observations of limb darkening measurements require one to collect data in the vicinity of and beyond the first zero or minimum of the visibility function .
the radial intensity profile of a star may be given ( hestroffer , 1997 ) by : @xmath418 where @xmath419 is the disk brightness at angle @xmath420 , @xmath372 the angle between the normal to the stellar surface and the direction to the observer , and @xmath421 is the limb - darkening factor depending on the stellar atmosphere . the visibility function in this case may be derived as , @xmath422 where @xmath423 and @xmath424 is the limb - darkened diameter of the source .
many interferometers can not measure low visibilities existing at high angular frequency ( i.e , when @xmath425 is large ) , beyond the first minimum of the visibility function .
reconstructions are , therefore , ambiguous and neither the diameter nor the limb - darkening factor may be accurately determined . usually , @xmath421 is an a priori information given by the stellar atmosphere model .
the diameter is , therefore , deduced from @xmath421 and the interferometric data .
figure 12 represents two visibility curves of the mira type variable star r leo , obtained by the fluor / iota combination in the k band at two different epochs , which show both the change in equivalent ud disk diameter and ( for the 1997 data ) the diffusion by circumstellar material whose signature is a visibility curve substantially different from that of a uniform disk ( perrin et al . 1999 ) . for two unresolved sources , e.g. , a binary system ,
the expression is formulated to , @xmath426 where @xmath427 is the angular separation between two sources and @xmath428 , @xmath429 the brightnesses of the source 1 and source 2 , respectively .
the visibility modulus corresponding to this function at @xmath47 is , therefore : @xmath430 it is then useful to use a technique called ` super - synthesis ' : the @xmath363 plane is swept during an observation lasting several hours , due to earth rotation . after a large variation of hour angle @xmath365 , several visibility moduli are , therefore , measured at different @xmath363 points to determine the parameters ( @xmath428 , @xmath429 and @xmath431 ) of the system by fitting the function described by the equation ( 117 ) .
the visibility function of the thin structures of the circumstellar shell may be computed by considering a coaxial uniform disk and a point source ; therefore , the function is written as , @xmath432 where @xmath433 is the diameter of the shell and @xmath434 the ratio of power radiated by the star . combining photometry with the measurement of limb - darkened stellar diameter yield the stellar emergent flux , @xmath435 or surface brightness and
is found from the relation , @xmath436 where @xmath437 is the measured absolute monochromatic flux received from the star at frequency @xmath438 .
the stellar effective temperature @xmath439 is defined in terms of the emergent flux by the stefan - boltzmann law .
integrating over all frequencies , @xmath440 where @xmath441 is stefan constant .
several ground - based lbois have been developed to obtain very high angular resolution informations of the stellar objects .
however , three of them , viz .
, ( i ) i2 t ( labeyrie , 1975 ) , plateau de calern , france , ( ii ) mark iii interferometer ( shao et al . , 1988 ) , mt .
wilson , usa , and ( iii ) infrared interferometer , soirdt ( gay and mekarnia , 1988 ) , plateau de calern , france are no longer in operation .
i. labeyrie et al .
( 1986 ) have developed gi2 t interferometer with two ` boule ' telescopes that run on north - south tracks with variable baseline of 12 to 65 m. this instrument at plateau de calern , france combines the features of the michelson design and the radio interferometers ; it operates in speckle mode .
it consists of a pair of 1.52 m telescopes on altitude - altitude mounts .
each telescope is housed in a sphere ( 3.5 m diameter ) made of concrete , which has three mirrors directing the horizontal afocal coud beam to the recombiner optics .
light beams from both the telescopes are superposed at the foci in order to produce young s fringes .
the driving system of the sphere consists of a pair of rings ; each ring is motorized by 3 actuators acting in 3 orthogonal directions within 2 different tangential planes .
the two rings alternately carry the sphere , which in turn , produce continuous motion with a resolution of the order of 1 @xmath344 m .
the main drawback comes from the slow pointing of the telescopes ; 4 or 5 stars can be tracked during the night .
the susi , sydney university , australia has a very long baseline ranging from 5 m to 640 m ( north - south ) that are achieved with an array of 11 input stations equipped with a siderostat and relay optics , located to give a minimal baseline redundancy ( davis et al .
1999a ) ; the intermediate baseline forms a geometric progression increasing in steps of @xmath19440% .
starlight is steered by two siderostats of 20 cm diameter into the evacuated pipe system that carries the light to the atmospheric refraction corrector ( at the central laboratory ) consisting of the pairs of counter - rotating risley prisms via a beam reducer .
it proceeds towards either the optical path length compensator ( oplc ) or is diverted towards the acquisition camera . on leaving oplc , the beams from the two arms of this interferometer may be switched over to one of the optical tables ( blue or red ) for recombination .
the coast , cambridge , uk uses four independent telescopes consisting of 50 cm siderostat flat feeding a fixed horizontal 40 cm cassegrain telescope with a magnification of 16 .
these are arranged in a y - layout with one telescope on each arm , movable to a number of fixed stations and one telescope at the center of the y ( baldwin et al .
light from each siderostat passes through pipes containing air at ambient pressure into the beam combining laboratory ( inside a tunnel ) .
the four beams emerging from the path compensator are each split at a dichroic ; the longer wavelength ( @xmath442650 nm ) of the visible band passes into the beam combining optics and the shorter ones are used for acquisition and autoguiding .
each output beam passes through an iris diaphragm and is focused by a long focus lens on to a fiber fed single - element avalanche photodiode detector for fringe detection .
the iota , situated at mt .
hopkins , arizona consists of three 45 cm collector assembly located at various stations on the l - shaped baseline ( 5 - 38 m ) that comprises a siderostat , an afocal cassegrain telescope and an active relay mirror ( traub et al . 2000 ) . in the recombining table , the optical differences are compensated by fixed and variable delays and the beams are recombined onto a beam splitter , producing two complementary interference signals ( carleton et al .
a fast autoguiding system is used to correct the atmospheric wavefront tilt errors .
two active delay - lines for three telescopes are provided ; a scanning piezo mirror is used to modulate the opd between the two telescopes . v. the npoi , located at lowell observatory , arizona , is designed to measure positions with precision comparable to that of hipparcos ( 1997 ) .
however , the aperture sizes limit the array to bright star astrometry . it is developed as y - shaped baseline configuration that includes sub - arrays for imaging , and for astrometry .
for the astrometric mode , four fixed siderostats ( 0.4 m diameter ) are used with the baselines extendable from 19 m to 38 m. the astrometric sub - array has a laser metrology system to measure the motions of the siderostats with respect to one another and to the bedrock . while for the imaging mode , 6 transportable siderostats ( 0.12 m diameter ) are used .
three siderostat positions are kept with equal space for each arm of the y. coherence of imaging configuration is maintained by phase bootstrapping .
the synchronous detection method applied to signals from several spectral channels , has been used as well ( benson et al . 1998 ) .
the palomar testbed interferometer ( pti ) , palomar observatory , california is an ir phase - tracking interferometer that was developed as a test - bench for the keck interferometer .
the main thrust is to develop techniques and methodologies for doing narrow angle astrometry ; therefore , it is designed to observe two stars simultaneously to measure the angle between them with high precision .
it uses coherent fringe demodulation and active fringe - tracking systems with an array detector at 2.2 @xmath344 m and active delay - lines with a range of @xmath44338 m. it is comprised of three 40 cm siderostats that are coupled to beam compressors .
these siderostats are used pairwise to provide baselines up to 110 m ( colavita et al . 1999 ) .
both phase and group - delay measurements for narrow angle astrometry are being carried out ( lawson et al . 2000 ) .
visibility is estimated from the fringe quadrature , either incoherently , or using source phase referencing to provide longer integration time ( colavita , 1999 ) .
heterodyne technique is generally used in radio - astronomy to reduce the high frequency signal to an intermediate one .
such a technique has the following advantages in the case of beam recombination in ir interferometry .
they are : a larger coherence length , a simplification of the transport of the signal from the collector to the recombiner ( coaxial cables instead of mirrors ) .
let @xmath444 and @xmath445 , be respectively the signals of a wave coming from a star and of an artificial source ( laser ) , which are expressed as , @xmath446 } , \\
{ \cal u}_{\it l}(t ) & = & a_{\it l0}.e^{-i\omega_{\it l}.t}.\end{aligned}\ ] ] the laser is the phase reference .
a detector like a photodiode , illuminated by the sources ( star + laser ) yields an electrical signal corresponding to the light intensity : @xmath447}+{\it a}_{\it l0}. e^{-i\omega_{\it l}.t } ) \nonumber\\ & & \times({\it a}_{\it s0}.e^{i[\omega_{\it s}.t+\psi]}+{\it a}_{\it l0}. e^{-i\omega_{\it l}.t } ) \nonumber\\ & & = { \it a}_{\it s0}^2+{\it a}_{\it l0}^2 + 2.{\it a}_{\it l0}.{\it a}_{\it s0}. \cos[(\omega_l-\omega_s)+\psi].\end{aligned}\ ] ] if @xmath448 and @xmath449 are close , the frequency of @xmath23 is low enough to fit in the bandwidth of the detector and its electronics ( a few ghz ) and @xmath23 carries the phase information from the radiation of the star . by correlating ( multiplying ) the signals @xmath450 and @xmath451 yielded by two apertures with heterodyning systems , one can extract a visibility term
. however , the lasers must have the same phases for the two apertures .
the heterodyne interferometry at 8 - 11.5 @xmath344 m spectral range was employed on the soirdt , observatoire de calern , france , an ir interferometer that consisted of a pair of 1 m telescopes with a 15 m east - west horizontal baseline ( gay and mekarnia , 1988 ) .
it was an interesting project but ultimately a failure and produced no significant scientific results .
in this project , beams were received in the central laboratory on a double cat s eye delay - line on a step by step movable carriage ; natural opd drift due to the earth - rotation was used for acquiring fringes ( rabbia et al .
heterodyne interferometry is also used on the recently developed mid - ir spatial interferometer ( isi ) , mt .
wilson , california ( townes et al . 1998 ) , which is an outstanding success with many scientific results .
it is well suited to study of circumstellar material around bright evolved stars .
it features three movable telescopes ; each telescope comprises a 1.65 m parabolic mirror and a 2 m flat mirror equipped with an automated guiding and tip - tilt control system at 2 @xmath344 m ( lipman et al .
the star light from each aperture is first mixed with a stable @xmath452 laser local oscillator , converting the signal to microwave frequencies , followed by pathlength matching and fringe detection in a correlator . on isi ,
two @xmath452 lasers are used , the phase of one being controlled by the other . here the interferometer noise is dominated by shotnoise of the laser and thermal background is negligible for setting the sensitivity limit ( hale et al .
isi utilizes earth rotation and periodic discrete changes of the baseline to obtain a wide range of effective baselines and map the visibility functions of the stellar objects .
i. the center for high angular resolution astronomy ( chara ) array at mt wilson , california comprises six fixed 1 m telescopes arranged in a y - shaped configuration with a maximum baseline of @xmath194350 m that will operate at optical and ir wavelengths ( mcalister et al . 1998 ) with a limiting resolution of 0.2 mas .
the key scientific goal of this interferometer is binary star astrometry , observations of stars with well - determined spectroscopic elements , and determination of the metal abundance .
\ii . the mitaka optical - ir array , national astronomical observatory , japan consists of several interferometers built one - by - one .
the first of the series was mira - i ( machida et al . 1998 ) that has 25 cm siderostats and a 4 m baseline .
its successor , mira - i.2 ( sato et al .
1998 ) has the same baseline and slightly larger siderostats ( 30 cm ) .
it features equipment encountered on many operating interferometers : beam compressors ( yielding 30 mm beams ) , delay - line operating in vacuum , tip - tilt correction system and laser metrology .
these instruments are specially designed for astrometry .
\iii . the large binocular telescope ( lbt ) , on mt .
graham , arizona that consists of two 8.4 m primary mirrors ( hill , 2000 ) is under construction , the lbt mirrors are co - mounted on a fully steerable alt - az mounting , in which variable delay lines for the path equalization are not needed . at near - ir wavelengths ,
a field of view of one arcminute or more is expected with unprecedented spatial resolution of order of 8 - 9 mas at @xmath453 m and a variable baseline of 0 - 23 m ( wehinger , 2001 ) .
another interferometer , the magdalene ridge observatory array , usa with three element ( [email protected] m plus 0.8 m ) has also received initial funding and is under development . the keck interferometer , mauna kea , usa ( colavita et al .
1998 ) , and ( ii ) vlti , eso , at paranal , chile ( derie et al .
2000 ) are of a heterogenous nature .
the problem comes from the recombination of two telescopes , one large and one small because the s / n ratio would be the one given by the small telescopes .
nevertheless , the recent success of obtaining interferometric fringes from the starlight by the two large telescopes at both keck ( http://www.jpl.nasa.gov/news ) and vlti ( glindemann and paresce , 2001 ) will have an enormous impact on developing future large optical arrays .
the keck interferometer comprises of two 10 m and four 1.8 m ` outrigger ' telescopes ; for imaging the main telescopes would be used with outriggers to fill in incomplete parts of the @xmath408 plane .
it will combine phased pupils provided by adaptive - optics for the main telescopes and fast tip / tilt correction on the outriggers .
beam recombination will be carried out by 5 two - way combiners at 1.5 - 2.4 @xmath344 m for fringe - tracking , astrometry , and imaging .
a project for a 10 @xmath344 m nulling - combiner for exo - zodiacal disk characterization is in the planning stages .
a differential phase technique to aim at detecting faint sources near a bright object is also under development ( akeson et al . 2000 ) . while in the case of vlti
, beams are received from the movable telescopes in a central laboratory for recombination and are made to interfere after introducing suitable optical delay - lines .
coud beams from these apertures are sent through delay - lines operating in rooms at atmospheric pressure but at accurately controlled temperature .
the beams reach an optical switch - yard to be directed to one of the four expected recombiners , e.g. , ( i ) a single mode fiber recombiner ( 2.2 @xmath344 m ) that intends to debug the upstream sub - system of vlti , ( ii ) a beam splitter based recombiner that operates at 10 @xmath344 m , ( iii ) a recombiner that operates between 1 @xmath344 m - 2.5 @xmath344 m , and ( iv ) a recombiner for narrow angle astrometry .
the next generation imaging interferometers with at least 15 or more elements should have the snapshot capability in an instantaneous mode .
beams from separated telescopes of such an interferometer must be recombined in the focal point as in the case of a fizeau interferometer that is optically equivalent to single large telescope masked with a multi - aperture screen so as to reproduce exactly the ensemble of collecting telescopes ( traub , 1986 ) .
development of an optical very large array ( ovla ) , an array of 27 telescopes of 1.5 m diameter was initiated more than a decade ago by labeyrie , lamaitre et al .
each telescope is housed in a fiberglass sphere ( dejonghe et al .
1998 ) that is mounted on a six - legged robot to enable it to move on the ground while fringes are acquired , eliminating the need for optical delay - lines .
a secondary mirror makes the beam afocal and compressed .
a third steerable flat mirror sends this beam out through a slit located on the sphere to the central station where all other beams coming from other telescopes are combined into a single high resolution image .
ovla has also been considered for different possible aperture diameters including 12 to 25 m ( labeyrie , 1998 ) .
a new telescope structure has been imagined for this class of very large collectors : the ` cage telescope ' , in which the sphere is replaced with an icosahedral truss steerable by a different mechanical system .
ridgway and roddier ( 2000 ) have proposed a project of developing ir_vla with a total baseline of @xmath1941000 m , consisting of 27 telescopes with each aperture of 4 m. another proposed project called ` large - aperture mirror array ' ( lama ) employs eighteen fixed 10 m liquid - mirror telescopes located within a circle of 60 m diameter which collect 50% of the light that falls within this area ( hickson , 2001 ) .
the ` densified - pupil multi - aperture imaging interferometry ' may provide direct images at their focal - plane , albeit it is a subtle technique .
conceptually it differs from the fizeau interferometers which become inefficient when the sub - aperture spacing is large compared to their size .
the reason is that most energy goes in a broad diffractive halo rather than in a narrow interference peak , which precludes obtaining usable snapshot images with kilometric or megametric arrays in space .
densifying the exit pupil , i.e. , distorting it to increase the relative size of the sub - pupils , in such a way that the pattern of sub - aperture centers is preserved , concentrates the halo and intensifies the image ( labeyrie , 1996 ) .
figure 13 depicts the concept of the hyper - telescope .
pedretti et al .
( 2000 ) have derived the integrated intensities of the central peaks of the images on the star capella that are obtained by taking two separate exposures of 100 s in the fizeau and densified - pupil mode of the hyper - telescope .
the comparison of these values showed an intensity gain of 24@xmath4433@xmath348 of the densified with respect to the fizeau configuration .
imaging arrays of huge size , possibly approaching a million kilometers to observe neutron stars , may become feasible in this way , called ` hyper - telescopes ' .
they provide an image with full luminosity in a narrow field of @xmath454 , where @xmath455 is the distance between the sub - apertures of the array . in practice for ground arrays ,
an elliptical track is one way of compensating the earth s rotation , but with delay lines , a periodic dilute - aperture can also be built at the scale of 10 kilometers . in space , it would be a periodic hexagonal paving in the case of hyper - telescope version proposed by labeyrie ( 2001 ) for space interferometer , terrestrial planet finder ( tpf ) .
this version has a fizeau focus followed by a small pupil densifier , and a coronagraph ( boccaletti et al . 2000 ) . unlike the ground - based elliptic ring where images are directly obtained at a recombination station located at a focus of the ellipse or periodic arrays on the ground , space arrays can be globally pointed .
the cophasing of the array may be done hierarchically ( pedretti and labeyrie , 1999 ) by cophasing triplets of beams ( yielding a honeycomb pattern in the image - plane ) , then triplets of triplets , etc .
piston errors are measurable from the triplet images . another way of analyzing the piston errors ( labeyrie , 1999a ) is an extension of the classical dispersed - fringes used since michelson : a set of monochromatic images recorded with a spectro - imager is organized as an @xmath456 data cube , and its 3-dimensional ft is calculated to extract piston errors in pairs or triplets of apertures .
the advantage of deploying lbi in space is that of observing can be done at any wavelength and for longer duration in the absence of atmospheric turbulence .
the difficulty comes from developing a technology featuring high precision positioning , as well as toughness required for space operation . a new generation of ultra - lightweight active mirrors ( burge et al .
2000 , angel , 2001 ) are essential to resolve the problems of size and weight . this new generation of scientific spacecraft , st3 ( gorham et al .
1999 ) , nasa s mission , scheduled for 2003 , consists of two independent free - flying elements launched into an earth - trailing heliocentric orbit .
one is a collector sending light from the observed object to the second element featuring another collector , an optical delay - line and a beam recombiner .
the aim of st3 is the demonstration and validation of technologies that might be used for future space - borne interferometers .
thus , the two elements of st3 should be able to move up to 1 km from each other , while being controlled by a laser metrology .
however , the designed delay - line of st3 can be up to 20 m of optical pathlength only .
this instrument will be used as an imaging interferometer for studying such objects as wolf - rayet or be stars ( linfield and gorham , 1999 ) .
space interferometry mission ( sim ) is also being designed and will be launched by nasa .
the main goal of this interferometer will be to collect the new high - precision astrometry results , including the possibility of jovian planet detection around stars up to 1 kilo - parsec distant and terrestrial planet detection around nearby stars ( unwin et al .
the design consists of one free - flyer with a 10 m boom supporting 30 cm collectors .
the expected angular accuracy is 1 @xmath344as in narrow - angle mode ( with a 1@xmath289 field of view ) and 4 @xmath344as in wide - angle mode .
the sensitivity for astrometry is @xmath457 after a four - hour integration .
this interferometer will work in the visible spectrum ( 0.4 to 0.9 @xmath344 m ) . in order to get an accurate knowledge of * b * for wide - angle astrometry without collector motions , it will feature two auxiliary interferometers , aimed at reference stars ( grid - locking ) .
the schematic design of the sim is depicted in figure 14 .
two new projects being pursued , darwin and tpf , are for an ozone search on extra - solar planets ( penny et al .
1998 ) , and for detecting extra - solar planet directly ( beichman , 1998 ) , respectively .
the diffraction - limited phase retrieval of a degraded image is indeed an art . in other branches of physics too , e.g. , electron microscopy , wavefront sensing , and crystallography , one often wishes to recover phase .
ao systems may also require image - processing algorithms since the real - time corrected image is often partial .
prior to using such algorithms , the basic operations to be performed are dead pixel removal , debiasing , flat fielding , sky or background emission subtraction , and suppression of correlated noise . in
what follows , the sub - sections vi - a to c describe methods to obtain components of the object ft and sub - section vi - d describes methods to reconstruct an image from these components , which can usually be described as a ` deconvolution ' . the shift - and - add ( saa ) technique ( lynds et al .
1976 , worden et al . 1976 )
aligns and co - adds the recorded short - exposure images ; the method is analogous to the tip - tilt mirror of conventional ao systems . the position of the brightest pixel , @xmath458 , is necessary to be located in each specklegram , @xmath459 , [ @xmath460 for all @xmath461 , followed by shifting the specklegram ( without any rotation ) to place this pixel at the center of the image space .
the saa image , @xmath462 , is obtained by averaging over the set of the shifted specklegrams , @xmath463 the large variations in the brightness of the brightest pixels are observed in a set of speckle images ; the contamination level may not be proportional to its brightest pixel ( bates and mcdonnell , 1986 ) .
the adjusted saa image , @xmath464 , can be defined as , @xmath465{\cal i}_k({\bf x } + { \bf x}_k)>,\ ] ] where @xmath466 $ ] is the weighting in relation with the brightness of the brightest pixel .
the choice of the same quantity can be made as @xmath467 .
an array of impulse is constructed by putting an impulse at each of the center of gravity with a weight proportional to the speckle intensity .
this impulse array is considered to be an approximation of the instantaneous psf and is cross - correlated with the speckle frame . disregarding the peaks lower than the pre - set threshold , the m@xmath319 speckle mask , @xmath468 ,
is defined by , @xmath469 the @xmath470 masked speckled image , @xmath471 , is expressed as , @xmath472 the lynds - worden - harvey image is obtained by averaging equation ( 127 ) . for direct speckle imaging , the saa image , @xmath462 , is a contaminated one containing two complications - a convolution , @xmath473 and an additive residual , @xmath474 - which means , @xmath475 where @xmath476 being the constant position vectors and @xmath477 , the positive constant .
it is essential to calibrate @xmath462 with an unresolved point source and reduce it in the same way to produce @xmath277 .
the estimate for the object , @xmath232 , is evaluated from the inverse ft of the following equation , @xmath478 which is the first approximation of the object irradiance .
this method is found to be insensitive to the telescope aberrations but sensitive to dominating photon noise .
another method called ` selective image reconstruction ' selects the few sharpest images that are recorded when the atmospheric distortion is naturally at minimum , from a large dataset of short - exposures ( dantowitz et al .
baldwin et al . ( 2001 ) have demonstrated the potential of such a technique .
the knox - thomson ( kt ) method ( knox and thomson , 1974 ) defines the correlation of @xmath273 and @xmath273 multiplied by a complex exponential factor with spatial frequency vector @xmath479 .
the approximate phase - closure is achieved by two vectors ( see figure 15 ) , @xmath47 and @xmath480 , assuming that the pupil phase is constant over @xmath479 .
let the general second - order moment be the cross spectrum , @xmath481 .
it takes significant values only if @xmath482 ; the typical value of @xmath483 is @xmath1940.2 - 0.5 @xmath267 .
invoking equation ( 70 ) , a 2-d irradiance distribution , @xmath273 and its ft , @xmath266 , is defined by the equation , @xmath484 in image space , the correlations of @xmath273 , is derived as , @xmath485 where @xmath486 are 2-d spatial co - ordinate vectors .
the kt correlation may be defined in fourier space as products of , @xmath266 , @xmath487 where @xmath488 , and @xmath489 are 2-d spatial frequency vectors .
@xmath479 is a small , constant offset spatial frequency .
a number of sub - planes is used by taking different values of @xmath479 .
the argument of the equation ( 133 ) provides the phase - difference between the two spatial frequencies separated by @xmath479 and is expressed as , @xmath490 therefore , the equation ( 133 ) translates into , @xmath491}. \end{aligned}\ ] ] the object phase - spectrum of the equation ( 135 ) is encoded in the term , @xmath492}$ ] . in a single image realization , it is corrupted by the random phase - differences due to the atmosphere - telescope otf , @xmath493}$ ] . if equation ( 135 ) is averaged over a large number of frames , the feature @xmath494 .
@xmath495 } \nonumber\\ & & < \widehat{\cal s}({\bf u}_1)\widehat{\cal s}^\ast({\bf u}_1 + { \bf \delta u } ) > , \end{aligned}\ ] ] from which , together with equation ( 73 ) , the object phase - spectrum , @xmath496 , can be determined .
the triple correlation ( tc ) is a generalization of closure phase technique ( section iv.b.2 . )
that is used in radio / optical interferometry where the number of closure phases is small compared to those available from bispectrum ( weigelt , 1977 , lohmann et al .
it is insensitive to ( i ) the atmospherically induced random phase errors , ( ii ) the random motion of the image centroid , and ( iii ) the permanent phase errors introduced by telescope aberrations ; any linear phase term in the object phase cancels out as well .
the images are not required to be shifted to common centroid prior to computing the bispectrum .
the other advantages are : ( i ) it provides information about the object phases with better s / n ratio from a limited number of frames , and ( ii ) it serves as the means of image recovery with diluted coherent arrays ( reinheimer and weigelt , 1987 ) .
the disadvantage of this technique is that it demands severe constraints on the computing facilities with 2-d data since the calculations are four - dimensional ( 4-d ) .
a triple correlation is obtained by multiplying a shifted object , @xmath497 , with the original object , @xmath273 , followed by cross - correlating the result with the original one ( for example , in the case of a close binary star , the shift is equal to the angular separation between the stars , masking one of the two components of each double speckle ) .
the calculation of the ensemble average tc is given by , @xmath498 where @xmath499 are 2-d spatial co - ordinate vectors .
the bispectrum is given by , @xmath500 where @xmath501 , @xmath502 , and @xmath503 .
the object bispectrum is given by , @xmath504 the modulus @xmath505 of the object ft @xmath237 can be evaluated from the object bispectrum @xmath506 . the argument of equation ( 138 ) gives the phase - difference and is expressed as , @xmath507 the object phase - spectrum is encoded in the term @xmath508 .
it is corrupted in a single realization by the random phase - differences due to the atmosphere - telescope otf , @xmath509}$ ] .
if sufficient number of specklegrams are averaged , one can overcome this shortcoming .
let @xmath510 be the phase of the object bispectrum ; then , @xmath511 equations ( 142 ) and ( 143 ) may be inserted into equation ( 140 ) , yielding the relations , @xmath512}\rightarrow , \\
\theta^{tc}_{\cal o}({\bf u}_1 , { \bf u}_2 ) & = & \psi_{\cal o}({\bf u}_1 ) - \psi_{\cal o}({\bf u}_1 + { \bf u}_2 ) + \psi_{\cal o}({\bf u}_2 ) .
\end{aligned}\ ] ] equation ( 145 ) is a recursive one for evaluating the phase of the object ft at coordinate @xmath513 .
the reconstruction of the object phase - spectrum from the phase of the bispectrum is recursive in nature . the object phase - spectrum at @xmath514 can be written as , @xmath515
if the object spectrum at @xmath516 and @xmath517 is known , the object phase - spectrum at @xmath518 can be computed .
the bispectrum phases are mod @xmath269 , therefore , the recursive reconstruction in equation ( 139 ) may lead to @xmath375 phase mismatches between the computed phase - spectrum values along different paths to the same point in frequency space . however , according to northcott et al .
( 1988 ) , phases from different paths to the same point can not be averaged to reduce noise under this condition .
a variation of the nature of computing argument of the term , @xmath519 , is needed to obtain the object phase - spectrum and the equation ( 146 ) translates into , @xmath520}.\ ] ] the values obtained using the unit amplitude phasor recursive re - constructor are insensitive to the @xmath375 phase ambiguities .
saha et al .
( 1999 ) have developed a code based on this re - constructor .
the least - square formulation of the phase reconstruction ( glindemann et al .
1991 ) and the projection - slice theorem of tomography and the radon transform ( northcott et al .
1988 ) have been successfully applied in the development of phase reconstruction from the bispectrum .
most deconvolution techniques , in which ` a priori information ' plays an essential role , can be simplified to the minimization /maximization of a criterion by using an iterative numerical method ( gerchberg and saxton , 1972 ) that bounces back and forth between the image - domain and fourier - domain constraints until two images are found that produce the input image when convolved together ( ayers and dainty , 1988 ) .
let the degraded image , @xmath273 , be used as the operand .
an initial estimate of the psf , @xmath277 , has to be provided .
the image is deconvolved from the guess psf by wiener filtering ( see section iv.a.1 ) , which is an operation of multiplying a suitable wiener filter ( constructed from @xmath521 , of the psf ) with @xmath266 .
the technique of wiener filtering damps the high frequencies and minimizes the mean square error between each estimate and the true spectrum .
this filtered deconvolution takes the form , @xmath522 the wiener filter , @xmath523 , is derived as , @xmath524 the term , @xmath279 can be replaced with a constant estimated as the rms fluctuation of the high frequency region in the spectrum where the object power is negligible .
the wiener filtering spectrum , @xmath525 , takes the form : @xmath526 the result , @xmath237 , is transformed back to image space , the negatives in the image and the positives outside a prescribed domain ( called object support ) are set to zero .
the average of negative intensities within the support are subtracted from all pixels .
the process is repeated until the negative intensities decrease below the noise .
a new estimate of the psf is next obtained by wiener filtering @xmath273 , with a filter constructed from the constrained object , @xmath232 ; this completes one iteration .
this entire process is repeated until the derived values of @xmath232 and @xmath277 converge to sensible solutions .
bid has the ability to retrieve the diffraction - limited image of an object from a single specklegram without the reference star data ( saha and venkatakrishnan , 1997 ) .
jefferies and christou , ( 1993 ) , have developed an algorithm which requires more than a single speckle frame for improving the convergence .
barnaby et al . ( 2000 ) used this algorithm , and parametric blind deconvolution ( pbd ) to post - process the data obtained with the ao system , and found that the secondary of 81 cnc was 0.12 @xmath141 brighter than the primary at 0.85 @xmath344 m .
fienup ( 1978 ) algorithm reconstructs an object using only the modulus of its ft . at the @xmath527 iteration , @xmath528 , an estimate of the object ft ,
is compared with the measured one and made to conform with the modulus at all fourier frequencies .
the inverse transform of the result yields an image @xmath529 .
this iteration is completed by forming a new estimate of the object that conforms to certain object - domain constraints , e.g. , positivity and finite extent , such that , @xmath530 , \\
\widehat{\cal g}_k^\prime{(\bf u ) } & = & |\widehat{\cal i}{(\bf u)}| e^{i\phi_k{(\bf u ) } } , \\ { \cal g}_k^\prime{(\bf x ) } & = & { \cal f}^{-1 } [ \widehat{\cal g}_k^\prime{(\bf u ) } ] , \\ { \cal g}_{k+1}({\bf x } ) & = & { \cal g}_k^{\prime } ( { \bf x } ) , \ , \ , \ , { \bf x } \notin \gamma , \nonumber\\ & & = 0 , \ , \ , \ , \ , \ , \ , \ , \ , \ , \ , { \bf x } \in \gamma,\end{aligned}\ ] ] where the region @xmath228 is the set of all points at which @xmath529 violates the object - domain constraints , and @xmath531 , and @xmath532 are estimates of @xmath533 , and the phase @xmath262 of @xmath534 , respectively . the above procedure may be accelerated if an estimate , @xmath535 is formed as , @xmath536 where @xmath537 is a constant feedback parameter .
i. the richardson - lucy ( richardson , 1972 , lucy , 1974 ) algorithm converges to the maximum likelihood solution for poisson statistics in the data which is appropriate for optical data with noise from counting statistics .
it forces the restored image to be non - negative and conserves flux both globally and locally at each iteration .
the magain , courbin , sohy , ( mcs ; magain et al .
1998 ) algorithm is based on the principle that sampled data can not be fully deconvolved without violating the sampling theorem ( shannon , 1949 ) that determines the maximal sampling interval allowed so that an entire function can be reconstructed from sampled data .
the sampled image should be deconvolved by a narrower function instead of the total psf so that the resolution of the deconvolved image is compatible with the adopted sampling .
the positivity constraint unlike the traditional deconvolution methods , is not mandatory ; accurate astrometric and photometric informations of the astronomical objects can be obtained .
\iii . the myopic iterative step preserving algorithm , ( mistral ; conan et al .
1998 ) , is based on the bayesian theorem that uses the probabilistic approach .
it incorporates a positivity constraint and some a priori knowledge of the object ( an estimate of its local mean and a model for the power spectral density etc . ) .
it also allows one to account for the noise in the image , the imprecise knowledge of the psf .
mistral has produced a few spectacular images after processing ao images .
a non - linear iterative image reconstruction algorithm , called pixon method has been developed by puetter and yahil ( 1999 ) that provides statistically unbiased photometry and robust rejection of spurious sources . unlike other bayesian methods
, this technique does not assign explicit prior probabilities to image models .
it minimizes complexity by smoothing the image model locally .
the model is then described using a few irregular - sized pixons containing similar amounts of information , rather than many regular pixels containing variable signal - to - noise data .
eke ( 2001 ) opines that it has the ability to detect sources in low snr data and to deconvolve a telescope beam in order to recover the internal structure of a source .
reconstruction of complex images involves the knowledge of complex visibilities .
interferometers with two apertures have limited possibilities for image reconstructions due to absence of phase visibility recovering .
the phase of a visibility may be deduced from closure - phase that has been applied in optical interferometry ( baldwin et al . 1998 ) .
process after data acquisition consists of phase calibration and visibility phase reconstruction from closure - phase terms by technique similar to bispectrum processing . from complex visibilities acquired from a phased interferometric array ,
it is possible to reconstruct the image by actually interpolating the function in the @xmath363 plane .
let the output of a synthesis array be a set of visibility functions @xmath538 that are obtained by averaging the quasi - sinusoidal response of each interferometer pair and hence , the resultant brightness distribution be given by the equation : @xmath539},\ ] ] where @xmath540 is usually called the dirty image which is the convolution of the true brightness distribution @xmath541 with the synthesized beam @xmath542 , @xmath543 is a weight associated with @xmath544 .
the map will show sufficient details when the @xmath363 coverage is good , but it is not the best representation of the sky and contains many artifacts notably the positive and negative side lobes around bright peaks . if the sampling of the @xmath363 plane is irregular and uneven , the dirty beam will have large sidelobes and confuse and obscure the structure of interest in the image .
several non - linear methods have been introduced to restore the unmeasured fourier components in order to produce a more physical map .
these methods work on the constraint that the image must be everywhere positive or zero and they make use of a priori knowledge about the extent of the source and statistics of the measurement processes .
since the restoration is not unique , the image restoration must select the best solution .
[ [ clean ] ] clean + + + + + the most commonly used approaches to deconvolve dirty image fall into two groups : clean algorithms ( hgbom , 1974 ) and maximum - entropy methods ( mem , ables , 1974 ) . to some extent ,
clean and mem are complimentary in their application . between them ,
clean is the most routinely used algorithm , particularly in radio astronomy , because it is both computationally efficient and intuitively easy to understand .
this procedure is a non - linear technique that applies iterative beam removing method .
it starts with a dirty image @xmath541 made by the linear fourier inversion procedure and attempts to decompose this image into a number of components , each of which contains part of the dirty beam @xmath542 .
one wishes to determine the set of numbers @xmath545 such that @xmath546 where @xmath547 is the residual brightness distribution after the decomposition .
the solution is considered satisfactory if @xmath547 is of the order of the expected noise .
the algorithm searches the dirty image for the pixel with largest absolute value @xmath548 and subtracts a dirty beam pattern centered on this point with amplitude @xmath549 , where @xmath550 is called the loop gain .
the residual map is searched for the next largest value and the second beam stage is subtracted and so on .
the iteration is stopped when the maximum residual is consistent with the noise level .
the iteration consists of a residual image that contains noise and low level source calculations plus a set of amplitudes and positions of the components removed .
these components can be considered as an array of delta function and convolved with a clean beams and added to produce a cleaned image .
the clean beam is usually chosen as a truncated elliptical gaussian about the same size of the main lobe of the dirty beam .
clean results in a map with the same resolution as the original dirty map without sidelobes .
[ [ maximum - entropy - method - mem ] ] maximum entropy method ( mem ) + + + + + + + + + + + + + + + + + + + + + + + + + + + + it is found that clean generally performs well on small compact sources , while mem does better on extended sources .
mem is employed in a variety of fields like medical imaging , crystallography , radio and x - ray astronomy as well .
this procedure governs the estimation of probability distributions when limited information is available .
in addition , it treats all the polarization component images simultaneously ( unlike clean which deconvolves different polarization component images independently ) and guarantees essential conditions on the image .
it makes use of the highest spatial frequency information by finding the smoothest image consistent with the interferometric data .
while enforcing positivity and conserving the total flux in the frame , smoothness is estimated by the ` entropy ' @xmath551 that is of the form , @xmath552 where @xmath553 $ ] represents the image to be restored , and @xmath554 $ ] is known as prior image .
it can be shown that @xmath555 ; the equality holds if @xmath556 . the value of @xmath551 is a measure of the similarity between @xmath557 and @xmath558 if the entropy @xmath551 is maximized without any data constraints . with data constraints ,
the maximum of @xmath551 will be less than its absolute maximum value zero , meaning that @xmath557 has been modified and deviated from its prior model @xmath558 .
mem solves the multi - dimensional constraints minimization problem .
it uses only those measured data and derives a brightness distribution which is the most random , i.e. , has the maximum entropy @xmath551 of any brightness distribution consistent with the measured data . maintaining an adequate fit to the data , it reconstructs the final image that fits the data within the noise level .
monnier et al .
( 2001 ) have reconstructed the dust shells around two evolved stars , ik tau and vy cma using @xmath363 coverage from the contemporaneous observations at keck - i and iota .
figure 16 depicts the mem reconstructions of the dust shells around these stars .
their results clearly indicate that without adequate spatial resolution , it is improbable to cleanly separate out the contributions of the star from the hot inner radius of the shell ( left panel in figure 16 ) .
they opined that image reconstructions from the interferometer data are not unique and yield results which depend heavily on the biases of a given reconstruction algorithm . by including the long baselines ( @xmath559 20 m ) data from the iota interferometer ( right panel ) , the fraction of the flux arising from the central star can be included in the image reconstruction process by using the mem prior
one can see for a dust shell such as ik tau , that additional iota data are critical in accurately interpreting the physical meaning of interferometer data .
the thick dashed lines show the expected dust shell inner radius from the data obtained at the isi .
[ [ self - calibration ] ] self calibration + + + + + + + + + + + + + + + + in case of producing images with accurate visibility amplitude and poor or absence of phases , self calibration can be used ( baldwin and warner , 1976 ) .
cornwell and wilkinson , 1981 ) introduced a modification by explicitly solving for the telescope - specific error as part of the reconstructing step .
the measured fourier phases are fit using a combination of intrinsic phases plus telescope phase errors .
if the field of observation contains one dominating internal point source which can be used as an internal phase reference , the visibility phase at other spatial frequencies is derived . a hybrid map can be made with the measured amplitudes together with model phase distributions . since the measured amplitudes differ from the single point source model , the hybrid map diffuses from the model map .
adding some new feature to the original model map , an improved model map is obtained in the next iteration . with clever selection of features to be added to the model in each iteration , the procedure converges .
[ [ linear - approach ] ] linear approach + + + + + + + + + + + + + + + another approach to the deconvolution problem is to formulate it as a linear system of the form : @xmath560 and then use algebraic techniques to solve this system .
the elements of * a * contain samples of the dirty beam , the elements of * b * are samples of the dirty image , while * x * contains components of the reconstructed image . without any additional constraints matrix * a * is singular : additional information has to be added to find a solution . assumptions that regularize the system include positivity and compact structure of the source .
an algebraic approach is not new in itself , but practical applications of such techniques in astronomy have become feasible only recently , e.g. , for the non - negative least squares ( nnls , briggs , 1995 ) . [ [ wipe ] ] wipe + + + + lannes et al .
( 1994 ) present another technique , wipe , based on a least squares approach , where support information is used to regularize the algorithm .
again , a linear system similar to that mentioned in the previous paragraph is solved , but using a technique which iterates between ( @xmath408 ) and image - planes .
unlike clean and mem , wipe suppresses excessive extrapolation of higher spatial frequencies during the deconvolution .
optical interferometry in astronomy is a boon for astrophysical studies .
the following sub - sections dwell on the astrophysical importance and the perspectives of interferometry .
single aperture interferometry has been contributing to study of the sun and solar system , and of a variety of stellar astrophysical problems .
the existence of solar features with sizes of the order of 100 km or smaller was found by means of speckle imaging ( harvey , 1972 , harvey and breckinridge , 1973 ) . from observations of photospheric granulation from disk center to limb at @xmath561 , wilken et al .
( 1997 ) found a decrease of the relative rms - contrast of the center - to - limb of the granular intensity .
a time series of high spatial resolution images reveal the highly dynamical evolution of sunspot fine structure , namely , umbral dots , penumbral filaments , facular points ( denker , 1998 ) .
small - scale brightening in the vicinity of sunspots , were also observed in the wings of strong chromospheric absorption lines .
these structures which are concomitant with strong magnetic fields show brightness variations close to the diffraction - limit of the vacuum tower telescope ( @xmath1940.16@xmath2 at 550 nm ) , observatorio del teide ( tenerife ) . with the phase - diverse speckle method , seldin et al .
( 1996 ) found that the photosphere at scales below 0.3@xmath2 is highly dynamic .
speckle imaging has been successful in resolving the heavenly dance of pluto - charon system ( bonneau and foy , 1980 ) , as well as in determining shapes of asteroids ( drummond et al .
reconstructions of jupiter with sub - arcsecond resolution have also been carried out by saha et al .
( 1997 ) .
studies of close binary stars play a fundamental role in measuring stellar masses , providing a benchmark for stellar evolution calculations ; a long term benefit of interferometric imaging is a better calibration of the main - sequence mass - luminosity relationship .
the parameter in obtaining masses involves combining the spectroscopic orbit with the astrometric orbit from the interferometric data .
continuous observations are necessary to be carried out in order to derive accurate orbital elements and masses , luminosities and distances .
the radiative transfer concerning the effects of irradiation on the line formation in the expanding atmospheres of the component that is distorted mainly by physical effects , viz.,(i ) rotation of the component , and ( ii ) the tidal effect can be modeled as well .
more than 8000 interferometric observations of stellar objects have been reported so far , out of which 75% are of binary stars ( hartkopf et al .
the separation of most of the new components discovered are found to be less than 0.25@xmath2 . from an inspection of the interferometric data , mason et al .
( 1999 ) have confirmed the binary nature of 848 objects , discovered by the hipparcos satellites ; prieur et al .
( 2001 ) reported high angular resolution astrometric data of 43 binary stars that were also observed with same satellite .
torres et al .
( 1997 ) derived individual masses for @xmath562 tau using the distance information from @xmath563 tau ; they found that the empirical mass - luminosity relation in good agreement with theoretical models .
gies et al .
( 1997 ) measured the radial velocity for the massive binary 15 mon . with the speckle spectrograph , baba , kuwamura , miura et al .
( 1994 ) have observed a binary star , @xmath564 and ( @xmath134 = 0.53@xmath2 ) and found that the primary ( a be star ) has h@xmath0 in emission while the companion has h@xmath0 in absorption .
the high angular polarization measurements of the pre - main sequence binary system z cma at 2.2 @xmath344 m by fischer et al .
( 1998 ) reveal that both the components are polarized ; the secondary showed an unexpected by large degree of polarization .
studies of multiple stars are also an important aspect that can reveal mysteries . for instance , the r136a was thought to be a very massive star with a mass of @xmath1942500m@xmath565 ( cassinelli et al . 1981 ) .
speckle imagery revealed that r136a was a dense cluster of stars ( weigelt and baier , 1985 , pehlemann et al .
r64 ( schertl et al . 1996 ) , hd97950 , and the central object of giant hii region starburst cluster ngc3603 ( hofmann et al .
1995 ) have been observed as well .
the star - like luminous blue variable ( lbv ) , @xmath566 carinae , an intriguing massive southern object , was found to be a multiple object ( weigelt and ebersberger , 1986 ) .
the polarimetric reconstructed images with a 0.11@xmath2 resolution in the h@xmath0 line of @xmath566 carina exhibit a compact structure elongated in consistence with the presence of a circumstellar equatorial disk ( falcke et al .
1996 ) .
many supergiants have extended gaseous envelope which can be imaged in their chromospheric lines .
the diameter of a few such objects , @xmath0 ori and mira , are found to be wavelength dependent ( bonneau and labeyrie , 1973 , labeyrie et al .
1977 , weigelt et al . 1996 ) .
recent studies have also confirmed the asymmetries on their surfaces ; the presence of hotspots are reported as well ( wilson et al . 1992 , haniff et al . 1995 , bedding , zijlstra et al .
1997 , tuthill et al .
1997 , monnier et al . 1999 ,
tuthill , monnier et al .
the surface imaging of long period variable stars ( tuthill , haniff et al .
1999 ) , vy cma reveals non - spherical circumstellar envelope ( wittkowski , langer et al .
monnier et al .
( 1999 ) have found emission to be one - sided , inhomogeneous and asymmetric in ir and derived the line - of - sight optical depths of its circumstellar dust shell .
the radiative transfer modeling of the supergiant nml cyg reveals the multiple dust shell structures ( blcker et al .
2001 ) .
high resolution imagery may depict the spatial distribution of circumstellar matter surrounding objects which eject mass , particularly in young compact planetary nebula ( ypn ) or newly formed stars in addition to t tauri stars , late - type giants or supergiants .
the large , older and evolved planetary nebula ( pn ) show a great variety of structure ( balick , 1987 ) that are ( a ) spherically symmetric ( a39 ) , ( b ) filamentary ( ngc6543 ) , ( c ) bipolar ( ngc6302 ) , and ( d ) peculiar ( a35 ) .
the structure may form in the very early phases of the formation of the nebula itself which is very compact and unresolved . by making maps at many epochs , as well as by following the motion of specific structural features
, one would be able to understand the dynamical processes at work .
the structures could be different in different spectral lines e.g. , ionization stratification in ngc6720 ( hawley and miller , 1977 ) , and hence maps can be made in various atomic and ionic emission lines too .
major results , such as , ( i ) measuring angular diameters of several ypns ( barlow et al .
1986 , wood et al .
( ii ) resolving five individual clouds around carbon star irc+10216 ( see figure 17 ) with a central peak surrounded by patchy circumstellar matter ( osterbart et al .
1996 , weigelt et al .
1998 ) , ( iii ) exhibiting two lobes of the evolved object , the red rectangle ( see figure 18 ) ( osterbart et al .
1996 ) , ( iv ) revealing a spiral pinwheel in the dust around wr104 ( tuthill , monnier et al .
1999 ) , and ( v ) depicting spherical dust shell around oxygen - rich agb star afgl 2290 ( gauger et al .
1999 ) are to be noted ; images of the young star , lkh@xmath0 101 in which the structure of the inner accretion disk is resolved have been reported as well ( tuthill et al .
detailed information that is needed for the modeling of the 2-d radiative transfer concerning the symmetry @xmath197 spherical , axial or lack of clouds , plumes etc . of the objects
@xmath197 can also be determined ( menshchikov and henning , 1997 , gauger et al .
1999 ) .
both novae and supernovae ( sn ) have a complex nature of shells viz . ,
multiple , secondary and asymmetric .
high resolution mapping may depict the events near the star and the interaction zones between gas clouds with different velocities .
soon after the explosion of the supernova sn1987a , various observers monitored routinely the expansion of the shell in different wavelengths ( nisenson et al .
1987 , papaliolios et al .
1989 , wood et al . 1989 ) .
the increasing diameter of the same in several spectral lines and the continuum was measured ( karovska et al . 1989 ) .
nulsen et al .
( 1990 ) have derived the velocity of the expansion as well and found that the size of this object was strongly wavelength dependent at the early epoch @xmath197 pre - nebular phase indicating stratification in its envelope . a bright source at 0.06@xmath2 away from the sn1987a with @xmath567 2.7 mag at h@xmath0 had also been detected .
recently , nisenson and papaliolios ( 1999 ) have detected a faint second spot , @xmath568 4.2 mag , on the opposite side of sn1987a with @xmath134 = 160 mas .
another important field of observational astronomy is the study of the physical processes , viz .
, temperature , density and velocity of gas in the active regions of active galactic nuclei ( agn ) ; optical imaging by emission lines on sub - arcsecond scales can reveal the structure of the narrow - line region .
the scale of narrow - line regions is well resolved by the diffraction - limit of a moderate - sized telescope .
the time variability of agns ranging from minutes to decades can also be studied .
the ngc 1068 is an archetype type 2 seyfert galaxy .
observations of this object corroborated with theoretical modeling like radiative transfer calculations have made significant contributions on its structure .
ebstein et al .
( 1989 ) found a bipolar structure of this object in the [ oiii ] emission line .
near - ir observations at the keck i telescope trace a very compact central core and extended emission with a size of the order of 10 pc on either side of an unresolved nucleus ( weinberger et al .
wittkowski , balega et al .
( 1998 ) have resolved central 2.2 @xmath344 m core by bispectrum speckle interferometry at the diffraction - limit of the special astrophysical observatory ( sao ) 6 m telescope , with a fwhm size of @xmath1942 pc for an assumed gaussian intensity distribution .
figure 19 depicts the reconstructed image of the agn , ngc1068 .
subsequent observations by wittkowski et al .
( 1999 ) indicate that this compact core is asymmetric with a position angle of @xmath19420@xmath289 and an additional more extended structure in n - s direction out to @xmath19425 pc .
quasars ( qso ) may be gravitationally lensed by stellar objects such as , stars , galaxies , clusters of galaxies etc . , located along the line of sight .
the aim of the high angular imagery of these qsos is to find their structure and components ; their number and structure as a probe of the distribution of the mass in the universe .
the capability of resolving these objects in the range of 0.2@xmath2 to 0.6@xmath2 would allow the discovery of more lensing events . the gravitational image of the multiple qso pg1115 +
08 was resolved by foy et al .
( 1985 ) ; one of the bright components , discovered to be double ( hege et al .
1981 ) , was found to be elongated that might be , according to them , due to a fifth component of the qso .
most of the results that obtained from the ground - based telescopes equipped with ao systems are in the near - ir band ; while results at visible wave lengths continue to be sparse .
the contributions are in the form of studying ( i ) planetary meteorology ( poulet and sicardy , 1996 , marco et al .
1997 , roddier et al .
1997 ) ; images of neptune s ring arcs are obtained ( sicardy et al . 1999 ) that are interpreted as gravitational effects by one or more moons , ( ii ) nucleus of m31 ( davidge , rigaut et al .
1997 ) , ( iii ) young stars and multiple star systems ( bouvier et al . 1997 ) , ( iv ) galactic center ( davidge , simons et al .
1997 ) , ( v ) seyfert galaxies , qso host galaxies ( hutchings et al .
1998 , 1999 ) , and ( vi ) circumstellar environment ( roddier et al . 1996 ) .
the images of the objects such as , ( a ) the nuclear region of ngc3690 in the interacting galaxy arp 299 ( lai et al .
1999 ) , ( b ) the starburst / agns , ngc863 , ngc7469 , ngc1365 , ngc1068 , ( c ) the core of the globular cluster m13 ( lloyd - hart et al . 1998 ) . and ( d ) r136 ( brandl et al . ( 1996 ) etc .
are obtained from the moderate - sized telescopes .
the highest ever angular resolution ao images of the radio galaxy 3c294 in the near - ir bands have been obtained at keck observatory ( quirrenbach et al . 2001 ) .
ao systems can also employed for studying young stars , multiple stars , natal disks , and related inward flows , jets and related outward flows , proto - planetary disks , brown dwarfs and planets .
roddier et al .
( 1996 ) have detected a binary system consisting of a k7-mo star with an m4 companion that rotates clockwise ; they suggest that the system might be surrounded by a warm unresolved disk .
the massive star sanduleak-66@xmath28941 in the lmc was resolved into 12 components by heydari and beuzit ( 1994 ) .
success in resolving companions to nearby dwarfs has been reported ( beuzit et al . 2001 , kenworthy et al . 2001 ) .
macintosh et al . ( 2001 ) measured the position of the brown dwarf companion to twa5 and resolved the primary into an 0.055@xmath2 double . the improved resolution of crowded fields like globular clusters would allow derivation of luminosity functions and spectral type , to analyze proper motions in their central area .
simon et al .
( 1999 ) have detected 292 stars in the dense trapezium star cluster of the orion nebula and resolved pairs to the diffraction - limit of a 2.2 m telescope .
optical and near - ir observations of the close herbig ae / be binary star nx pup ( brandner et al .
1995 ) , associated with the cometary globular cluster i , schller et al .
( 1996 ) estimated the mass and age of both the components and suggest that circumstellar matter around the former could be described by a viscous accretion disk .
line and continuum fluxes , and equivalent widths are also derived for the massive stars in the arches cluster ( blum et al .
2001 ) .
stellar populations in galaxies in near - ir region provides the peak of the spectral energy distribution for old populations .
bedding , minniti et al .
( 1997 ) have observed the sgr a window at the galactic center of the milky way .
they have produced an ir luminosity function and color - magnitude diagram for 70 stars down to @xmath56919.5 mag .
figure 20 depicts the adonis k@xmath570 image of the sgr a window .
images have been obtained of the star forming region messier 16 ( currie et al . 1996 ) , the reflection nebula ngc2023 revealing small - scale structure in the associated molecular cloud , close to the exciting star , in orion ( rouan et al .
close et al .
( 1997 ) mapped near - ir polarimetric observations of the reflection nebula r mon resolving a faint source , 0.69@xmath2 away from r mon and identified it as a t tauri star .
monnier et al .
( 1999 ) found a variety of dust condensations that include a large scattering plume , a bow shaped dust feature around the red supergiant vy cma ; a bright knot of emission 1@xmath2 away from the star is also reported .
they argued in favor of the presence of chaotic and violent dust formation processes around the star .
imaging of proto - planetary nebula ( ppn ) , frosty leo and the red rectangle by roddier et al .
( 1995 ) revealed a binary star at the origin of these ppns .
imaging of the extragalactic objects particularly the central area of active galaxies where cold molecular gas and star formation occur is an important program . from the images of nucleus of ngc1068 ,
rouan et al . (
1998 ) , found several components that include : ( i ) an unresolved conspicuous core , ( ii ) an elongated structure , and ( iii ) large and small - scale spiral structures .
lai et al . (
1998 ) have recorded images of markarian 231 , a galaxy 160 mpc away demonstrating the limits of achievements in terms of morphological structures of distant objects .
aretxaga et al .
( 1998 ) reported the unambiguous detection of the host galaxy of a normal radio - quiet qso at high - redshift in k - band ; detection of emission line gas within the host galaxies of high redshift qsos has been reported as well ( hutchings et al .
observations by ledoux et al .
( 1998 ) of broad absorption line quasar apm 08279 + 5255 at z=3.87 show the object consists of a double source ( @xmath134 = 0.35@xmath2 @xmath443 0.02@xmath2 ; intensity ratio = 1.21 @xmath443 0.25 in h band ) .
they proposed for a gravitational lensing hypothesis which came from the uniformity of the quasar spectrum as a function of the spatial position .
search for molecular gas in high redshift normal galaxies in the foreground of the gravitationally lensed quasar q1208 + 1011 has also been made ( sams et al . 1996 ) .
ao imaging of a few low and intermediate redshift quasars has been reported recently ( mrquez et al . 2001 ) .
high resolution stellar coronagraphy is of paramount importance in ( i ) detecting low mass companions , e.g. , both white and brown dwarfs , dust shells around asymptotic giant branch ( agb ) and post - agb stars ,
( ii ) observing nebulosities leading to the formation of a planetary system , ejected envelops , accretion disk , and ( iii ) understanding of structure ( torus , disk , jets , star forming regions ) , and dynamical process in the environment of agns and qsos . by means of coronagraphic techniques the environs of a few interesting objects have been explored .
they include : ( i ) a very low mass companion to the astrometric binary gliese 105 a ( golimowski et al .
1995 ) , ( ii ) a warp of the circumstellar disk around the star @xmath537 pic ( mouillet et al .
1997 ) , ( iii ) highly asymmetric features in ag carina s circumstellar environment ( nota et al .
1992 ) , ( iv ) bipolar nebula around the lbv r127 ( clampin et al .
1993 ) , and ( v ) the remnant envelope of star formation around pre - main sequence stars ( nakajima and golimowski , 1995 ) .
the main objective of lbois is to measure the diameters , distances , masses and luminosities of stars , to detect the morphological details , such as granulations , oblateness of giant stars , and the image features , i.e. , spots and flares on their surfaces .
eclipsing binaries are also good candidates ; for they provide information on circumstellar envelopes such as , the diameter of inner envelope , color , symmetry , radial profile etc . as stated earlier ( in section vii.a.2 )
, good spectroscopic and interferometric measurements are required to derive precise stellar masses since they depend on @xmath571 .
a small variation on the inclination @xmath572 implies a large variation on the radial velocities .
most of the orbital calculations that are carried out with speckle observations are not precise to provide masses better than 10% ( pourbaix , 2000 ) .
the results obtained so far with lbois are from the area of stellar angular diameters with implications for emergent fluxes , effective temperatures , luminosities and structure of the stellar atmosphere , dust and gas envelopes , binary star orbits with impact on cluster distances and stellar masses , relative sizes of emission - line stars and emission region , stellar rotation , limb darkening , astrometry etc .
( saha , 1999 , saha and morel , 2000 , quirrenbach , 2001 and references therein ) . the angular diameter for more than 50 stars has been measured ( dibenedetto and rabbia , 1987 , mozurkewich et al . 1991 , dyck et al . 1993 , nordgren et al . 2000 , perin et al .
1999 , kervella et al .
2001 , van belle et al .
2001 ) with accuracy better than 1% in some cases .
interesting results that have been obtained using i2 t and mark iii interferometers are : ( i ) measuring diameters , effective temperatures of giant stars ( faucherre et al .
1983 , dibenedetto and rabbia , 1987 ) , ( ii ) resolving the gas envelope of the be star @xmath228 cas in the h@xmath0 line ( thom et al . 1986 ) , and structure of circumstellar shells ( bester et al . 1991 ) , and ( iii ) determining orbits for spectroscopic , and eclipsing binaries ( armstrong et al .
1992 , shao and colavita , 1994 ) .
the gi2 t is being used regularly to observe the be stars , lbvs , spectroscopic and eclipsing binaries , wavelength dependent objects , diameters of bright stars , and circumstellar envelopes . however , the scientific programs are restricted by the low limiting visible magnitude down to 5 ( seeing and visibility dependent ) . the first successful result that is reported on resolving the rotating envelope of hot star , @xmath228 cas came out of this instrument in 1989 .
mourard et al .
( 1989 ) observed this star with a spectral resolution of 1.5 centered on h@xmath0 .
as many as @xmath194300,000 short - exposure images were recorded by a photon - counting camera , cp40 ( blazit , 1986 ) . they have digitally processed each image , which contained @xmath194100 photons , using the correlation technique .
the results were co - added to reduce the effect of atmospheric seeing and photon noise , according to the principle of speckle interferometry ( labeyrie , 1970 ) . with the central star as a reference
, they have determined the relative phase of the shell visibility and showed clearly the rotation of the envelope .
this result demonstrates the potential of observations that combines spectral and spatial resolution . through subsequent observations on later dates ,
stee et al . ( 1995 , 1998 ) derived the radiative transfer model . using the data obtained since 1988 with this instrument , berio , stee et al .
( 1999 ) have found the evidence of a slowly prograde of rotating density pattern in the said star s equatorial disk .
indeed @xmath228 cas has been a favorite target to the gi2 t , with further systematic monitoring of multiple emission lines , the formation , structure , and dynamics of other be stars can be addressed .
the other noted results obtained in recent times with this instrument include the mean angular diameter and accurate distance estimate of @xmath573 cep ( mourard et al .
1997 ) , subtle structures in circumstellar environment such as jets in the binary system @xmath537 lyr ( harmanec et al . 1996 ) , clumpiness in the wind of p cyg ( vakili et al .
1997 ) , and detection of prograde one - armed oscillation in the equatorial disk of the be star @xmath217 tau ( vakili et al . 1998 ) . with the susi instrument , davis et al .
( 1998 , 1999b ) have determined the diameter of @xmath573 cma with an accuracy of @xmath4431.8% . from the data obtained at the coast , aperture - synthesis maps of the double - lined spectroscopic binary @xmath0 aur ( baldwin et al .
1996 ) depict the milli - arcsecond orbital motion of the system over a 15 day interval .
images of @xmath0 ori reveals the presence of a circularly symmetric data with an unusual flat - topped and limb darkening profile ( burns et al .
young et al . ( 2000 ) have found a strong variation in the apparent symmetry of the brightness distribution as function of wavelength .
variations of the cycle of pulsation of the mira variable r leo have been measured ( burns et al .
1998 ) . with iota ,
the angular diameters and effective temperatures have been measured for carbon stars ( dyck et al .
1996 ) , mira variables ( van belle et al .
1996 , 1997 ) , cool giant stars ( perrin et al .
1998 , 1999 , cepheids ( kervella et al . 2001 ) , and dust shell of ci cam ( traub et al . 1998 ) .
millan - gabet et al .
( 2001 ) have resolved circumstellar structure of herbig ae / be stars in near - ir .
figure 21 depicts the examples of h - band visibility data and models for 2 sources .
the lower right panel in figure 21 illustrates the observed lack of visibility variation with baseline rotation , consistent with circumstellar emission from dust which is distributed either in a spherical envelope or in a disk viewed almost face - on .
figure 22 summarizes the existing set of measurements of near - ir sizes of said herbig sources ( credit : r. millan - gabet and j. d. monnier ) , using the data from danchi et al .
( 2001 ) , millan - gabet et al .
( 2001 ) , and tuthill et al .
the agreement observed in most cases has motivated in part a revision of disk physics in models of herbig ae / be systems ( natta et al .
2001 ) . in the new models ,
the gas in the inner disk is optically thin so that dust at the inner edge is irradiated frontally , and expands forming a ` puffed - up ' inner wall . due to the extra heating ,
compared to the irradiation of a flat disk traditionally considered , this model results in larger near - ir sources , essentially corresponding to the ` puffed up ' inner wall , which appears as a bright ring to the interferometer .
monnier et al .
( 2001 ) have reported the results of decomposing the dust and stellar signatures from the evolved stars ( figure 16 ) . with pti , malbet et al .
( 1998 ) have resolved the young stellar object ( yso ) , fu ori in the near - ir with a projected resolution better than 2 au .
measurements of diameters and effective temperatures of cool stars and cepheids have been reported ( van belle et al .
1999 , lane et al . 2000 ) .
the visual orbit for the spectroscopic binary @xmath574 peg with interferometric visibility data has also been derived ( boden et al . 1999 ) .
direct observations of an oblate photosphere of a main sequence star , @xmath0 aql ( van belle et al .
2001 ) have been carried out .
ciardi et al .
( 2001 ) have resolved the stellar disk of @xmath0 lyr in the near - ir .
employing npoi , hummel et al .
( 1998 ) have determined the orbital parameters of two spectroscopic binaries , @xmath217 uma ( mizar a ) , @xmath566 peg ( matar ) and derived masses and luminosities .
the limb - darkened diameters have also been measured of late - type giant stars ( hajian et al .
1998 , pauls et al .
1998 , wittkowski et al . , 2001 ) , and cepheid variables ( nordgren et al .
2000 ) .
the observing programs of isi have been aimed at determining the spatial structure and temporal evolution of dust shell around long period variables . from the data obtained with this instrument at 11.15 @xmath344 m , danchi et al .
( 1994 ) showed that the radius of dust formation depends on the spectral type of the stars .
lopez et al . (
1997 ) have found a strong dependence on the pulsation phase .
observations have been made of nml cyg , @xmath0 sco , as well as of changes in the dust shell around mira and ik tau , ( hale et al .
1997 , lopez et al .
1997 ) , mid - ir molecular absorption features of ammonia and silane of irc+10216 and vy cma ( monnier et al .
2000 ) .
recent observations with the two of the vlt telescopes have measured the angular diameter of the blue dwarf @xmath0 eri , which was found to be 1.92 mas ( glindemann and paresce , 2001 ) .
subsequently they ( i ) derived the diameters of a few red dwarf stars , and ( ii ) determined the variable diameters of a few pulsating cepheid stars , as well as the measurement of the core of @xmath566 carinae .
the future of high resolution optical astronomy lies with the new generation arrays but its implementation is a challenging task .
numerous technical challenges for developing such a system will require careful attention .
nevertheless , steady progress has enabled scientists to expand their knowledge of astrophysical processes . with improved technology ,
the interferometric arrays of large telescopes may provide snap - shot images at their recombined focus and yield improved images , spectra of quasar host galaxies , and astrometric detection and characterization of extra - solar planets . the expected limiting magnitude of a hyper - telescope imaging technique
is found to be 8.3 @xmath141 by numerical solution if 10-cm apertures are used and 20 @xmath141 for 10-m apertures ( pedretti and labeyrie , 1999 ) .
the limit is expected to increase with the carlina array ( labeyrie , 1999b ) , a 100-element hyper - telescope with a diameter of 200 m , shaped like arecibo radio telescope in puerto rico .
high precision astrometry also helps in establishing the cosmic distance scale ; measurements of proper motion can confirm stars as members of cluster ( known distance ) that may elucidate the dynamics of the galaxy .
the quest for extra - solar planets ( wallace et al .
1998 ) is a challenge for narrow angle astrometry .
very valuable astrometric results from space have already been obtained by the hipparcos satellite ( perryman , 1998 ) .
hipparcos used phase - shift measurements of the temporal evolution of the photometric level of two stars seen drifting through a grid .
the successor of hipparcos , gaia ( lindengren and perryman , 1996 ) , will probably use the same technique with improvements , yielding more accurate results on a larger number of objects .
however , only space - borne interferometers will achieve very high precision angular measurements .
as many as seventy - six jovian - size planets orbiting stars have been identified by the doppler - fizeau effect ( mayor and queloz , 1995 , butler and marcy , 1996 , http://exoplanets.org ) . for one of them , an atmosphere containing
sodium ( observed in the sodium resonance doublet at 589.3 nm ) has been detected in absorption as the planet transits its parent star hd209458 ( charbonneau et al . , 2001 ) .
it may also be possible to detect smaller planets by measuring the stellar photocenter motion due to the wobble .
such photocenter measurements will require diffraction - limited imaging even for the best possible candidates .
interferometry can be used to measure the ` wobble ' in the position of a star caused by the transverse component of a companion s motion .
a planet orbiting around a star causes a revolution of the star around the center of gravity defined by the two masses .
like galaxy velocity , this periodical short motion has a radial counterpart measurable from the ground by spectrometry .
the aim of the space interferometers like darwin and tpf is the discovery and characterization of terrestrial planets around nearby stars ( closer than 15 pc ) by direct detection ( i.e. , involving the detection of photons from the planet and not from the star as it is done with doppler - fizeau effect detection or wobble detection ) .
the difficulties for achieving earth - like planet detection come from ( i ) minimizing scattered light from the parent star and ( ii ) the presence of exo - zodiacal light ( infrared emission from the dust surrounding the parent star ) .
interferometric nulling technique will be useful to address the first issue .
the knowledge of the chemical composition of any planetary atmosphere gives hints about the likelihood to find carbon - based life .
lovelock ( 1965 ) has suggested that the simultaneous presence on earth of a highly oxidized gas , like @xmath575 , and highly reduced gases , like @xmath576 and @xmath577 is the result of the biochemical activity .
however , finding spectral signatures of these gases on an extra - solar planet would be very difficult .
an alternative life indicator would be ozone ( @xmath578 ) , detectable as an absorption feature at 9.6 @xmath344 m . on earth , ozone is photochemically produced from @xmath575 and , as a component of the stratosphere , is not masked by other gases . finding ozone would , therefore , indicate a significant quantity of @xmath579 that should have been produced by photosynthesis ( lger et al .
moreover , for a star - like the sun , detecting ozone can be done 1000 times faster than detecting @xmath575 at 0.76 @xmath344 m : estimates made by angel and woolf ( 1997 ) show that the requirements for planet detection in the visible with an 8 m telescope are not detectable with current technology .
space - borne interferometry projects for years spanning from 2020 to 2050 already exist ; such projects must be regarded as drafts for future instruments . for the post - tpf era
, nasa has imagined an enhanced version featuring four 25 m telescopes and a @xmath580 spectrometer .
this interferometer would be able to detect an extra - solar planet lines of gases directly produced by biochemical activity .
the next step proposed by nasa is an array of 25 telescopes , 40 m diameter each , that would yield 25@xmath34825 pixel images of an earth - like planet at 10 pc , revealing its geography and eventually oceans or chlorophyll zones .
a comparable project has been proposed by labeyrie ( 1999a ) .
it consists of 150 telescopes , 3 m diameter each , forming an interferometer with a 150 km maximum baseline .
such an instrument equipped with a highly efficient coronagraph would give a 40@xmath34840 pixel image of an earth - like planet at 3 pc .
figure 23 depicts a simulated image of an earth - like planet detection ( labeyrie , 1999b ) .
developing lbi for lunar operation consisting of 20 to 27 off - axis parabolic segments carried by robotic hexapodes that are movable during observing run has also been suggested by arnold et al .
earth - bound astronomical observations are strongly affected by atmospheric turbulence that set severe limits on the angular resolution , which in optical domain , is rarely better than @xmath194 0.5@xmath2 .
a basic understanding of interference phenomenon is of paramount importance to other branches of physics and engineering as well . in recent years
, a wide variety of applications of speckle patterns has been found in many areas . though the statistical properties of the speckle pattern is complicated , a detailed analysis of this pattern is useful in information processing .
image - processing is a very important subject . a second - order moment ( power spectrum )
analysis provides only the modulus of the object ft , whereas a third - order moment ( bispectrum ) analysis yields the phase allowing the object to be fully reconstructed . a more recent attempt to go beyond the third order ,
e.g. , fourth - order moment ( trispectrum ) , illustrates its utility in finding optimal quadratic statistics through the weak gravitational lensing effect ( hu , 2001 ) .
this algorithm provides a far more sensitive test than the bispectrum for some possible sources of non - gaussianity ( kunz et al . , 2001 ) , however , its implementation in optical imaging is a difficult computational task .
deconvolution method is an important topic as well .
it applies to imaging in general which covers methods spanning from simple linear deconvolution algorithms to complex non - linear algorithms . in astronomy
, the field of research that has probably benefited the most from high angular resolution techniques using single telescopes , and will still benefit in the future , is undoubtedly the origin and evolution of stellar systems .
this evolution starts with star formation , including multiplicity , and ends with the mass loss process which recycles heavier elements into the interstellar medium .
large - scale star formation provides coupling between small - scale and large - scale processes .
stellar chemical evolution or nucleo - synthesis that is a result of star formation activity further influences evolutionary process .
high resolution observations are fruitful for the detection of proto - planetary disks and possibly planets ( either astrometrically , through their influence on the disk , or even directly ) .
the technique is also being applied to studies of starburst and seyfert galaxies , agns , and quasars .
studies of the morphology of stellar atmospheres , the circumstellar environment of nova or supernova , ypn , long period variables ( lpv ) , rapid variability of agns etc .
are also essential . in spite of the limited capability of retrieving fully diffraction - limited images of the objects ,
ao systems are now offered to users of large telescopes .
ao observations have contributed to the study of the solar system , and added to the results of space borne instruments , for examples , monitoring of the volcanic activity on io or of the cloud cover on neptune , the detection of neptune s dark satellites and arcs , and the ongoing discovery of companions to asteroids etc . ; they are now greatly contributing to the study of the sun itself as well ( antoshkin et al .
2000 , rimmele , 2000 ) .
combination of ao systems with speckle imaging may enhance the results ( dayton et al . 1998 ) . by the end of the next decade ( post 2010 ) , observations using the ao system on a new generation telescope like owl ,
will revolutionize the mapping of ultra - faint objects like blazars that exhibit the most rapid and the largest amplitude variations of all agn ( ulrich et al .
1997 ) , extra - solar planets etc . ; certain aspects of galactic evolution like chemical evolution in the virgo cluster of galaxies can be studied as well .
a host of basic problems needs a very high angular resolution for their solution . among others , an important fundamental problem that can be addressed with
the interferometry and ao is the origin and evolution of galaxies .
the upcoming large facilities with phased arrays of multiple 8 - 10 m sub - apertures will provide larger collecting areas and higher spatial resolution simultaneously than the current interferometers .
these instruments fitted with complete ao systems would be able to provide imaging and morphological informations on the faint extragalactic sources such as , galactic centers in the young universe , deep fields , and host galaxies .
measurement of such objects may be made feasible by the instruments with a fairly complete @xmath408 coverage and large field of view .
the derivations of motions and parallaxes of galactic centers seem to be feasible with phase reference techniques .
origin of faint structures close to non - obscured central sources can also be studied in detail with interferometric polarization measurement .
the capabilities of the proposed large arrays offer a revolution in the study of compact astronomical sources from the solar neighborhood to cosmological distances .
the aim ranges from detecting other planetary systems to imaging the black hole driven central engines of quasars and active galaxies ; gamma - ray bursters may be the other candidate .
another important scientific objective is the recording of spectra to derive velocity and to determine black hole masses as a function of redshift . at the beginning of the present millennium
, several such arrays will be in operation both on the ground as well as in space .
space - borne interferometers are currently planned to detect planets either astrometrically ( sim ) or directly ( tpf ) .
projects like darwin and other ambitious imaging instruments may also come to function .
the author expresses his gratitude to a. labeyrie , s. t. ridgway , and p. a. wehinger for comments on the review , and indebtedness to v. coud du foresto , p. r. lawson , and s. morel for valuable communications .
thanks are also due to t. r. bedding , a. boccaletti , p. m. hinz , r. millan - gabet , j. d. monnier , r. osterbart , and m. wittkowski for providing the images , figures etc . , and granting permission for their reproduction , as well as to h. bradt and p. hickson for reading the manuscript . the services rendered by v. chinnappan , k. r. subramanian , and b. a. varghese are gratefully acknowledged
. ables j. g. , 1974 , astron .
* 15 * , 383 .
aime c. , 2000 , j. opt . a : pure & appl . opt . , * 2 * , 411 .
akeson r. , m. swain , and m. colavita , 2000 , spie , * 4006 * , 321 .
anderson j. a. , 1920 , astrophys .
j. , * 51 * , 263 . angel j. r. p. , 1994 , nature , * 368
angel j. r. p. , 2001 , nature , * 409 * , 427
. angel j. r. , and n. j. woolf , 1997 , astrophy .
j. , * 475 * , 373 .
antoshkin l. v. et al .
, 2000 , spie , * 4007 * , 232 .
aretxaga i. , d. mignant , j. melnick , r. terlevich , and b. boyle , 1998 , mon . not .
r. astron .
* 298 * , l13 .
armstrong j. t. , c. a. hummel , and d. mozurkewich , 1992 , proc .
eso - noao conf .
eds . , j. m. beckers & f. merkle , eso , germany , 673 .
armstrong j. et al . , 1998 ,
j. , * 496 * , 550 .
arnold l. , a. labeyrie , and d. mourard , 1996 , adv .
space res . *
18 * , 1148 .
ayers g. r. , and j. c. dainty , 1988 , opt .
* 13 * , 457 .
baba n. , s. kuwamura , n. miura , and y. norimoto , 1994 , astrophys . j. , * 431 * , l111 .
baba n. , s. kuwamura , and y. norimoto , 1994 , appl .
, * 33 * , 6662 .
baba n. , h. tomita , and n. miura , 1994 , appl . opt .
, * 33 * , 4428 .
babcock h. w. , 1953 , pub .
pac , * 65 * , 229 .
babcock h. w. , 1990 , science,*249 * , 253 .
bagnuolo jr .
w. g. , b. d. mason , d. j. barry , w. i. hartkopf , and h. a. mcalister , 1992 , astron .
j. , * 103 * , 1399 .
baldwin j. et al . , 1996 ,
astrophys . ,
* 306 * , l13 .
baldwin j. , r. boysen , c. haniff , p. lawson , c. mackay , j. rogers , d. st - jacques , p. warner , d. wilson , and j. young , 1998 , spie . , * 3350 * , 736 .
baldwin j. e. , c. a. haniff , c. d. mackay , and p. j. warner , 1986 , nature , * 320 * , 595 .
baldwin j. , r. tubbs , g. cox , c. mackay , r. wilson , and m. andersen , 2001 , astron .
astrophys . ,
* 368 * , l1 .
baldwin j. e. , and p. j. warner , 1978 , mon . not .
r. astron .
, * 182 * , 411 .
balick b. , 1987 , astron .
j. , * 94 * , 671 .
barlow m. j. , b. l. morgan , c. standley , and h. vine , 1986 , mon . not .
r. astron .
* 223 * , 151 .
barnaby d. , e. spillar , j. christou , and j. drummond , 2000 , astron .
j. , * 119 * , 378 .
bates r. , and m. mcdonnell , 1986 , ` image restoration and reconstruction ' , oxford eng . sc . , clarendon press .
beckers j. m. , 1982 , opt .
* 29 * , 361 .
bedding t. r. , d. minniti , f. courbin , and b. sams , 1997 , astron .
, * 326 * , 936 . bedding t. , a. zijlstra , o. von der lhe , j. robertson , r. marson , j. barton , and b. carter , 1997 , mon . not .
r. astron .
* 286 * , 957 .
beichman c. , 1998 , spie , * 3350 * , 719 .
benson j. , d. mozurkewich , and s. jefferies , 1998 , spie , * 3350 * , 493 .
berger j. , p. haguenauer , p. kern , k. perraut , f. malbet , i. schanen , m. severi , r. millan - gabet , and w. traub , 2001 , astron .
, * 376 * , l31 .
berio p. , d. mourard , d. bonneau , o. chesneau , p. stee , n. thureau , and f. vakili , 1999 , j. opt .
a. , * 16 * , 872 .
berio p. , p. stee , f. vakili , d. mourard , d. bonneau , o. chesneau , n. thureau , d. le mignant , and r. hirata , 1999 , astron .
, * 345 * , 203 .
bester m. , w. c. danchi , c. g. degiacomi , and c. h. townes , 1991 , astrophys .
j. , * 367 * , l27 . beuzit j. et al . , 2001 , astro - ph/0106277 .
blazit a. , 1986 , spie , * 702 * , 259 .
blcker t. , y. balega , k. -h .
hofmann , and g , weigelt , 2001 , astron .
( to appear ) .
blum r. , d. schaerer , a. pasquali , m. heydari - malayeri , p. conti , and w. schmutz , 2001 , astron .
j. , ( to appear ) .
boccaletti a. , 2001 , private communication .
boccaletti a. , c. moutou , d. mouillet , a. lagrange , and j. augereau , 2001 , astron .
astrophys . , * 367 * , 371 .
boccaletti a. , p. riaud , c. moutou , and a. labeyrie , 2000 , icarus , * 145 * , 628 .
boden a. et al . , 1999 ,
j. , * 515 * , 356 .
bonneau d. , and r. foy , 1980 , astron .
astrophys . , * 92 * , l1 .
bonneau d. , and a. labeyrie , 1973 , astrophys .
j , * 181 * , l1 .
born m. , and e. wolf , 1984 , principles of optics , pergamon press .
bouvier j. , f. rigaut , and d. nadeau , 1997 , astron .
astrophys . , * 323 * , 139 . bracewell r. n. , 1978 , nature , * 274 * , 780 . brandl b. , b. j. sams , f. bertoldi , a. eckart , r. genzel , s. drapatz , r. hofmann , m. lowe , and a. quirrenbach , 1996 , astrophys . j. , * 466 * , 254 .
brandner w. , j. bouvier , e. grebel , e. tessier , d. de winter , and j. l. beuzit , 1995 , astron .
astrophys . ,
* 298 * , 816 .
briggs d. s. , 1995 , ph .
d. thesis , new mexico institute of mining and technology .
bruns d. , t. barnett , and d. sandler , 1997 , spie .
, * 2871 * , 890 .
burge j. h. , b. cuerdem , and j. r. p. angel , 2000 , spie , * 4013 * , 640 .
burns d. et al . , 1997 ,
mon . not .
r. astron .
, * 290 * , l11 . burns d. et al . , 1998 , mon . not .
r. astron .
* 297 * , 467 .
butler r. , and g. marcy , 1996 , astrophys .
j , * 464 * , l153 .
cadot o. , y. couder , a. daerr , s. douady , and a. tsinocber , 1997 , phys .
, * 56 * , 427 .
callados m. , and m. vzquez , 1987 , astron .
astrophys . , * 180 * , 223 . carleton n. et al . , 1994 , spie , * 2200 * , 152
. cassinelli j. , j. mathis , and b. savage , 1981 , science , * 212 * , 1497 .
charbonneau d. , t. brown , r. noyes , and r. gilfiland , 2001 , astrophys .
j. , ( to appear ) .
ciardi d. , g. van belle , r. akeson , r. thompson , e. a. lada , and s. howell , 2001 , astrophys .
j. ( to appear ) .
clampin m. , j. croker , f. paresce , and m. rafal , 1988 , rev .
* 59 * , 1269 .
clampin m. , a. nota , d. a. golimowski , c. leitherer , and a. ferrari , 1993 , astrophys .
j. , * 410 * , l35 .
close l. , f. roddier , j. hora , j. graves , m. northcott , c. roddier , w. hoffman , a. doyal , g. fazio , and l. deutsch , 1997 , astrophys . j. , * 489 * , 210 .
colavita m. , 1999 , pub .
, * 111 * , 111 . colavita m. et al . , 1998 , spie , * 3350 * , 776 . colavita m. et al
. , 1999 , astron .
j. , * 117 * , 505 .
conan j. -m . , l. m. mugnier , t. fusco , v. michau , and g. rousset , 1998 , appl .
, * 37 * , 4614 .
conan r. , a. ziad , j. borgnino , f. martin , and a. tokovinin , 2000 , spie , * 4006 * , 963 .
cooper d. , d. bui , r. bailey , l. kozlowski , and k. vural , 1993 , spie , * 1946 * , 170 .
cornwell t. j. , and p. n. wilkinson , 1981 , mon . not .
r. astron .
soc . , * 196 * , 1067 .
coud du foresto v. , g. perrin , and m. boccas , 1995 , astron .
astrophys . , * 293 * , 278 .
coulman c. e. , 1974 , solar phys . ,
* 34 * , 491 .
coulman c. e. , 1985 , annu .
astrophys . ,
* 23 * , 19 .
currie d. , k. kissel , e. shaya , p. avizonis , d. dowling , and d. bonnacini , 1996 , the messenger , no . * 86 * , 31 .
danchi w. c. , m. bester , c. degiacomi , i. greenhill , and c. townes , 1994 , astron .
j. , * 107 * , 1469 .
danchi w. c. , p. g. tuthill , and j. d. monnier , 2001 , astrophys .
j. ( to appear ) .
dantowitz r. , s. teare , and m. kozubal , 2000 , astron .
j. , * 119 * , 2455 .
davidge t. j. , f. rigaut , r. doyon , and d. crampton , 1997 , astron .
j. , * 113 * , 2094 .
davidge t. j. , d. a. simons , f. rigaut , r. doyon , e. e. becklin , and d. crampton , 1997 , astron .
j. , * 114 * , 2586 .
davis j. , and w. j. tango , 1996 , pub .
, * 108 * , 456 .
davis j. , w. tango , a. booth , and j. obyrne , 1998 , spie , * 3350 * , 726 .
davis j. , w. tango , a. booth , t. ten brummelaar , r. minard , and s. owens , 1999a , mon . not .
r. astron .
soc . , * 303 * , 773
. davis j. , w. j. tango , a. j. booth , e. d. thorvaldson , and j. giovannis , 1999b mon . not .
r. astron .
soc . , * 303 * , 783 .
dayton d. , s. sandven , j. gonglewski , s. rogers , s. mcdermott , and s. browne , 1998 , spie , * 3353 * , 139 .
dejonghe j. , l. arnold , o. lardire , j. -p berger , c. cazal , s. dutertre , d. kohler , and d. vernet , 1998 , spie , * 3352 * , 603 .
denker c. , 1998 , solar phys . , * 81 * , 108 .
derie f. , m. ferrai , e. brunetto , m. duchateau , r. amestica , and p. aniol , 2000 , spie , * 4006 * , 99 .
dibenedetto g. p. , and y. rabbia , 1987 , astron .
astrophys . , * 188 * , 114 .
diericks p. , and r. gilmozzi , 1999 , proc .
` extremely large telescopes ' , eds . ,
t. andersen , a. ardeberg , and r. gilmozzi , 43 .
drummond j. , a. eckart , and e. hege , 1988 , icarus , * 73 * , 1 .
dyck h. , j. benson , and s. ridgway , 1993 , pub .
pac , * 105 * , 610 .
dyck h. , g. van belle , and j. benson , 1996 , astron .
j. , * 112 * , 294 . ebstein s. , n. p. carleton , and c. papaliolios , 1989 , astrophys .
j , * 336 * , 103 . eke v. , 2001 , mon . not .
r. astron .
soc . , * 320 * , 106 .
elias n. m. , 2001 , astrophys .
j. , * 549 * , 647 .
falcke h. , k. davidson , k. -h .
hofmann , and g. weigelt , 1996 , astron .
astrophys . * 306 * , l17 .
faucherre m. , d. bonneau , l. koechlin , and f. vakili , 1983 , astron .
* 120 * , 263 .
fienup j. r. , 1978 , opt .
lett . , * 3 * , 27 .
fischer o. , b. stecklum , and c. leinert , 1998 , astron .
astrophys . , * 334 * , 969 . fizeau h. , 1868 , c. r. acad .
paris , * 66 * , 934 .
fomalont e. b. , and m. c. h. wright , 1974 , in galactic and extra - galactic radio astronomy , eds .
, g. l. verschuur , and k. i. kellerman , 256 .
foy r. , d. bonneau , and a. blazit , 1985 , astron .
astrophys . ,
* 149 * , l13 .
foy r. , and a. labeyrie , 1985 , astron .
astrophys . ,
* 152 * , l29 .
fried d. l. , 1966 , j. opt .
am . , * 56 * , 1372 .
fugate r. et al . , 1994 ,
a. , * 11 * , 310 .
gauger a. , y. y. balega , p. irrgang , r. osterbart , and g. weigelt , 1999 , astron .
astrophys . ,
* 346 * , 505 .
gay j. , and d. mekarnia , 1988 , proc .
eso - noao conf . ed . ,
f. merkle , eso , frg , 811 .
gerchberg r. w. , and w. o. saxton , 1972 , optik , * 35 * , 237 .
gies r. , b. mason , w. bagnuolo , m. haula , w. i. hartkopf , h. mcalister , m. thaller , w. mckibben , and l. penny , 1997 , astrophys .
j. , * 475 * , l49 .
glindemann a. , 1997 , pub .
pac , * 109 * , 682 .
glindemann a. , r. g. lane , and j. c. dainty , 1991 , proc .
` digital signal processing ' , eds . , v. cappellini & a. g. constantinides , 59 .
glindemann a. , and f. paresce , 2001 , http://www.eso.org/outreach .
golimowski d. a. , t. nakajima , s. r. kulkarni , and b. r. oppenheimer , 1995 , astrophys .
j , * 444 * , l101 .
gonsalves s. a. , 1982 , opt .
* 21 * , 829 .
goodman j. w. , 1975 , ` laser speckle and related phenomena ' , ed . , j. c. dainty , springer - verlag , berlin , 9 .
goodman j. w. , 1985 , ` statistical optics ' , wiley , n. y. gorham p. w. , 1998 , spie , * 3350 * , 116 .
gorham p. , w. folkner , and g. blackwood , 1999 , asp conf .
, * 194 * , eds .
s. unwin , and r. stachnik , isbn : 1 - 58381 - 020-x , 359 .
greenwood d. p. , 1977
am . , * 67 * , 390 . grieger f. , f. fleischman , and g. weigelt , 1988 , proc
. eso - noao conf .
f. merkle , eso , frg , 225 .
haguenauer p. , m. sevei , i. schanen - duport , k. rousselet - perraut , j. berger , y. duchne , m. lacolle , p. kern , f. melbet , and p. benech , 2000 , spie , * 4006 * , 1107 .
hajian a. et al . , 1998 ,
j. , * 496 * , 484 .
hale d. , m. bester , w. danchi , w. fitelson , s. hoss , e. lipman , j. monnier , p. tuthill , and c. townes , 2000 , astrohys , j. , * 537 * , 998 .
hale d. et al . , 1997 ,
j. , * 490 * , 826 .
hanbury brown r. , 1974 , ` the intensity interferometry , its applications to astronomy ' , taylor & francis , london .
hanbury brown r. , and r. twiss , 1958 , proc .
a , * 248 * , 222 .
hanbury brown r. , r. c. jennison , and m. k. das gupta , 1952 , nature , * 170 * , 1061 .
haniff c. , m. scholz , and p. tuthill , 1995 , mon . not .
r. astron .
soc . , * 276 * , 640 .
et al . , 1996 , astron .
astrophys . , * 312 * , 879 . hartkopf w. i. , h. a. mcalister , and b. d. mason , 1997 , chara contrib
4 , ` third catalog of interferometric measurements of binary stars ' , w.i . hartley m. , b. mcinnes , and f. smith , 1981 , q. j. astr .
, * 22 * , 272 .
harvey j. w. , 1972 , nature , * 235 * , 90 .
harvey j. w. , and j. b. breckinridge , 1973 , astrophys .
j. , * 182 * , l137
. hawley s. a. , and j. s. miller , 1977 , astrophys .
j , * 212 * , 94 .
hege e. , e. hubbard , p. strittmatter , and s. worden , 1981 , astrophys .
j. , * 248 * , 1 .
hestroffer d. , 1997 , astron .
astrophys . ,
* 327 * , 199 .
heydari m. , and j. beuzit , 1994 , astron .
astrophys . ,
* 287 * , l17 .
hickson p. , 2001
, private communication .
hill j. m. , 2000 , spie , * 4004 * , 36 .
hinz p. , r. angel , w. hoffmann , d. mccarthy , p. mcguire , m. cheselka , j. hora , and n. woolf , 1998 , nature , * 395 * , 251 .
hinz p. , w. hoffmann , and j. hora , 2001 , astrophys .
( to appear ) .
the hipparcos catalogue , 1997 , esa , sp-1200 .
hofmann k. -h . , w. seggewiss , and g. weigelt , 1995 , astron .
astrophys . , * 300 * , 403 . hgbom j. , 1974 , astron .
suppl . , * 15 * , 417 .
hu w. , 2001 , astro - ph/0105117 hummel c. , d. mozurkevich , j. armstrong , a. hajian , n. elias , and d. hutter , 1998 , astron . j. , * 116 * , 2536 .
hutchings j. , d. crampton , s. morris , d. durand , and e. steinbring , 1999 , astron .
j. , * 117 * , 1109 .
hutchings j. , d. crampton , s. morris , and e. steinbring , 1998 , pub .
pac , * 110 * , 374 .
hutchings j. , s. morris , and d. crampton , 2001 , astron .
j. , * 121 * , 80 .
ishimaru a. , 1978 , ` wave propagation and scattering in random media ' , academic press , n. y. jefferies s. , and j. christou , 1993 , astrophys .
j , * 415 * , 862 .
jennison r. c. , 1958 , mon . not .
r. astron .
soc . , * 118 * , 276 . karovska m. , l. koechlin , p. nisenson , c. papaliolios , and c. standley , 1989 , astrophys .
j , * 340 * , 435 .
kenworthy m. et al . , 2001 ,
( to appear ) .
kervella p. , v. coud du foresto , g. perrin , m. schller , w. traub , and m. lacasse , 2001 , astron .
astrophys . ,
* 367 * , 876 .
knox k. , and b. thompson , 1974 , astrophys .
j , * 193 * , l45 .
koechlin l. , p. r. lawson , d. mourard , a. blazit , d. bonneau , f. morand , p. stee , i. tallon - bosc , and f. vakili , 1996 , appl .
, * 35 * , 3002 .
kolmogorov a. , 1941a , in ` turbulence ' , eds .
, s. k. friedlander & l. topper , 1961 , wiley - interscience , n. y. , 151 .
kolmogorov a. , 1941b , in ` turbulence ' , eds .
, s. k. friedlander & l. topper , 1961 , wiley - interscience , n. y. , 156 .
kolmogorov a. , 1941c , in ` turbulence ' , eds .
, s. k. friedlander & l. topper , 1961 , wiley - interscience , n. y. , 159 .
korff d. , 1973 , j. opt .
* 63 * , 971 .
kunz m. , a. banday , p. castro , p. ferreira , and k. grski , 2001 , astrophys .
j. lett . , ( to appear ) .
labeyrie a. , 1970 , astron .
astrophys . ,
* 6 * , 85 .
labeyrie a. , 1975 , astrophys .
j , * 196 * , l71 .
labeyrie a. , 1995 , astron .
astrophys . , * 298 * , 544 . labeyrie a. , 1996 , astron .
suppl . , * 118 * , 517 . labeyrie a. , 1998 , spie , * 3350 * , 960
. labeyrie a. , 1999a , asp conf .
, * 194 * , eds .
s. unwin , and r. stachnik , 350 .
labeyrie a. , 1999b , science , * 285 * , 1864 .
labeyrie a. , 2000 , private communication .
labeyrie a. , 2001 , private communication .
labeyrie a. , l. koechlin , d. bonneau , a. blazit , and r. foy , 1977 , astrophys . j. , * 218 * , l75 .
labeyrie a. , g. lamaitre , and l. koechlin , 1986 , spie , * 628 * , 323 .
labeyrie a. , g. schumacher , m. dugu , c. thom , p. bourlon , f. foy , d. bonneau , and r. foy , 1986 , astron .
astrophys . ,
* 162 * , 359 .
lai o. , d. rouan , f. rigaut , r. arsenault , and e. gendron , 1998 , astron .
astrophys . ,
* 334 * , 783 .
lai o. , d. rouan , f. rigaut , f. doyon , and f. lacombe , 1999 , astron .
astrophys . ,
* 351 * , 834 .
lane b. , m. kuchner , a. boden , m. crooch - eakman , and s. r. kulkarni , 2000 , nature , * 407 * , 485 .
lannes a. , e. anterrieu , and k. bouyoucef , 1994 , j. mod .
, * 41 * , 1537 .
lawson p. r. , 1994 , pub .
, * 106 * , 917 .
lawson p. r. , 1995 , j. opt .
am a. , * 12 * , 306 .
lawson p. , j. baldwin , p. warner , r. boysen , c. haniff , j. rogers , d. saint - jacques , d. wilson , and j. young , 1998 , spie , * 3350 * , 753 .
lawson p. , m. colavita , p. dumont , and b. lane , 2000 , spie , * 4006 * , 397 .
ledoux c. , b. thodore , p. petitjean , m. n. bremer , g. f. lewis , r. a. ibata , m. j. irwin , and e. j. totten , 1998 , astron .
astrophys . ,
* 339 * , l77 . lee j. , b. bigelow , d. walker , a. doel , and r. bingham , 2000 , pub .
pac , * 112 * , 97 .
lefvre h. c. , 1980 , electron .
* 16 * , 778 .
lger a. , m. pirre , and f. j. marceau , 1993 , astron .
astrophys . ,
* 277 * , 309 .
liang j. , d. r. williams , and d. t. miller , 1997 , j. opt .
a , * 14 * , 2884 .
lindengren l. , and m. a. c. perryman , 1996 , astron .
suppl . , * 116 * , 579
. linfield r. , and p. gorham , 1999 , asp conf .
, * 194 * , eds .
s. unwin , and r. stachnik , isbn : 1 - 58381 - 020-x , 224 .
lloyd - hart m. , 2000 , pub .
pac , * 112 * , 264 .
lloyd - hart m. , j. r. angel , t. groesbeck , t. martinez , b. jacobsen , b. mcleod , d. mccarthy , e. hooper , e. hege , and d. sandler , 1998 , astrophys . j. , * 493 * , 950 .
lipman e. , m. bester , w. danchi , and c. townes , 1998 , spie , * 3350 * , 933 .
liu y. c. , and a. w. lohmann , 1973 , opt .
* 8 * , 372 .
lohmann a. , g. weigelt , and b. wirnitzer , 1983 , appl .
, * 22 * , 4028 .
lopez b. et al . , 1997 ,
j. , * 488 * , 807 .
lopez b. , 1991 , ` last mission at la silla , april 19 @xmath197 may 8 , on the measure of the wavefront evolution velocity ' , eso internal report
. lovelock j. e. , 1965 , nature , * 207 * , 568 .
lucy l. , 1974 , astron .
j. , * 79 * , 745 .
lynds c. , s. worden , and j. harvey , 1976 , astrophys .
j , * 207 * , 174 .
machida y. et al . , 1998 ,
spie , * 3350 * , 202 .
macintosh b. , c. max , b. zuckerman , e. becklin , d. kaisler , p. lowrence , a. weinberger , j. christou , g. schneider , and s. acton , 2001 , astro - ph/0106479 .
mackay c. d. , r. n. tubbs , r. bell , d. burt , p. jerram , and i. moody , 2001 , spie , * 4306 * , ( in press ) .
magain p. , f. courbin , and s. sohy , 1998 , astrophys .
j. , * 494 * , 472 .
malbet f. et al . , 1998 ,
j. , * 507 * , l149 .
marco o. , t. encrenaz , and e. gendron , 1997 , planet sp .
, * 46 * , 547 .
mrquez i. , p. petitjean , b. thodore , m. bremer , g. monnet , and j. beuzit , 2001 , astron .
astrophys . , * 371 * , 97
. masciadri e. , j. vernin , and p. bougeault , 1999 , astron .
* 137 * , 203 .
mason b. d. , 1996 , astron .
j. , * 112 * , 2260 .
mason b. , c. martin , w. i. hartkopf , d. barry , m. germain , g. douglass , c. worley , g. wycoff , t. ten brummelaar , and o. franz , 1999 , astron . j. , * 117 * , 1890 . mayor m. , and d. queloz , 1995 , nature , * 378 * , 355 .
mcalister h. , w. bagnuolo , t. ten brummelaar , w. i. hartkopf , m. shure , l. sturmann , n. turner , and s. ridgway , 1998 , spie , * 3350 * , 947 .
mendel l. , and e. wolf , 1995 , ` optical coherence and quantum optics ' , cambridge university press , cambridge .
mennesson b. , j. -m .
mariotti , v. coud du foresto , g. perrin , s. ridgway , c. ruilier , w. traub , m. lacasse , and g. maz , 1999 , astron .
astrophys . , * 346 * , 181
. menshchikov a. , and t. henning , 1997 , astron .
astrophys . ,
* 318 * , 879 .
michelson a. a. , 1891 , nature , * 45 * , 160 .
michelson a. , and f. pease , 1921 , astrophys .
j , * 53 * , 249 .
millan - gabet r. , p. schloerb , and w. traub , 2001 , astrophys .
j. , * 546 * , 358 . monnier j. , w. danchi , d. hale , p. tuthill , and c. townes , 2000 , astrophys .
j , * 543 * , 868 .
monnier j. , et al .
, 2001 , aas meeting , * 198 * , 63.02 .
monnier j. , p. tuthill , b. lopez , p. cruzalbes , w. danchi , c. haniff , 1999 , astrophys .
j , * 512 * , 351 .
morel s. , 2000 , private communication .
morel s. , and l. koechlin , 1998 , spie , * 3350 * , 257 .
morel s. , w. traub , j. bregman , r. mah , and c. wilson , 2000 , spie , * 4006 * , 506 .
mouillet d. , j. d. larwood , j. c. papaloizou , and a. m. lagrange , 1997 , mon . not .
r. astron .
, * 292 * , 896 .
mourard d. , d. bonneau , l. koechlin , a. labeyrie , f. morand , p. stee , i. tallon - bosc , and f. vakili , 1997 , astron .
astrophys . , * 317 * , 789 . mourard d. , i. bosc , a. labeyrie , l. koechlin , and s. saha , 1989 , nature , * 342 * , 520
. mozurkewich d. , k. johnston , r. simon , d. hutter , m. colavita , m. shao , and x. pan , 1991 , astron .
j. , * 101 * , 2207 .
nakajima t. , 1994 , astrophys j. , * 425 * , 348 .
nakajima t. , and d. golimowski , 1995 , astron .
j. , * 109 * , 1181 .
natta a. , t. prusti , r. neri , d. wooden , and v. grinin , 2001 , astron .
astrophys . ,
* 371 * , 186 .
nelkin m. , 2000 , am .
j. phys . ,
* 68 * , 310 .
nisenson p. , and c. papaliolios , 1999 , astrophys .
j , * 518 * , l29 .
nisenson p. , c. papaliolios , m. karovska , and r. noyes , 1987 , astrophys .
j , * 320 * , l15 .
nordgren t. , j. armstrong , m. gierman , r. hindsley , a. hajian , j. sudol , and c. hummel , 2000 , astrophys .
j. , * 543 * , 972 . northcott m. j. , g. r. ayers , and j. c. dainty , 1988 , j. opt .
a , * 5 * , 986 .
nota a. , c. leitherer , m. clampin , p. greenfield , and d. a. golimowski , 1992 , astron . j. , * 398 * , 621 .
nulsen p. , p. wood , p. gillingham , m. bessel , m. dopita , and c. mccowage , 1990 , astrophys .
j , * 358 * , 266 .
osterbart r. , y. balega , g. weigelt , and n. langer , 1996 , proc .
180 , eds .
, h. habing & g. lamers , kluwer academic pub .
netherlands , 362 .
padilla c. , v. karlov , l. matson , k. soosaar , and t. ten brummelaar , 1998 , spie , * 3350 * , 1045 .
papaliolios c. , m. karovska , l. koechlin , p. nisenson , c. standley , and s. heathcote , 1989 , nature , * 338 * , 565 .
papaliolios c. , p. nisenson , and s. ebstein , 1985 , appl .
, * 24 * , 287 .
pauls t. , d. mozurkewich , j. armstrong , c. hummel , j. benson , and a. hajian , 1998 , spie , * 3350 * , 467 .
paxman r. , t. schulz , and j. fienup , 1992 , j. opt .
* 9 * , 1072 .
paxman r. , j. seldin , m. lfdahl , g. scharmer , and c. keller , 1996 , astrophys .
j. , * 466 * , 1087 .
pedretti e. , and a. labeyrie , 1999 , astron .
suppl . , * 137 * , 543 .
pedretti e. , a. labeyrie , l. arnold , n. thureau , o. lardire , a. boccaletti , and p. riaud , 2000 , astron .
suppl . , * 147 * , 285 .
pehlemann e. , k. -h hofmann , and g. weigelt , 1992 , astron .
astrophys . , * 256 * , 701 .
penny a. , a. lger , j. mariotti , c. schalinski , c. eiora , r. laurance , m. fridlund , 1998 , spie , * 3350 * , 666 .
perrin g. , 1997 , astron .
* 121 * , 553 .
perrin g. , v. coud du foresto , s. ridgway , j. -m .
marrioti , w. traub , n. carleton , m. lacasse , 1998 , astron .
astrophys . ,
* 331 * , 619 .
perrin g. , v. coud du foresto , s. ridgway , b. mennesson , c. ruilier , j. -m .
marrioti , w. traub , and m. lacasse , 1999 , astron .
astrophys . ,
* 345 * , 221 .
perryman m. a. c. , 1998 , nature , * 340 * , 111 . poulet f. , and b. sicardy , 1996 , bull .
soc . , * 28 * , 1124 .
pourbaix d. , 2000 , astron .
suppl . , * 145 * , 215 . prieur j. , e. oblak , p. lampens , m. kurpinska - winiarska , e. aristidi , l. koechlin , and g. ruymaekers , 2001 , astron .
astrophys . , * 367 * , 865
. puetter r. , and a. yahil , 1999 , astro - ph/9901063 .
quirrenbach a. , 2001 , annu .
astrophys . ,
* 39 * , 353 .
quirrenbach a. , d. mozurkewich , d. buscher , c. hummel , and j. armstrong , 1996 , astron .
astrophys . , * 312 * , 160 .
quirrenbach a. , j. roberts , k. fidkowski , w. de vries , and w. van breugel , 2001 , astrophys .
j. ( to appear ) .
rabbia y. , d. mekarnia , and j. gay , 1990 , spie , * 1341 * , 172 .
racine r. , 1984 , iau colloq .
79 , eds . ,
m. ulrich & kjr , 235 .
racine r. , g. herriot , and r. mcclure , 1996 , proc .
` adaptive optics ' ed . , m. cullum , eso , germany , 35 .
ragazzoni r. , and d. bonaccini , 1996 , proc . ` adaptive optics ' , ed . , m. cullum , 17 .
ragazzoni r. , e. marchetti , and g. valente , 2000 , nature , * 403 * , 54 .
reinheimer t. , and g. weigelt , 1987 , astron .
astrophys . , * 176 * , l17 .
richardson w. h. , 1972 , j. opt .
* 62 * , 55 .
ridgway s. t. , and f. roddier , 2000 , spie , * 4006 * , 940 .
rimmele t. r. , 2000 , spie , * 4007 * , 218 .
robbe s. , b. sorrente , f. cassaing , y. rabbia , and g. rousset , 1997 , astron .
. suppl . * 125 * , 367 .
robertson n. a. , 2000 , class .
grav . , * 17 * , 19 .
roddier c. , and f. roddier , 1988 , proc .
nato - asi workshop , eds .
, d. m. alloin & j. -m .
mariotti , 221 .
roddier c. , f. roddier , m. j. northcott , j. e. graves , and k. jim , 1996 , astrophys .
j , * 463 * , 326 .
roddier f. , 1981 , progress in optics , * xix * , 281 .
roddier f. , 1988 , phys . rep .
, * 170 * , 97 .
roddier f. , 1999 , ` adaptive optics in astronomy ' , ed .
, f. roddier , cambridge univ . press .
roddier f. , c. roddier , a. brahic , c. dumas , j. graves , m. northcott , and t. owen , 1997 , planet sp .
, * 45 * , 1031 . roddier f. , c. roddier , j. e. graves , and m. j. northcott , 1995 , astrophys
. j , * 443 * , 249 .
roggemann m. c. , b. m. welsh , and r. q. fugate , 1997 , rev .
modern phys . ,
* 69 * , 437 .
rouan d. , d. field , j. -l .
lemaire , o. lai , g. p. forts , e. falgarone , and j. -m .
deltorn , 1997 , mon . not .
r. astron .
soc . , * 284 * , 395 . rouan d. , p. riaud , a. boccaletti , y. clnet , and a. labeyrie , 2000 , pub .
pac . , * 112 * , 1479 .
rouan d. , f. rigaut , d. alloin , r. doyon , o. lai , d. crampton , e. gendron , and r. arsenault , 1998 , astron .
astrophys . , * 339 * , 687 . rousset g. , 1999 , ` adaptive optics in astronomy ' , ed .
, f. roddier , cambridge univ .
press , 91 .
rousset g. , j. c. fontanella , p. kem , p. gigan , f. rigaut , p. lna , p. boyer , p. jagourel , j. p. gaffard , and f. merkle , 1990 , astron .
astrophys . ,
* 230 * , l29 .
rousselet - perraut k. , f. vakili , and d. mourard , 1996 , opt .
* 35 * , 2943 .
ryan s. , and p. wood , 1995 , pub .
austr . , * 12 * , 89 . saha s. k. , 1999 , bull
ind . , * 27 * , 443 . saha s. k. , and v. chinnappan , 1999 , bull
ind . , * 27 * , 327 . saha s. k. , and d. maitra , 2001 , ind
. j. phys . , * 75b * , 391 . saha s. k. , and s. morel , 2000 , bull .
ind . , * 28 * , 175 . saha s. k. , r. rajamohan , p. vivekananda rao , g. som sunder , r. swaminathan , and b. lokanadham , 1997 , bull
ind . , * 25 * , 563 . saha s. k. , r. sridharan , and k. sankarasubramanian , 1999 , ` speckle image reconstruction of binary stars ' , presented at xix asi meeting , bangalore . saha s. k. , and p. venkatakrishnan , 1997 , bull
ind . , * 25 * , 329
. sams b. j. , k. schuster , and b. brandl , 1996 , astrophys .
j. , * 459 * , 491 .
sato k. et al . , 1998 ,
spie , * 3350 * , 212 .
schertl d. , k. -h .
hofmann , w. seggewiss , and g. weigelt , 1996 , astron .
astrophys . , * 302 * , 327 . schller m. , w. brandner , t. lehmann , g. weigelt , and h. zinnecker , 1996 , astron .
astrophys . , * 315 * , 445 . seldin j. , r. paxman , and c. keller , 1996 , spie . , * 2804 * , 166 . seldin j. , and r. paxman , 1994 , spie . , * 2302 * , 268
. serabyn e. , 2000 , spie , * 4006 * , 328 .
shannon c. j. , 1949 , proc .
ire , * 37 * , 10 .
shao m. , and m. m. colavita , 1992 , astron .
astrophys . , * 262 * , 353 . shao m. , and m. m. colavita , 1994 , proc .
158 , eds .
, j. g. robertson and w. j. tango , 413 .
shao m. , and d. staelin , 1977 , j. opt .
am . , * 67 * , 81 . shao m. et al . , 1988 , astron
. astrophys . , * 193 * , 357 . sicardy b. , f. roddier , c. roddier , e. perozzi , j. e. graves , o. guyon , and m. j. northcott , 1999 , nature , * 400 * , 731 .
simon m. , l. close , and t. beck , 1999 , astron .
j. , * 117 * , 1375 .
stee p. , de arajo , f. vakili , d. mourard , i. arnold , d. bonneau , f. morand , and i. tallon - bosc , 1995 , astron .
astrophys . , * 300 * , 219 .
stee p. , f. vakili , d. bonneau , and d. mourard , 1998 , astron .
astrophys . ,
* 332 * , 268 .
tallon m. , r. foy , and a. blazit , 1988 , proc .
eso conf . ed .
ulrich , eso , frg , 743 .
tango w. j. , and r. q. twiss , 1980 , prog .
, * 17 * , 239 .
tatarski v. i. , 1967 , ` wave propagation in a turbulent medium ' , dover , new york .
tatarski v. i. , 1993 , j. opt .
a , * 56 * , 1380 .
taylor g. l. , 1921 , in ` turbulence ' , eds .
, s. k. friedlander & l. topper , 1961 , wiley - interscience , new york , 1 .
thom c. , p. granes , and f. vakili , 1986 , astron .
astrophys . ,
* 165 * , l13 .
thompson l. a. , and c. s. gardner , 1988 , nature , * 328 * , 229 .
timothy j. g. , 1993 , spie . , * 1982
torres g. , r. stefanik , and d. latham , 1997 , astrophys .
j , * 485 * , 167 .
townes c. h. , m. bester , w. danchi , d. hale , j. monnier , e. lipman , a. everett , p. tuthill , m. johnson , and d. walters , 1998 , spie , * 3350 * , 908 .
traub w. a. , 1986 , appl .
, * 25 * , 528 .
traub w. a. , 2000 , course notes on ` principles of long baseline interferometry ' , ed .
, p. r. lawson , 31 .
traub w. a. et al . , 2000 , spie , * 4006 * , 715 .
traub w. a. , r. millam - gabet , and m. garcia , 1998 , bull .
soc . , * 193 * , 52.06 .
troxel s. e. , b. m. welsh , and m. c. roggemann , 1994 , j. opt .
a , * 11 * , 2100 .
tuthill p. , c. haniff , and j. baldwin , 1997 , mon . not .
r. astron .
* 285 * , 529 .
tuthill p. , c. haniff , and j. baldwin , 1999 , mon . not .
r. astron .
* 306 * , 353 .
tuthill p. g. , j. d. monnier , and w. c. danchi , 1999 , nature , * 398 * , 487 .
tuthill p. g. , j. d. monnier , and w. c. danchi , 2001 , nature , * 409 * , 1012 .
tuthill p. g. , j. d. monnier , w. c. danchi , and wishnow , 2000 , pub .
, * 116 * , 2536 .
ulrich m. , l. maraschi , and c. urry , 1997 , annu .
astrophys . ,
* 35 * , 445 .
unwin s. , s. turyshev , and m. shao , 1998 , spie , * 3350 * , 551 .
vakili f. , d. mourard , d. bonneau , f. morand , and p. stee , 1997 , astron .
astrophys . , * 323 * , 183 . vakili f. , d. mourard , p. stee , d. bonneau , p. berio , o. chesneau , n. thureau , f. morand , a. labeyrie , and i. tallon - bosc , 1998 , astron .
astrophys . ,
* 335 * , 261 .
van belle g. et al .
, 1999 , astron .
j. , * 117 * , 521 .
van belle g. , d. ciardi , r. thompson , r. akeson , and e. a. lada , 2001 , astro - ph/0106184 .
van belle g. , h. dyck , j. benson , and m. lacasse , 1996 , astron .
j. , * 112 * , 2147 .
van belle g. , h. dyck , r. thomson , j. benson , and s. kannappan , 1997 , astron .
j. , * 114 * , 2150 .
von der lhe o. , 1984 , j. opt .
a , * 1 * , 510 .
walkup j. f. , and j. w. goodman , 1973 , j. opt .
am , * 63 * , 399 .
wallace j. et al . , 1998 ,
spie , * 3350 * , 864 .
wehinger p. a. , 2001 , private communication .
weigelt g. , 1977 , opt .
communication , * 21 * , 55 . weigelt g. , and g. bair , 1985 , astron .
astrophys . ,
* 150 * , l18 .
weigelt g. , y. balega , t. blcker , a. fleischer , r. osterbart , and j. winters , 1998 , astron .
, * 333 * , l51 .
weigelt g. , y. balega , k. -h .
hofmann , and m. scholz , 1996 , astron .
astrophys . ,
* 316 * , l21 .
weigelt g. , and j. ebersberger , 1986 , astron .
astrophys . ,
* 163 * , l5 .
weinberger a. , g. neugebauer , and k. matthews , 1999 , astron .
j. , * 117 * , 2748 .
wilken v. , c. r. de boer , c. denker , and f. kneer , 1997 , astron .
astrophys . ,
* 325 * , 819 .
wilson r. , j. baldwin , d. busher , and p. warner , 1992 , mon . not .
r. astron .
* 257 * , 369 .
wittkowski m. , y. balega , t. beckert , w. duschi , k. hofmann , and g. weigelt , 1998 , astron .
astrophys . ,
* 329 * , l45 .
wittkowski m. , y. balega , k. hofmann , and g. weigelt , 1999 , mitteilungen der astronomischen gesellschaft ( agm ) .
, * 15 * , 83 .
wittkowski m. , c. hummel , k. johnston , d. mozurkewich , a. hajian , and n. white , 2001 , astron .
astrophys . ,
* 377 * , 981 .
wittkowski m. , n. langer , and g. weigelt , 1998 , astron .
, * 340 * , l39 .
wood p. , m. bessel , and m. dopita , 1986 , astrophys .
j , * 311 * , 632 .
wood p. , p. nulsen , p. gillingham , m. bessel , m. dopita , and c. mccowage , 1989 , astrophys .
j , * 339 * , 1073 .
worden s. p. , c. r. lynds , and j. w. harvey , 1976 , j. opt .
am . , * 66 * , 1243 .
wyngaard j. c. , y. izumi , and s. a. collins , 1971 , j. opt . soc .
am . , * 60 * , 1495 .
young a. t. , 1974 , astrophys .
j. , * 189 * , 587 .
young j. et al . , 2000 ,
mon . not .
r. astron .
soc . , * 315 * , 635 .
zago l. , 1995 , http://www.eso.org/gen-fac/pubs/astclim/lz-thesis/node4-html . | the present ` state of the art ' and the path to future progress in high spatial resolution imaging interferometry is reviewed .
the review begins with a treatment of the fundamentals of stellar optical interferometry , the origin , properties , optical effects of turbulence in the earth s atmosphere , the passive methods that are applied on a single telescope to overcome atmospheric image degradation such as speckle interferometry , and various other techniques .
these topics include differential speckle interferometry , speckle spectroscopy and polarimetry , phase diversity , wavefront shearing interferometry , phase - closure methods , dark speckle imaging , as well as the limitations imposed by the detectors on the performance of speckle imaging .
a brief account is given of the technological innovation of adaptive - optics ( ao ) to compensate such atmospheric effects on the image in real time .
a major advancement involves the transition from single - aperture to the dilute - aperture interferometry using multiple telescopes .
therefore , the review deals with recent developments involving ground - based , and space - based optical arrays . emphasis is placed on the problems specific to delay - lines , beam recombination , polarization , dispersion , fringe - tracking , bootstrapping , coherencing and cophasing , and recovery of the visibility functions .
the role of ao in enhancing visibilities is also discussed .
the applications of interferometry , such as imaging , astrometry , and nulling are described .
the mathematical intricacies of the various ` post - detection ' image - processing techniques are examined critically .
the review concludes with a discussion of the astrophysical importance and the perspectives of interferometry .
# 10=-.025em0 - 0 .05em0 - 0 -.025em.0433em0 |
let @xmath4 be a transcendental entire function , and denote by @xmath5 , the @xmath6th iterate of @xmath0 .
the _ fatou set _
@xmath7 is defined to be the set of points @xmath8 such that @xmath9 forms a normal family in some neighborhood of @xmath10 .
the complement of @xmath7 is called the _
julia set _
@xmath11 of @xmath0 .
an introduction to the properties of these sets can be found in @xcite .
this paper concerns the _ escaping set _ of @xmath0 , defined as follows : @xmath12 for a transcendental entire function @xmath0 , the escaping set was first studied by eremenko @xcite who proved that @xmath13 and that @xmath14 much of the work on @xmath15 has been motivated by eremenko s conjecture @xcite that every component of @xmath15 is unbounded .
significant progress was made on eremenko s conjecture in @xcite where it was shown that @xmath15 always has at least one unbounded component .
this result was proved by considering the _ fast escaping set _
@xmath1 , which was introduced by bergweiler and hinkkanen in @xcite and which now plays a key role in transcendental dynamics . in this paper
we define @xmath16 here , @xmath17 @xmath18 denotes iteration of @xmath19 with respect to the variable @xmath20 , and @xmath21 can be taken to be any value such that @xmath22 for @xmath23 or , equivalently , such that @xmath24 as @xmath25 . for simplicity , we will only write down this restriction on @xmath26 in formal statements of results elsewhere this should be assumed to be true .
note that @xmath22 whenever @xmath27 ; otherwise , by montel s thorem , we would have @xmath28 .
we prove the basic properties of @xmath1 in section 2 these include the fact that the set @xmath1 is independent of the choice of @xmath26 .
note that different definitions of @xmath1 were used in @xcite , @xcite , @xcite , @xcite , @xcite and @xcite .
we show that all these definitions are equivalent in section 2 .
the definition used here is of a similar form to the definition of the fast escaping set in a direct tract used in @xcite and @xcite .
we now introduce certain subsets of @xmath1 based on the above definition .
let @xmath0 be a transcendental entire function , let @xmath29 and let @xmath21 be such that @xmath22 for @xmath23 .
the @xmath30th _ level _ of @xmath1 ( with respect to @xmath26 ) is the set @xmath31 we also put @xmath32 note that , unlike the sets @xmath15 and @xmath1 , each of the levels of @xmath1 is a _ closed _ set .
also , since @xmath33 for all @xmath34 , we have @xmath35 for all @xmath36 .
this implies that @xmath1 can be written as an expanding union of closed sets : @xmath37 in this paper we show that this concept of the levels of @xmath1 leads to significant new results concerning the properties of @xmath1 .
further , it leads to considerable improvements to many of the results in our earlier papers @xcite , @xcite and @xcite .
this paper is written in such a way that it provides a self - contained account of the main properties of @xmath1 , so some proofs of known results are included . in almost all cases ,
the proofs here are more straightforward than the original ones .
in @xcite we showed that all the components of @xmath1 are unbounded .
this is the strongest result for general entire functions that has been obtained towards eremenko s conjecture as , for any transcendental entire function , @xmath38 ( see theorem [ e ] ) and so , as stated earlier , there is at least one unbounded component of @xmath15 . here
we need the following more precise version of this result .
[ main1 ] let @xmath0 be a transcendental entire function and let @xmath21 be such that @xmath22 for @xmath23 .
then , for each @xmath29 , each component of @xmath39 is closed and unbounded ; in particular , each component of @xmath1 is unbounded . in section 3 of this paper
we give two proofs of theorem [ main1 ] .
first we outline a constructive proof similar to the proof of ( * ? ? ?
* theorem 1 ) , the ideas of which have had significant applications ( see @xcite , @xcite and @xcite ) . then we give a new shorter proof by contradiction . in section 4
, we consider fatou components that meet @xmath1 .
we prove the following , which is a key result of the paper and is used to deduce many of the other results .
given its many applications , it has a surprisingly simple proof which consists of combining the new concept of the levels of @xmath1 with known distortion properties of the iterates of a function in a fatou component .
[ main2 ] let @xmath0 be a transcendental entire function , let @xmath21 be such that @xmath22 for @xmath23 , and let @xmath29 . if @xmath2 is a fatou component that meets @xmath39 , then * @xmath40 ; * if , in addition , @xmath2 is simply connected , then @xmath41 .
_ note that theorem [ main2 ] implies that , if @xmath2 is a fatou component in @xmath1 , then the boundary of @xmath2 is also in @xmath1 .
this contrasts with the situation of a fatou component in @xmath15 which may have boundary points that are not in @xmath15 .
for example , if @xmath42 , then @xmath7 is connected and is in @xmath15 @xcite .
in fact , for this function , @xmath7 is a baker domain and so is contained in the slow escaping set @xcite .
since the boundary of @xmath7 is always equal to @xmath11 , the boundary of this baker domain contains many points whose behaviour is very different under iteration , such as fast escaping points and periodic points .
we conclude section 4 with an example which shows that it is not possible to replace @xmath43 by @xmath30 in theorem [ main2 ] part ( a ) . in section 5
, we consider the relationship between @xmath1 and @xmath11 .
the main result in this section is the following .
[ main3 ] let @xmath0 be a transcendental entire function , let @xmath21 be such that @xmath22 for @xmath23 , and let @xmath29 .
all the components of @xmath44 are unbounded if and only if @xmath0 has no multiply connected fatou components .
it follows immediately from theorem [ main3 ] that if @xmath0 has no multiply connected fatou components , then _ all _ the components of @xmath45 are unbounded . in @xcite
we used a relatively complicated argument to prove the weaker result that if @xmath0 has no multiply connected fatou components , then there is at least one unbounded component of @xmath46 .
here , theorem [ main3 ] follows from theorem [ main2 ] and theorem [ main1 ] in a straightforward manner , showing the value of introducing the concept of the levels of @xmath1 . in sections 68 ,
we consider the case when @xmath1 has a structure which we now describe as an ` infinite spider s web ' .
the first examples of functions for which @xmath1 has this structure were given in ( * ? ? ?
* theorem 2 ) .
it has become apparent that there are many transcendental entire functions for which @xmath1 has this structure and that such functions have a number of strong dynamical properties .
because of the increasing significance of this structure , we introduce the following terminology . a set @xmath47 is an _
( infinite ) spider s web _ if @xmath47 is connected and there exists a sequence of bounded simply connected domains @xmath48 with @xmath49 , for @xmath50 , @xmath51 , for @xmath50 and @xmath52 .
_ if @xmath15 is a spider s web , then @xmath15 is connected ( and unbounded ) and so eremenko s conjecture holds .
+ in ( * ? ? ?
* theorem 2 ) we showed that , if @xmath0 is a transcendental entire function , then @xmath1 and @xmath15 are spiders webs whenever @xmath0 has a multiply connected fatou component .
in fact @xmath53 is also a spider s web for such a function .
this is a corollary of the following more general result , proved in section 6 .
a version of theorem [ main4 ] was proved in @xcite .
[ main4 ] let @xmath0 be a transcendental entire function and let @xmath21 be such that @xmath22 for @xmath23 .
if @xmath54 has a bounded component , then each of @xmath53 , @xmath1 and @xmath15 is a spider s web .
it follows from theorem [ main4 ] that if @xmath53 is a spider s web , then so are @xmath1 and @xmath15 .
we now know that there are many functions for which @xmath53 is a spider s web ; see theorem [ main8 ] .
in fact all of the functions for which we know that @xmath15 is a spider s web have the stronger property that @xmath53 is a spider s web .
an approximation to @xmath53 for the function @xmath55 is shown in figure 1 .
( this function was mentioned in ( * ? ? ?
* section 6 ) . )
this structure of @xmath53 is novel , and clearly very different to the familiar structure of a cantor bouquet , which arises when describing the escaping sets of many entire functions . in this diagram ,
the set of black points is an approximation to @xmath53 , for some @xmath21 .
the large grey region on the right is the immediate basin of an attracting fixed point , and the fine structure of @xmath53 around this basin extends to the rest of the plane , though the computer can not handle the arithmetic at most of these points .
the range shown is given by @xmath56 .
spider s web , width=302 ] in section 7 , we show that functions for which @xmath53 is a spider s web have many strong dynamical properties . we begin by proving the following result which follows easily from theorem [ main2 ] .
part ( a ) is new and a version of part ( b ) was proved in @xcite , though the proof given here is more straightforward .
let @xmath0 be a transcendental entire function , let @xmath21 be such that @xmath22 for @xmath23 , and let @xmath53 be a spider s web .
* if @xmath0 has no multiply connected fatou components , then each of @xmath57 is a spider s web . *
the function @xmath0 has no unbounded fatou components .
part ( b ) of theorem [ main6 ] allows us to make a connection between the existence of an @xmath53 spider s web and a conjecture of baker about functions of small growth ; see section 8 .
we also use the concept of the levels of @xmath1 to prove the following result about the complement of @xmath1 .
[ holes ] let @xmath0 be a transcendental entire function , let @xmath21 be such that @xmath22 for @xmath23 , and let @xmath53 be a spider s web . *
all the components of @xmath58 are compact . *
every point of @xmath11 is the limit of a sequence of points , each of which lies in a distinct component of @xmath58 .
the property of @xmath1 proved in part ( a ) of this result contrasts with the fact that all the components of @xmath59 , @xmath29 , are open .
this result demonstrates that , if @xmath53 is a spider s web , then @xmath1 has a very intricate structure .
further results about the intricate nature of @xmath1 when @xmath53 is a spider s web have been obtained by osborne @xcite ; for example , the result in part ( b ) can be strengthened to state that singleton components of @xmath58 are dense in @xmath11 .
we end section 7 by proving the following result whose proof also uses the concept of the levels of @xmath1 .
[ main7 ] let @xmath0 be a transcendental entire function , let @xmath21 be such that @xmath22 for @xmath23 , and let @xmath53 be a spider s web .
* each point in @xmath15 belongs to an unbounded continuum in @xmath15 on which all points escape to infinity uniformly . * if @xmath60 is a component of @xmath58 , then either @xmath61 or all points in @xmath60 escape to infinity uniformly .
_ theorem [ main7 ] part ( a ) answers a question asked by rempe in @xcite .
it is shown in @xcite that the conclusion of this part holds for many functions in the eremenko - lyubich class @xmath62 ( defined in section 3 ) .
rempe @xcite has shown the existence of a transcendental entire function in the class @xmath62 for which the conclusion of theorem [ main7 ] part ( a ) does not hold .
these results show that , if @xmath53 is a spider s web , then @xmath0 has many interesting dynamical properties . in section 8 ,
we investigate which families of functions have these properties .
previously it was thought that it is relatively unusual for the escaping set to have this structure .
indeed , the following result shows that many commonly studied entire functions do _ not _ have an @xmath53 spider s web .
[ sw2 ] let @xmath0 be a transcendental entire function , let @xmath21 be such that @xmath22 for @xmath23 , and let @xmath53 be a spider s web .
then there is no path to infinity on which @xmath0 is bounded and so * @xmath0 does not belong to the class @xmath62 ; * @xmath0 has no exceptional points ( that is , points with a finite backwards orbit ) .
however , we show that there are in fact many functions for which @xmath53 is a spider s web . thus , for all these functions , the sets @xmath53 , @xmath1 and @xmath15 are connected , and @xmath0 has no unbounded fatou components , and so both eremenko s conjecture and baker s conjecture hold .
[ main8 ] let @xmath0 be a transcendental entire function and let @xmath21 be such that @xmath22 for @xmath23 .
then @xmath53 is a spider s web if one of the following holds : * @xmath0 has a multiply connected fatou component ; * @xmath0 has very small growth ; that is , there exist @xmath63 and @xmath64 such that @xmath65 where @xmath66 is the @xmath67th iterated logarithm ; * @xmath0 has order less than @xmath68 and regular growth ; * @xmath0 has finite order , fabry gaps and regular growth ; * @xmath0 has a sufficiently strong version of the pits effect and has regular growth .
parts ( a ) to ( c ) follow from results in @xcite and @xcite .
parts ( d ) and ( e ) are new
there are many functions which belong to these classes .
( we will define fabry gaps , regular growth and the pits effect in section 8 . )
we conclude section 8 by showing that @xmath53 is a spider s web if and only if @xmath69 is a spider s web .
this enables us to construct many new examples of functions for which @xmath53 is a spider s web , by considering iterates of the functions described in theorem [ main8 ] .
in particular , this makes it easy to construct examples of such functions for which the order is infinite .
_ finally , many of the new results about @xmath1 obtained here can be generalised in some form to transcendental meromorphic functions with direct tracts ; see @xcite for results concerning the fast escaping set in a direct tract .
in this section we prove the basic properties of the fast escaping set and also show that there are several equivalent definitions of this set .
let @xmath0 be a transcendental entire function and let @xmath70 .
recall that the fast escaping set @xmath1 is defined in terms of the maximum modulus function @xmath19 . for simplicity , from now on we will write this as @xmath71 provided that it is clear from the context which function @xmath0 is being considered .
we often use the following facts about @xmath71 without comment : @xmath72 @xmath73 many results in this paper use the following property of the maximum modulus function , which was proved in this form in ( * ? ? ?
* lemma 2.2 ) .
[ convex ] let @xmath0 be a transcendental entire function .
then there exists @xmath21 such that , for all @xmath74 and all @xmath75 , @xmath76 it follows from lemma [ convex ] together with that , if @xmath77 , then @xmath78 now let @xmath21 be such that @xmath22 for @xmath23 . then @xmath79 recall that @xmath80 also , for each @xmath29 , @xmath81 and the fast escaping set is @xmath82 the following set relations follow immediately from the definition of @xmath83 and are often used without comment : @xmath84 @xmath85 we now give the following elementary result concerning @xmath1 .
parts ( a)(c ) were stated in @xcite and part ( d ) was proved in @xcite . [
basic ] let @xmath0 be a transcendental entire function , let @xmath21 be such that @xmath22 for @xmath23 , and let @xmath86 .
then * @xmath1 is completely invariant under @xmath0 ; * @xmath1 is independent of @xmath26 ; * @xmath87 ; * if @xmath88 , where @xmath89 , @xmath90 , then @xmath91 . _ remark . _
the set @xmath92 , which consists of the points that ` zip ' to infinity , was studied in @xcite .
( some authors denote this set by @xmath93 . ) _ ( a ) _ the complete invariance of @xmath1 under @xmath0 follows directly from and .
_ ( b ) _ we take @xmath94
. clearly @xmath95 , for @xmath29 , and so @xmath96 now note that , by , there exists @xmath97 such that @xmath98 and so @xmath99 together with , these set relations show that @xmath1 is independent of @xmath26 . _
( c ) _ to prove that @xmath100 , we use the fact that , since @xmath0 is a transcendental entire function , it follows from and that @xmath101 as @xmath25 .
thus , for each @xmath102 , there exists @xmath103 such that @xmath104 for all @xmath105 and @xmath106 .
so , for @xmath107 , @xmath108 , @xmath109 and hence , if @xmath110 , then @xmath111 the result now follows since we can choose @xmath112 to be arbitrarily large . _
( d ) _ this property follows immediately from the equivalent definition of @xmath1 as the set @xmath113 given in @xcite ; see corollary [ equal ] .
it was shown in @xcite that eremenko s construction in @xcite of points in @xmath15 actually gives points that are in @xmath1 .
in fact his construction , which is based on wiman - valiron theory , can readily be adapted to give points that are in @xmath1 and also have other strong dynamical properties .
we summarise the main properties of such points in the following result .
points @xmath114 with these properties are particularly useful and we often refer to them as _ eremenko points_. [ e ] let @xmath0 be a transcendental entire function and let @xmath115 .
there exists @xmath21 such that , if @xmath116 , then there exists @xmath117 with @xmath118 further , we can choose @xmath114 such that there exist sequences @xmath119 and @xmath120 and sets @xmath121 such that * for @xmath34 , @xmath122 * for @xmath50 , @xmath123 * @xmath124 , where @xmath125 and , for @xmath50 , @xmath126 is a component of @xmath127 ; * @xmath128 is a positive increasing sequence with @xmath129 as @xmath25 .
_ by modifying the definitions of the discs and annuli in theorem [ e ] appropriately , each disc @xmath130 can be chosen to lie in a thin annulus @xmath131 contained in and concentric with @xmath132 . for any @xmath133
, the modulus of the annulus @xmath131 can be chosen to be less than @xmath134 for @xmath6 sufficiently large .
note that @xmath130 can be allowed to lie in @xmath135 but , with this modification , the point @xmath114 does not necessarily satisfy @xmath136 , for @xmath50 .
+ we now show how eremenko points can be used to demonstrate that the definitions of the fast escaping set used in earlier papers are equivalent to the definition used in this paper .
the original definition of the fast escaping set was given by bergweiler and hinkkanen in @xcite .
for clarity , we refer to the set given by their definition as @xmath137 until we have proved that it is equal to the set @xmath1 defined earlier : @xmath138 where @xmath26 is any value such that @xmath139 . in @xcite we showed that @xmath137 is equal to the set @xmath140 where @xmath141 is any open disc meeting @xmath11 , and @xmath142 denotes the union of @xmath2 and its bounded complementary components .
we now show that all of these definitions of the fast escaping set are equivalent . in order to do this ,
we prove the following result .
[ subsets ] let @xmath0 be a transcendental entire function and let @xmath143 .
if @xmath21 is sufficiently large , then @xmath144 the only set relation that is not clear in the above list is the first one . to prove this
, we note that , by theorem [ e ] , if @xmath26 is sufficiently large , then @xmath145 contains a neighbourhood of an eremenko point .
the result now follows from the properties of such a point described in parts ( a ) and ( c ) of theorem [ e ] .
the equivalence of the various definitions of the fast escaping set follows immediately from lemma [ subsets ] . note that the last set relation in lemma [ subsets ] is required since the definition of @xmath137 involves a strong inequality .
[ equal ] let @xmath0 be a transcendental entire function . then @xmath146 .
_ remarks . _ 1 .
the proof of corollary [ equal ] is much simpler than the proof that @xmath147 given in @xcite .
\2 . in @xcite
we based the proofs of several results about @xmath1 and @xmath15 on the set @xmath148 where @xmath141 is any open disc meeting @xmath11 .
this set is analogous to the set @xmath53 but we do not have @xmath149 ( compare with ) and so more complicated proofs are required when working with @xmath150 instead of @xmath53 .
+ we now use lemma [ subsets ] to prove the following result . [ eq ]
let @xmath0 be a transcendental entire function and let @xmath97 .
if @xmath21 is sufficiently large , then @xmath151 and hence @xmath152 . let @xmath153 be such that @xmath22 for @xmath23 , and let @xmath97 . if @xmath154 , then @xmath155 so @xmath156 and hence @xmath157 .
thus @xmath158 and hence @xmath159 .
conversely , suppose that @xmath21 is sufficiently large .
then , by lemma [ subsets ] , @xmath160 if @xmath157 , then @xmath161 so , by , @xmath162 this implies that @xmath163 for all @xmath50 and hence @xmath164 . thus @xmath165 and hence @xmath166 .
we conclude this section by giving a condition that is sufficient to ensure that certain points are in @xmath1 . in forthcoming work
@xcite we give a weaker condition that is sufficient to ensure that points are in @xmath1 provided that @xmath0 has sufficiently regular growth .
[ q(f ) ] let @xmath0 be a transcendental entire function and , for @xmath115 , @xmath70 , let @xmath167 .
then @xmath168 where @xmath21 is sufficiently large to ensure that @xmath169 , for @xmath74 . by hypothesis , @xmath170 ( the existence of such an @xmath26 follows from . )
further , by , there exists @xmath171 such that , for @xmath172 , @xmath173 it follows from that there exists @xmath174 such that @xmath175 .
so , by , @xmath176 so , if there exists @xmath108 such that @xmath177 , for @xmath50 , then @xmath178 and hence @xmath179
. this result can be applied , for example , to the function @xmath180 to show that , for @xmath26 sufficiently large , @xmath181 \subset a(f)$ ] , since @xmath182 ) \subset ( -\infty , -r]$ ] and @xmath183
in this section we show that each component of each of the levels of @xmath1 is unbounded and hence all the components of @xmath1 are unbounded .
this is sufficient to prove theorem [ main1 ] since each of the levels is a closed set and hence each component of each of the levels must also be closed .
we give two proofs of theorem [ main1 ] .
the first is an outline of the constructive proof that all the components of @xmath1 are unbounded that we gave in @xcite and the second is a new proof by contradiction .
let @xmath184 , for some @xmath29 .
then , for all @xmath50 with @xmath185 , @xmath186 say . now let @xmath187 denote the component of @xmath188 that contains @xmath189 .
then @xmath187 is closed and also unbounded , since @xmath190 is entire .
moreover , @xmath191 indeed @xmath192 implies that @xmath193 so @xmath194 . hence @xmath195 is a closed connected subset of @xmath196 which contains @xmath189 and @xmath197 . now let @xmath198 be the component of @xmath199 which contains @xmath189 .
then @xmath198 is closed in @xmath200 and unbounded ; see @xcite .
finally , we note that @xmath201 since , if @xmath202 , then @xmath193 and so @xmath203 suppose that there exists a bounded component @xmath60 of @xmath83 , for some @xmath29 .
then , since @xmath83 is closed , there exists a jordan curve @xmath204 that surrounds @xmath60 ; see @xcite .
for each @xmath50 with @xmath205 , we put @xmath206 since @xmath207 , it follows from the maximum principle that @xmath208 .
further , for @xmath50 with @xmath205 , @xmath209 is closed and @xmath210 .
thus @xmath211 this , however , is a contradiction since @xmath212
in this section we consider which types of fatou components can meet @xmath1 and prove various results about such components .
in particular , we prove theorem [ main2 ] .
we begin this section with a distortion lemma that we use to prove various properties of @xmath1 .
the original proof of this lemma is due to baker ( * ? ? ?
* lemmas 1 and 2 ) and a good account of the result is also given in ( * ? ? ?
* lemma 7 ) .
[ dist ] let @xmath0 be a transcendental entire function , let @xmath213 be a fatou component and let @xmath60 be a compact subset of @xmath2 .
* there exist @xmath214 , @xmath103 such that @xmath215 * if , in addition , @xmath2 is simply connected , then there exist @xmath214 , @xmath103 such that @xmath216 the following corollary of lemma [ dist ] was proved in @xcite .
let @xmath0 be a transcendental entire function and let @xmath2 be a fatou component with @xmath217 .
then @xmath2 is a wandering domain .
let @xmath213 be a periodic fatou component of period @xmath218 , say , and let @xmath219 .
it follows from lemma [ dist ] part ( a ) applied to the entire function @xmath220 with @xmath221 that there exist @xmath214 , @xmath103 such that , for each @xmath105 , @xmath222 and hence @xmath223 .
thus , by theorem [ basic ] part ( c ) , @xmath224 .
there are various types of wandering domains that are known to lie in @xmath1 .
in particular , the following result due to baker ( * ? ? ? * theorem 3.1 ) can be used to show that @xmath2 belongs to @xmath1 if @xmath2 is multiply connected .
[ baker ] let @xmath0 be a transcendental entire function and let @xmath2 be a multiply connected fatou component .
then * each @xmath225 is bounded , * @xmath226 surrounds @xmath225 for large @xmath6 , * @xmath227 as @xmath228 .
_ note that if @xmath2 is a bounded fatou component of a transcendental entire function @xmath0 then , for @xmath50 , @xmath225 is also a fatou component of @xmath0 ; see @xcite .
we showed in @xcite that , for any multiply connected fatou component @xmath2 we have @xmath229 .
the following more precise result holds .
[ baker1 ] let @xmath0 be a transcendental entire function and let @xmath2 be a multiply connected fatou component .
then , for some @xmath103 and @xmath21 , we have @xmath230 and hence @xmath229 .
suppose that @xmath143 , where @xmath26 is sufficiently large to apply lemma [ subsets ] , and that @xmath2 is a multiply connected fatou component .
then , by lemma [ baker ] , there exists @xmath103 such that @xmath231 and @xmath232 then @xmath233 so , by and , @xmath234 hence , by lemma [ subsets ] , @xmath235 so @xmath236 as required .
alternatively , theorem [ baker1 ] can be proved by using the following result due to zheng @xcite .
[ zheng ] let @xmath0 be a transcendental entire function . if @xmath0 has a multiply connected fatou component @xmath2 , then there exist annuli @xmath237 such that @xmath238 and @xmath239 as @xmath228 .
_ remarks . _ 1 .
zheng proved a version of this result for the larger class of transcendental meromorphic functions with at most finitely many poles , where @xmath2 is a wandering domain with the properties listed in lemma [ baker ] .
\2 . in @xcite
( see also @xcite ) we show that lemma [ zheng ] can be strengthened to give much larger annuli inside multiply connected fatou components .
+ it follows from lemma [ zheng ] and theorem [ e ] that , if @xmath2 is a multiply connected fatou component of a transcendental entire function @xmath0 and @xmath6 is sufficiently large , then @xmath225 contains an eremenko point .
together with theorem [ e ] part ( a ) and lemma [ baker ] , this is sufficient to prove theorem [ baker1 ] .
there are also examples of simply connected fatou components in @xmath1 .
bergweiler @xcite has constructed an example of a transcendental entire function with simply connected bounded wandering domains that lie between multiply connected wandering domains , and whose closures are in @xmath1 .
these appear to be the only known examples of fatou components in @xmath1 that are not multiply connected .
this suggests the following question . +
* question 1 * can there be any unbounded fatou components in @xmath1 ?
+ for the examples of fatou components that meet @xmath1 described above , it is the case that the closure of the fatou component is also in @xmath1 .
we now show that this property is true in general by proving theorem [ main2 ] .
let @xmath2 be a fatou component and let @xmath240 , where @xmath241 .
then @xmath213 , by normality , and latexmath:[\[\label{u1 } _ ( a ) _ to prove that @xmath40 , we suppose that @xmath243 .
it follows from lemma [ dist ] part ( a ) , and properties of the maximum modulus function that there exist @xmath214 and @xmath103 such that @xmath244 thus @xmath245 . since @xmath246 is a closed set , this implies that @xmath40 as claimed .
_ ( b ) _ now suppose that @xmath2 is simply connected .
to prove that @xmath247 , we use proof by contradiction .
suppose that there exists @xmath243 with @xmath248 .
then there exist @xmath103 and @xmath75 such that @xmath249 say . since @xmath250 , we can assume that @xmath251 as @xmath25 .
together with , this implies that , for @xmath50 , @xmath252 this , however , contradicts lemma [ dist ] part ( b ) since , by lemma [ convex ] , @xmath253 thus @xmath254 .
since @xmath83 is a closed set , this implies that @xmath247 .
the next result follows immediately from theorem [ main2 ] part ( b ) , and will be used later in the paper .
let @xmath0 be a transcendental entire function , let @xmath21 be such that @xmath22 for @xmath23 , and let @xmath29 . if all the components of @xmath7 are simply connected , then @xmath255 . _ remark .
_ it follows from montel s theorem that , if @xmath0 is a transcendental entire function and @xmath29 , then all the interior points of @xmath83 are in @xmath7 . + we conclude this section by showing that the conclusion of corollary [ ajf ] can fail if there is a multiply connected fatou component .
suppose that @xmath0 is a transcendental entire function and that @xmath2 is a multiply connected component of @xmath7 .
then , by lemma [ zheng ] and theorem [ e ] , there exist @xmath103 and @xmath21 such that @xmath256 and such that @xmath257 contains an eremenko point @xmath189 with @xmath258 , for @xmath50 . since @xmath259 and , by , @xmath260 it follows that @xmath261 and so @xmath262 .
this example also shows that , in theorem [ main2 ] part ( a ) , @xmath246 can not be replaced by @xmath83 .
in this section we consider points in @xmath1 that also belong to the julia set and prove theorem [ main3 ] .
we begin by showing that @xmath1 has properties corresponding to the properties of @xmath15 proved by eremenko @xcite that we listed in .
proofs of parts of this result appeared in @xcite and in @xcite .
the proof of part ( a ) given here is particularly straightforward .
[ ej ] let @xmath0 be a transcendental entire function . then * @xmath263 ; * @xmath264 ; * @xmath265 . _
( a ) _ we saw in theorem [ e ] that eremenko s construction shows that there are always points in @xmath1 .
further , if such a point is in a component @xmath2 of @xmath7 , then it follows from theorem [ main2 ] that @xmath266 .
since @xmath267 is also in @xmath11 , this shows that @xmath263 .
_ ( b ) _ we begin by noting that it follows from theorem [ main2 ] that @xmath268 .
we now show that @xmath269 , by following the argument used by eremenko to show that @xmath270 . first note that @xmath1 is completely invariant and so , for @xmath50 , @xmath271 and hence @xmath272 since @xmath1 contains infinitely many points , it follows from montel s theorem that @xmath273 and so @xmath274 .
since @xmath1 contains no periodic points , all interior points of @xmath1 must belong to @xmath7 .
thus @xmath269 and hence @xmath264 as claimed . _
( c ) _ similarly , since @xmath45 is completely invariant and contains infinitely many points , @xmath275 .
clearly , since @xmath11 is closed , @xmath276 and so @xmath277 .
we saw in the previous section that there exist transcendental entire functions @xmath0 such that @xmath278 .
there are many functions , however , for which @xmath279 .
in particular , this is the case for functions in the eremenko - lyubich class @xmath280 where the set @xmath281 consists of the critical values and the finite asymptotic values of @xmath0 ; indeed it was shown by eremenko and lyubich @xcite that @xmath282 if @xmath283 .
for many functions in the class @xmath62 , the escaping set consists of a family of curves to infinity ; see @xcite and @xcite . in @xcite we showed that , in many cases , @xmath1 consists of these curves with the exception of some of the endpoints .
in particular , this is the case if @xmath0 is a finite composition of functions of finite order in the class @xmath62 .
note that , for such functions , it is possible for both @xmath1 and @xmath15 to be connected .
we now show that this is the case for the function @xmath284 this function has order 1 and is in the class @xmath62 since its only singular values are the critical values @xmath285 and @xmath286 , which are taken at odd and even multiples of @xmath287 respectively .
both singular values are in @xmath15 so @xmath288 by ( * ? ? ?
* theorems 7 and 12 ) .
this function has the properties that @xmath0 is @xmath289-periodic , @xmath290 , for @xmath70 , and @xmath291 $ ] , for @xmath292 .
so , if we set @xmath293 , then @xmath294 for @xmath295 and , for @xmath296 , @xmath297 @xmath298 @xmath299 thus @xmath300 clearly @xmath47 is connected and contains @xmath301 , for every @xmath302 .
we now note that @xmath0 maps each half - strip of the form @xmath303 univalently onto the cut plane @xmath304 , with each vertex of @xmath305 mapping to @xmath286 and @xmath306 mapping to @xmath285 .
further , @xmath0 has a continuous extension to @xmath307 , mapping it onto @xmath308 .
therefore , for each @xmath296 , the set @xmath309 forms a connected subset of @xmath310 and @xmath311 , so @xmath312 is connected . repeating this argument
gives that @xmath313 is a connected subset of @xmath1 which contains the backwards orbit of any point in @xmath314 .
since @xmath288 and the backwards orbit of any non - exceptional point is dense in @xmath11 , it follows that @xmath315 .
thus @xmath316 and hence both @xmath1 and @xmath15 are connected . in figure 2 ,
the set of black points is an approximation to part of the set @xmath1 for this function , consisting of curves that are successive pre - images under @xmath0 of the positive real axis .
the range shown is given by @xmath317 .
the original motivation behind this example was to find a function @xmath0 for which there is an unbounded continuum meeting only finitely many levels of @xmath1 the imaginary axis is such an unbounded continuum when @xmath318 . for @xmath319,width=302 ]
there are now known to be many examples of functions in the class @xmath62 for which @xmath1 and @xmath15 are both connected .
for example , rempe @xcite has shown that @xmath15 is connected if @xmath320 , where @xmath321 , and it follows from his proof that @xmath1 is also connected for these functions .
these results have recently been extended to larger classes of exponential functions in @xcite and @xcite .
we now prove the main result of this section .
we must show that all the components of @xmath44 are unbounded if and only if @xmath0 has no multiply connected fatou components .
we begin by noting that if @xmath2 is a multiply connected fatou component of @xmath0 , then it follows from lemma [ baker ] that there are no unbounded components of @xmath11 .
now suppose that @xmath0 has no multiply connected fatou components and let @xmath322 . we know from theorem [ main1 ] that @xmath189 belongs to a closed unbounded component @xmath198 of @xmath83 .
further , it follows from corollary [ ajf ] and the remark following it that @xmath323 we now let @xmath112 denote the component of @xmath324 containing @xmath189 and show that , if @xmath112 is bounded , then we get a contradiction .
so suppose that @xmath112 is bounded .
since @xmath324 is closed , there exists a jordan curve @xmath325 that surrounds @xmath112 ; see @xcite .
since @xmath198 is unbounded and connected , and @xmath326 , we have @xmath327 . since @xmath328
, we have @xmath329 and so , by , @xmath330 . also , by , @xmath331 and so @xmath332 belongs to a multiply connected fatou component .
we have assumed , however , that there are no such components .
thus @xmath112 must in fact be unbounded . since @xmath333 this completes the proof .
in the final three sections , we consider the case when @xmath1 is connected and , moreover , is a spider s web .
recall from the introduction that a connected set @xmath47 is a spider s web if there exists a sequence of bounded simply connected domains @xmath48 with @xmath49 , @xmath51 , for @xmath50 , and @xmath334 . in this section
we prove theorem [ main4 ] which states that , if @xmath21 is such that @xmath22 for @xmath23 , and @xmath54 has a bounded component , then each of @xmath53 , @xmath1 and @xmath15 is a spider s web .
suppose that there exists a bounded component @xmath335 of @xmath54
. then @xmath335 is simply connected ( since all the components of @xmath53 are unbounded by theorem [ main1 ] ) .
further , @xmath336 , since @xmath53 is closed . now let @xmath337 .
then , for each @xmath50 , it follows from and that @xmath338 further , if @xmath6 is sufficiently large , then @xmath339 must surround @xmath285 .
otherwise , for all @xmath50 , we would have @xmath340 thus contradicting the fact that @xmath341 .
since @xmath342 for @xmath50 , it follows from that there exists a sequence @xmath343 such that @xmath344 , @xmath345 and @xmath346 .
since each component of @xmath53 is unbounded , we see that each component of @xmath53 will meet @xmath347 for @xmath348 sufficiently large .
thus @xmath53 must be connected and indeed must be a spider s web .
similarly , since each component of @xmath1 is unbounded , @xmath1 must be a spider s web .
it remains to show that @xmath15 is a spider s web .
first note that @xmath1 is connected and is therefore contained in a component @xmath349 of @xmath15 .
further , @xmath349 is a spider s web .
we now show that @xmath350 .
first suppose that @xmath351 .
it follows from lemma [ ej ] part ( b ) that @xmath352 .
thus @xmath353 is connected .
since @xmath354 and @xmath349 is a component of @xmath15 , it follows that @xmath355 .
thus @xmath356 to complete the proof we show that @xmath357 .
first , we show that , if @xmath358 is a fatou component in @xmath15 , then @xmath359 . we can assume that @xmath360 lies in a simply connected component of @xmath7 for each @xmath50 since otherwise @xmath361 by theorem [ baker1 ] . if @xmath362 , then the claim is proved . otherwise , there exist @xmath363 , @xmath97 and a sequence @xmath343 such that @xmath364 , for @xmath365 , and @xmath366 .
since @xmath367 , there exists @xmath365 such that @xmath368 .
since @xmath369 lies in a simply connected component of @xmath7 , there are no curves in @xmath369 surrounding points in @xmath11 .
so @xmath370 thus @xmath371 .
since @xmath372 , we have @xmath359 by the complete invariance of @xmath15 .
so , whenever @xmath358 is a fatou component in @xmath15 , we have @xmath373 . since @xmath374
, it follows from that @xmath375 .
now @xmath376 is a connected subset of @xmath15 that meets @xmath349 and so @xmath377 . hence @xmath378 .
together with , this shows that @xmath350 and so @xmath15 is a spider s web .
the next result follows easily from theorem [ main4 ] and proves theorem [ main8 ] part ( a ) .
[ mc ] let @xmath0 be a transcendental entire function and let @xmath21 be such that @xmath22 for @xmath23 . if @xmath0 has a multiply connected fatou component , then each of @xmath53 , @xmath1 and @xmath15 is a spider s web .
if @xmath0 has a multiply connected fatou component @xmath2 , then it follows from theorem [ baker1 ] that , for @xmath6 sufficiently large , @xmath379 .
thus , by lemma [ baker ] , @xmath54 has a bounded component .
the result then follows from theorem [ main4 ] .
we give more examples of functions for which @xmath53 , @xmath1 and @xmath15 are spiders webs in section 8 .
in fact , all the examples of functions for which we know that either @xmath1 or @xmath15 is a spider s web also have the property that @xmath53 is a spider s web .
this suggests the following questions .
+ * question 2 * can @xmath1 be a spider s web when @xmath53 is not a spider s web ? + * question 3 * can @xmath15 be a spider s web when @xmath1 is not a spider s web ? +
in this section we show that , if @xmath53 is a spider s web , then several strong results about the structure of the escaping set hold , namely theorem [ main6 ] , theorem [ holes ] and theorem [ main7 ] .
we begin by proving some basic properties of such spiders webs .
[ sw1 ] let @xmath0 be a transcendental entire function , let @xmath21 be such that @xmath22 for @xmath23 , and let @xmath29 .
* if @xmath335 is a bounded component of @xmath59 , then @xmath380 and @xmath190 is a proper map of @xmath335 onto a bounded component of @xmath381 , for each @xmath50 . * if @xmath59 has a bounded component , then @xmath83 is a spider s web and hence every component of @xmath59 is bounded .
* @xmath53 is a spider s web if and only if @xmath83 is a spider s web .
* let @xmath94 .
then @xmath53 is a spider s web if and only if @xmath382 is a spider s web . _
( a ) _ let @xmath335 be a bounded component of @xmath59
. since @xmath83 is closed , @xmath380 .
thus , for each @xmath50 , @xmath383 and @xmath384 .
it follows that @xmath385 , so @xmath190 is a proper map of @xmath335 onto @xmath386 and @xmath386 is a bounded component of @xmath387 .
_ ( b ) _ let @xmath335 be a bounded component of @xmath59 .
then , as in part ( a ) , @xmath388 and , for @xmath6 sufficiently large , @xmath389 surrounds @xmath285 , since @xmath383 .
since each component of @xmath83 is unbounded by theorem [ main1 ] , it now follows ( as in the proof of theorem [ main4 ] ) that @xmath83 is a spider s web .
the final assertion of part ( b ) follows from the definition of a spider s web . _
( c ) _ suppose that @xmath53 is a spider s web . if @xmath390 then any component of @xmath59 is contained in a component of @xmath54 and is therefore bounded .
so , by part ( b ) , @xmath83 is a spider s web . if @xmath391 and @xmath335 is a component of @xmath54 then , by part ( a ) , @xmath392 is a bounded component of @xmath59 and so , by part ( b ) , @xmath83 is a spider s web .
the converse result follows by similar arguments . _
( d ) _ suppose that @xmath382 is a spider s web .
since @xmath94 , each component of @xmath393 is contained in a component of @xmath394 , so it follows from part ( b ) that @xmath395 is also a spider s web .
now suppose that @xmath395 is a spider s web .
there exists @xmath396 such that @xmath397 and so each component of @xmath394 is contained in a component of @xmath398 .
it then follows from parts ( b ) and ( c ) that @xmath382 is also a spider s web . in order to prove the main results of this section we introduce some terminology .
let @xmath0 be a transcendental entire function and let @xmath21 be such that @xmath22 for @xmath23 .
if @xmath53 is a spider s web then , for each @xmath34 , * @xmath399 denotes the component of @xmath400 that contains 0 ; * @xmath401 .
we say that @xmath402 is the _ sequence of fundamental holes _ for @xmath53 and @xmath403 is the _ sequence of fundamental loops _ for @xmath53 .
before proving the main results of this section , we prove some preliminary results about these holes and loops . [ sw3 ] let @xmath0 be a transcendental entire function and let @xmath21 be such that @xmath22 for @xmath23 .
suppose that @xmath53 is a spider s web and @xmath404 and @xmath405 are the sequences of fundamental holes and loops for @xmath53 .
then * @xmath406 and @xmath407 , for @xmath34 ; * @xmath408 , for @xmath34 ; * for @xmath50 and @xmath409 , @xmath410 * there exists @xmath103 such that , for @xmath105 and @xmath409 , @xmath411 * if @xmath29 and @xmath335 is a component of @xmath59 , then @xmath412 and @xmath413 , for @xmath6 sufficiently large ; * if there are no multiply connected fatou components , then @xmath414 for @xmath34 . _
( a ) _ this follows from and lemma [ sw1 ] part ( a ) .
_ ( b ) _ let @xmath34 .
we know that @xmath415 .
also , @xmath399 contains @xmath285 .
so , by definition , @xmath416 . _
( c ) _ let @xmath50 and @xmath409 .
first note that @xmath417 and that @xmath418 , so @xmath419 .
the result now follows from lemma [ sw1 ] part ( a ) . _
( d ) _ since @xmath420 is bounded , it is clear that there exists @xmath103 such that @xmath421 and hence @xmath422 for @xmath105 .
the result now follows from part ( c ) . _
if @xmath29 and @xmath335 is a component of @xmath59 , then it follows from lemma [ sw1 ] part ( a ) that , for each @xmath50 , @xmath386 is a component of @xmath381 .
thus , for each @xmath50 , either @xmath412 or @xmath386 lies outside @xmath423 , in which case @xmath424 since @xmath425 , must fail for large values of @xmath6 .
thus , if @xmath6 is sufficiently large , @xmath412 and so @xmath413 , by lemma [ sw1 ] part ( a ) . _
( f ) _ note that , for each @xmath34 , @xmath426 .
if there are no multiply connected fatou components , then it follows from corollary [ ajf ] that , for each @xmath34 , @xmath427 and so @xmath414 .
we are now in a position to prove the main results of this section .
let @xmath0 be a transcendental entire function , let @xmath21 be such that @xmath22 for @xmath23 , and suppose that @xmath53 is a spider s web . _
( a ) _ if @xmath0 has no multiply connected fatou components , then it follows from lemma [ sw3 ] part ( f ) that @xmath428 where @xmath405 is the sequence of fundamental loops for @xmath53 . since each component of @xmath429 is unbounded , by theorem [ main3 ]
, this implies that @xmath429 is a spider s web .
similarly , @xmath45 must be a spider s web .
it follows from theorem [ ej ] part ( c ) that @xmath430 thus each of @xmath431 and @xmath11 is also connected and moreover , by , is a spider s web .
_ ( b ) _ we must show that @xmath0 has no unbounded fatou components .
if @xmath0 has a multiply connected fatou component , then this follows from lemma [ baker ] .
if @xmath0 has no multiply connected fatou components , then the result follows from part ( a ) . next we prove theorem [ holes ] which concerns the components of @xmath58 .
let @xmath0 be a transcendental entire function , let @xmath21 be such that @xmath22 for @xmath23 , and suppose that @xmath53 is a spider s web . _
( a ) _ let @xmath60 be a component of @xmath58 .
we must show that @xmath60 is compact .
for each @xmath108 , we let @xmath432 denote the component of @xmath433 that contains @xmath60 . clearly
moreover , it follows from lemma [ sw3 ] parts ( d ) and ( e ) that there exists @xmath103 such that @xmath435 for @xmath108 , and so @xmath436 thus @xmath437 is a compact connected set in @xmath58 .
since @xmath438 and @xmath60 is a component of @xmath58 , it follows that @xmath439 and that @xmath60 is compact .
_ ( b ) _ we now show that every point @xmath440 is the limit of a sequence of points , each of which lies in a distinct component of @xmath58 .
we begin by considering the case that @xmath441 .
we know from theorem [ ej ] part ( b ) that @xmath264 and so there exists a sequence @xmath442 of points in @xmath58 with @xmath443 .
if the points @xmath444 belong to only finitely many components of @xmath58 then , by part ( a ) , @xmath189 also belongs to one of these components and is therefore in @xmath58 .
this , however , is a contradiction .
so the result holds if @xmath441 .
we now consider the case that @xmath445 .
we know from theorem [ ej ] part ( c ) that @xmath265 .
so there exists a sequence @xmath446 of points in @xmath45 with @xmath447 .
we know that the result holds for each point @xmath448 and this is sufficient to show that the result also holds for @xmath189 . we now prove theorem [ main7 ] , which gives various subsets of @xmath15 on which the iterates of @xmath0 escape to infinity uniformly .
let @xmath0 be a transcendental entire function , let @xmath21 be such that @xmath22 for @xmath23 , and suppose that @xmath53 is a spider s web .
let @xmath354 .
we show that @xmath189 belongs to an unbounded continuum in @xmath15 on which all points escape to infinity uniformly .
if @xmath184 , for some @xmath29 , then we know from theorem [ main1 ] that @xmath189 belongs to an unbounded component of @xmath83 . clearly , all points in this component escape to infinity uniformly .
now suppose that @xmath449 .
we let @xmath60 denote the component of @xmath58 containing @xmath189 and , for @xmath108 , let @xmath432 denote the component of @xmath433 that contains @xmath60 , and let @xmath450 .
note that @xmath451 is connected .
it follows from and that @xmath452 we now let @xmath453 note that @xmath454 we claim that @xmath455 is an unbounded continuum in @xmath15 on which all points escape to infinity uniformly .
we begin by showing that @xmath112 is an unbounded continuum .
first , it follows from lemma [ sw1 ] part ( c ) that @xmath456 is a spider s web and hence an unbounded continuum , for each @xmath108 .
in particular , @xmath457 is an unbounded continuum .
we now claim that @xmath458 is connected , for @xmath108 .
this is the case because a separation of @xmath458 would lead to a separation of @xmath459 , since @xmath460 and @xmath461 .
but there can be no separation of @xmath459 since this set is connected .
further , it follows from that @xmath462 is connected , for each @xmath108 .
thus @xmath463 is unbounded and connected .
it follows that @xmath464 is connected because any open set that contains @xmath60 must also contain a set @xmath450 for some @xmath108 , by and , and so meet @xmath465 .
thus @xmath466 is an unbounded continuum as claimed .
we now show that all the points in @xmath112 escape to infinity uniformly .
let @xmath404 and @xmath405 denote the sequences of fundamental holes and loops for @xmath53 . recall that @xmath467 and so , for each @xmath97 , there exists @xmath468 such that @xmath469
we claim that @xmath470 and so all points in @xmath471 escape to infinity uniformly . to prove this , we let @xmath97 and @xmath472 and consider the images of the sets @xmath432 for @xmath108 .
we begin by noting that it follows from lemma [ sw1 ] part ( a ) and that , for @xmath108 , @xmath473 also @xmath474 by the second relation in and the first in we have @xmath475 also , since @xmath476 , we have @xmath477 and so it follows from and that @xmath478 because @xmath479 and @xmath480 is contained in a component of @xmath481 different from @xmath482 for such @xmath30 . since @xmath483 , it follows from that @xmath484 since @xmath485 , it also follows from that @xmath486 together with and , this is sufficient to show that holds . to conclude the proof
, we note that all points escape to infinity uniformly on @xmath487 .
together with , this implies that all points escape to infinity uniformly on @xmath466 .
we begin this section by proving theorem [ sw2 ] which shows in particular that , if @xmath0 belongs to the widely studied eremenko - lyubich class @xmath62 , then @xmath53 is _ not _ a spider s web .
if @xmath283 , then there must be a path to @xmath197 on which @xmath490 is constant ( see @xcite ) , so @xmath53 is not a spider s web .
finally , if @xmath0 has an exceptional value @xmath491 , then @xmath492 , for some @xmath493 and some entire function @xmath494 ( see @xcite ) , so @xmath495 is a transcendental meromorphic function with one pole , at @xmath491 .
hence @xmath496 has asymptotic value @xmath197 by iversen s theorem ( see , for example , @xcite ) , so @xmath0 has asymptotic value @xmath491 and is therefore bounded on a path to @xmath197 .
the rest of this section is devoted to showing that there are large classes of functions for which @xmath53 is a spider s web whenever @xmath22 for @xmath497 ; in particular , we prove parts ( b)(e ) of theorem [ main8 ] .
( recall that we proved theorem [ main8 ] part ( a ) at the end of section 6 . )
we obtain these examples by using the following result .
[ gamman ] let @xmath0 be a transcendental entire function and let @xmath21 be such that @xmath22 for @xmath23 .
then @xmath53 is a spider s web if and only if there exists a sequence @xmath498 of bounded simply connected domains such that , for all @xmath34 , @xmath499 and @xmath500 now suppose that there exists a sequence @xmath502 of bounded simply connected domains satisfying and , and let @xmath335 denote the component of @xmath54 containing @xmath503
. we will show that @xmath335 is bounded and hence , by lemma [ sw1 ] part ( b ) , that @xmath53 is a spider s web . in order to do this
, we suppose that @xmath335 is not contained in @xmath504 and obtain a contradiction . for simplicity
, we put @xmath505 , for @xmath34 .
if @xmath335 is not a subset of @xmath504 , then there exists a compact connected set @xmath60 such that @xmath506 it follows from and that @xmath507 is a compact connected set with @xmath508 thus , by , @xmath509 using and repeatedly in this way , we deduce that , for each @xmath50 , @xmath510 since , for each @xmath50 , @xmath511 is compact and @xmath512 , we have @xmath513 .
it follows from that @xmath514 and from that @xmath515 this is a contradiction and so our original supposition that @xmath335 is not a subset of @xmath504 must have been incorrect .
thus @xmath516 and is therefore bounded as required .
[ m(r ) ] let @xmath0 be a transcendental entire function and let @xmath21 be such that @xmath22 for @xmath23 . then @xmath53 is a spider s web if there exists a sequence @xmath518 such that , for @xmath34 , @xmath519 and @xmath520 much work has been done on establishing when the conditions of this corollary are satisfied in connection with a conjecture of baker that functions of small growth have no unbounded fatou components .
( for further details on the history of this problem see the survey article @xcite . )
the strongest results on this problem are given in @xcite and @xcite .
recall from theorem [ main6 ] part ( b ) that if @xmath53 is a spider s web , then @xmath0 has no unbounded fatou components .
thus , if @xmath53 is a spider s web , then both eremenko s conjecture ( that all the components of @xmath15 are unbounded ) and baker s conjecture hold .
it follows from the discussion at the end of the first section of @xcite that the conditions of corollary [ m(r ) ] are satisfied if @xmath0 is a transcendental entire function and there exist @xmath63 and @xmath64 such that @xmath521 here @xmath66 denotes the @xmath67th iterated logarithm function .
thus , for such functions , @xmath53 is a spider s web . as discussed in (
* section 6 ) , there are many examples of such functions for which there are no multiply connected fatou components .
firstly , bergweiler and eremenko showed in @xcite ( see also @xcite ) that there are functions of arbitrarily small growth ( and hence functions satisfying ) for which the julia set is the whole plane .
further , baker @xcite and boyd @xcite independently showed that there are functions of arbitrarily small growth ( and hence functions satisfying ) for which every point in the fatou set tends to @xmath285 under iteration .
we now define the order and lower order of a transcendental entire function since these concepts play a key role in discussing further examples of functions for which @xmath53 is a spider s web .
the _ order of growth _ of a transcendental entire function @xmath0 is defined by @xmath522 and the _ lower order of growth _ is @xmath523 functions satisfying have order zero .
it follows from the results in ( * ? ? ?
* section 6 ) that there are many functions of order less than @xmath68 with regular growth for which @xmath53 is a spider s web .
( note that we do not use this terminology in that paper there we say that for these functions the set @xmath150 is connected .
this set has an analogous definition to the set @xmath53 see the remarks after corollary [ equal ] . )
let @xmath525 be as in part ( a ) .
then , by part ( b ) , there exists a sequence @xmath528 satisfying with @xmath530 , for @xmath34 .
so , by , for each @xmath34 , there exists @xmath531 with @xmath532 by , we have @xmath533 , for @xmath34 . thus the conditions of corollary [ m(r ) ] are satisfied and hence @xmath53 is a spider s web .
note that corollary [ regular ] generalises ( * ? ? ?
* theorem 5 ) which states that , if @xmath0 is a transcendental entire function of order less than 1/2 , then @xmath0 has no unbounded fatou components whenever , for all @xmath524 , there exists a real function @xmath534 defined on @xmath535 , where @xmath64 , such that , for @xmath536 , we have @xmath537 and @xmath538 indeed , if @xmath0 has order less than @xmath68 , then @xmath0 satisfies corollary [ regular ] part ( a ) for all sufficiently large values of @xmath67 this was proved by baker ( * ? ? ?
* satz 1 ) and also follows from the version of the @xmath539 theorem proved by barry @xcite .
if the function @xmath534 above exists for such a value of @xmath67 , then corollary [ regular ] part ( b ) holds by taking @xmath540
. then corollary [ regular ] implies that @xmath53 is a spider s web , so @xmath0 has no unbounded fatou components by theorem [ main6 ] part ( b ) . for a discussion of various functions satisfying the regularity condition , see ( * ? ? ?
* section 7 ) .
for example , we show there that holds for a suitable function @xmath534 if there exists @xmath541 such that the function @xmath542 satisfies @xmath543 the condition is used by anderson and hinkkanen in @xcite and it is satisfied if @xmath0 has finite order and positive lower order see @xcite .
so , if @xmath0 has finite order and positive lower order , then @xmath0 satisfies corollary [ regular ] part ( b ) .
we now discuss several classes of functions for which corollary [ regular ] part ( a ) is satisfied for all sufficiently large values of @xmath67 .
it follows from corollary [ regular ] that , for functions in these classes , @xmath53 is a spider s web whenever the growth is regular , as defined by for an appropriate value of @xmath67 . + * class 1 * as mentioned above , it follows from the @xmath539 theorem that any function of order less than @xmath68 satisfies corollary [ regular ] part ( a ) for all sufficiently large values of @xmath67 .
many examples of functions of order less than @xmath68 with regular growth are given in ( * ? ? ?
* section 6 ) .
these examples include the function @xmath544 which has order @xmath545 .
the spider s web @xmath53 for this function is illustrated in section 1 .
+ _ remark . _ rod halburd ( personal communication ) has pointed out that , more generally , functions of the form @xmath546 are entire with order @xmath547 .
these functions also have regular growth , so if @xmath548 , then @xmath53 is a spider s web .
further to this , sixsmith @xcite has observed that , for many entire functions @xmath549 , the function @xmath550 has an @xmath53 spider s web .
+ * class 2 * many entire functions for which the power series expansion has wide enough ` gaps ' also satisfy corollary [ regular ] part ( a ) for all sufficiently large values of @xmath67 .
a transcendental entire function @xmath0 has _ fabry gaps _ if @xmath551 and @xmath552 as @xmath553 .
it was shown by fuchs @xcite that , if @xmath0 has finite order and fabry gaps , then , for each @xmath554 , @xmath555 holds for all @xmath20 outside a set @xmath47 for which @xmath556 together with lemma [ convex ] , this result shows that functions of finite order with fabry gaps satisfy corollary [ regular ] part ( a ) for each @xmath524 .
we now give an example of such a function @xmath0 for which corollary [ regular ] part ( b ) is also satisfied for each @xmath524 , and hence @xmath53 is a spider s web .
let @xmath0 be the transcendental entire function given by @xmath557 ^ 2 } , \mbox { where } c>0.\ ] ] then @xmath0 is a function of order @xmath558 with fabry gaps and regular growth , and so @xmath53 is a spider s web .
note that , for @xmath559 , the power series defining @xmath0 is obtained from the taylor series for @xmath560 by discarding all terms except those for which the power of @xmath10 is of the form @xmath561 .
the function @xmath0 clearly has fabry gaps , and it is not difficult to check that it has order and lower order equal to @xmath558 .
these can be calculated either directly or by using the following formulae ( see @xcite ) for the order and lower order of an entire function of the form @xmath562 : since @xmath0 has finite order and fabry gaps , it satisfies corollary [ regular ] part ( a ) for each @xmath524 .
since @xmath0 has finite order and positive lower order , it follows from the discussion after corollary [ regular ] that the growth of @xmath0 is sufficiently regular to satisfy corollary [ regular ] part ( b ) for each @xmath524 .
thus , by corollary [ regular ] , @xmath53 is a spider s web .
there are also functions of _ infinite _ order with gap series for which corollary [ regular ] part ( a ) is satisfied for each @xmath524 .
hayman @xcite showed that fuch s result above , for functions of finite order with fabry gaps , holds for functions of any order provided that a stronger gap series condition is satisfied , which can be stated as follows : @xmath564 for some @xmath565 and sufficiently large values of @xmath348 .
thus , by the same argument as above , such functions also satisfy corollary [ regular ] part ( a ) for each @xmath524 .
an explicit example of a transcendental entire function of infinite order which satisfies and with sufficiently regular growth to ensure that @xmath53 is a spider s web is given by sixsmith @xcite .
_ it was proved by wang @xcite that functions with gap series of the type described here and with regular growth of a certain type have no unbounded fatou components .
this is also implied by our results since functions for which @xmath53 is a spider s web have no unbounded fatou components . +
* class 3 * many functions which show the pits effect also satisfy corollary [ regular ] part ( a ) for all sufficiently large values of @xmath67 . roughly speaking
, a function shows the pits effect if it is very large except in small regions , known as pits , around the zeros of @xmath0 . if the pits are sufficiently small and spaced out and the function is sufficiently large outside the pits , then corollary [ regular ] part ( a ) is satisfied .
the pits effect has been studied by many authors and , in most cases , it is sufficiently strong for corollary [ regular ] part ( a ) to be satisfied .
here we look in detail at the functions studied by littlewood and offord in @xcite since it follows from their results that , in a certain sense , for almost all functions of finite order and positive lower order , @xmath53 is a spider s web
. thus functions for which @xmath53 is a spider s web are not unusual , as was originally thought .
other papers giving examples of functions which show a suitably strong version of the pits effect include @xcite and @xcite .
suppose that @xmath566 is a transcendental entire function of order @xmath567 with positive lower order and that @xmath568 where the @xmath569 take the values @xmath570 with equal probability .
then @xmath53 is a spider s web for almost all functions in @xmath112 .
it was shown by littlewood and offord in @xcite that almost all functions in the set @xmath112 show the pits effect .
more precisely , they proved , among other things , that , for each @xmath554 , there exist @xmath571 and a set @xmath572 of measure less than @xmath573 such that if @xmath574 , @xmath575 , @xmath576 and @xmath577 is the largest @xmath6 satisfying this equality , then 1 . if @xmath10 is not in a pit and @xmath578 , then @xmath579 2 .
if @xmath10 is in a pit and @xmath580 , then the radius of the pit is at most @xmath581 ; 3 .
the number of pits in the annulus @xmath582 is at most @xmath583 .
it follows from ( * ? ? ?
* theorem 6 and the proof of ( 6.20 ) ) and also from ( * ? ? ?
* theorem 6.23 and lemma 6.15 ) that @xmath584 and @xmath585 for @xmath586 , where @xmath47 is a set of finite logarithmic measure ; that is , @xmath587 .
we now take @xmath573 with @xmath588 , where @xmath589 is the lower order of @xmath0 .
it follows from ( a ) above together with that , if @xmath10 is not in a pit and @xmath578 , then @xmath590 for all values of @xmath20 except those in a set of finite logarithmic measure .
further , it follows from ( b ) and ( c ) above together with that the set of values of @xmath20 for which @xmath591 meets a pit has finite logarithmic measure and so holds except in a set of finite logarithmic measure . together with lemma [ convex ] , this is sufficient to show that , for each function in @xmath592 , corollary [ regular ] part ( a ) is satisfied for all sufficiently large values of @xmath67 .
the result now follows since @xmath573 can be chosen to be arbitrarily small and functions in the class @xmath112 have finite order and positive lower order , and hence satisfy corollary [ regular ] part ( b ) for each @xmath524 .
_ it is of interest to know which of the above examples of functions with @xmath53 spiders webs have no multiply connected fatou components . in forthcoming work
@xcite we show that any transcendental entire function for which @xmath19 satisfies the condition , introduced by anderson and hinkkanen , has _ no _ multiply connected fatou components , and we give sufficient conditions on @xmath0 for to hold .
we end this section by proving the following result which allows us to construct many more examples of functions for which @xmath53 is a spider s web .
for example , we can deduce from theorem [ main99 ] and a theorem of plya ( * ? ? ?
* theorem 2.9 ) that , if @xmath0 has non - zero order and @xmath53 is a spider s web , then @xmath593 has infinite order and @xmath594 is a spider s web .
suppose that @xmath53 is a spider s web .
it follows from lemma [ sw1 ] part ( b ) that each component of @xmath54 is bounded .
we know from theorem [ eq ] that @xmath158 and so each component of @xmath595 is bounded .
thus , by lemma [ sw1 ] part ( b ) , @xmath69 is a spider s web . to prove the converse
, we use the result of theorem [ eq ] that if @xmath596 is sufficiently large , then @xmath597 .
it follows from lemma [ sw1 ] part ( d ) that , if @xmath69 is a spider s web then @xmath598 is also a spider s web and hence , by lemma [ sw1 ] part ( b ) , that each component of @xmath599 is bounded .
thus each component of @xmath600 is also bounded and so , by lemma [ sw1 ] part ( b ) , @xmath601 is a spider s web .
thus , by lemma [ sw1 ] part ( d ) , @xmath53 is a spider s web .
* acknowledgements . *
we thank walter bergweiler and lasse rempe for stimulating discussions , and dan nicks , john osborne and dave sixsmith for their helpful comments .
we also thank dominique fleischmann for producing the picture of the @xmath53 spider s web . | let @xmath0 be a transcendental entire function and let @xmath1 denote the set of points that escape to infinity ` as fast as possible ' under iteration . by writing @xmath1 as a countable union of closed sets , called ` levels ' of @xmath1
, we obtain a new understanding of the structure of this set .
for example , we show that if @xmath2 is a fatou component in @xmath1 , then @xmath3 and this leads to significant new results and considerable improvements to existing results about @xmath1 . in particular ,
we study functions for which @xmath1 , and each of its levels , has the structure of an ` infinite spider s web ' .
we show that there are many such functions and that they have a number of strong dynamical properties .
this new structure provides an unexpected connection between a conjecture of baker concerning the components of the fatou set and a conjecture of eremenko concerning the components of the escaping set . |
a particularly interesting aspect of the physics programme of the @xmath15 factories for the exploration of the kobayashi maskawa mechanism of cp violation @xcite is given by @xmath1 and @xmath4 modes ( see @xcite and references therein ) . in the summer of 2003 , the last missing element of the @xmath1 system , @xmath16 , was eventually observed by the babar @xcite and belle collaborations @xcite , with a surprisingly prominent rate , pointing to large corrections to the predictions of the qcd factorization approach @xcite . concerning the @xmath4 system , the last missing transition , @xmath17 , was already observed by the cleo collaboration in 2000 @xcite , and is now well established . immediately after that measurement , a puzzling pattern in the ratios @xmath2 and
@xmath3 of the charged and neutral @xmath4 rates has been pointed out @xcite .
this `` @xmath4 puzzle '' has survived over the years and was recently reconsidered by several authors ( see , for instance , @xcite ) .
it is an important feature of the observables @xmath2 and @xmath3 that they are significantly affected by ( colour - allowed ) electroweak ( ew ) penguins @xcite . on the other hand ,
the ratio @xmath8 of the cp - averaged @xmath5 and @xmath9 branching ratios , which is expected to be only marginally affected by ( colour - suppressed ) ew penguins , does not show an anomalous behaviour . since ew penguins
offer an attractive avenue for new physics ( np ) to manifest itself @xcite , the @xmath4 puzzle may indicate np in the ew penguin sector .
should this actually be the case , also several rare @xmath18 and @xmath15 decays may show np effects @xcite , thereby complementing the @xmath4 puzzle in a valuable manner . in @xcite , we developed a strategy to address these exciting issues systematically , with the following logical structure : * using the isospin flavour symmetry of strong interactions and assuming the range for the angle @xmath19 of the unitarity triangle ( ut ) that follows in the standard model ( sm ) from the ckm fits @xcite , we may extract a set of hadronic parameters characterizing the @xmath1 system from the experimental results for the corresponding cp - averaged branching ratios and the cp - violating observables of @xmath20 .
we found large deviations of the hadronic parameters from the predictions of qcd factorization .
moreover , we could predict the cp asymmetries of the @xmath7 decay in the sm . *
if we use the @xmath21 flavour symmetry and neglect penguin annihilation and exchange topologies , which can be probed through @xmath22 and @xmath23 modes , the hadronic @xmath1 parameters allow us to determine their @xmath4 counterparts .
assuming again the sm , as in the @xmath1 analysis , we may predict all observables offered by the @xmath4 system , including also cp asymmetries .
we found agreement with the experimental picture for @xmath8 , whereas the situation in the @xmath3@xmath2 plane was _ not _ in accordance with experiment .
this discrepancy could be resolved through np effects in the ew penguin sector , requiring a significant enhancement of the parameter @xmath24 measuring their strength relative to the tree contributions , and a np phase @xmath25 , which vanishes in the sm , around @xmath26 . * assuming a more specific ( but popular ) scenario @xcite@xcite , where np enters the ew penguin sector through @xmath27 penguins , system in a slightly different scenario involving an additional @xmath28 boson .
] we obtain an interesting interplay between the parameters @xmath24 and @xmath25 following from the resolution of the @xmath4 puzzle and several rare @xmath15 and @xmath18 decays .
this allowed us to explore the impact of the data for @xmath29 and @xmath30 processes , constraining the enhancement of @xmath24 to be smaller than that suggested by the @xmath4 data , thereby favouring a smaller value of @xmath2 and a larger value of @xmath3 .
in fact , this pattern has been confirmed to a large extent by the new data .
taking these constraints into account , there may still be prominent np effects in the rare - decay sector , the most spectacular ones in @xmath31 and @xmath32 , exhibiting branching ratios that could be enhanced with respect to the sm by factors of @xmath33 and @xmath34 , respectively .
in addition , we discussed the determination of @xmath19 ( and the other two ut angles @xmath35 and @xmath36 ) , where we obtained a result in agreement with the ckm fits @xcite , had a closer look at the @xmath37-meson decays @xmath38 and @xmath39 , and performed a couple of consistency checks of the @xmath21 flavour symmetry , which did not indicate large corrections .
as there were several exciting experimental developments thanks to the babar and belle collaborations since we wrote our original papers @xcite , it is interesting to confront our strategy with the most recent data , although the picture is still far from being settled .
the most important aspects are the following : * several new results for the @xmath1 and @xmath4 branching ratios @xcite@xcite . * first results for the direct cp asymmetry of @xmath7 @xcite .
* updates for the cp - violating observables of @xmath20 @xcite , as well as for the cp asymmetries of several @xmath4 modes @xcite .
* observation of direct cp violation in @xmath5 @xcite , representing a new milestone in the exploration of cp violation .
some implications of this measurement have been discussed in @xcite and @xcite , concerning the sm and supersymmetry , respectively .
* observation of @xmath40 , i.e. of the first @xmath41 penguin decay , and an emerging signal for its charged counterpart @xmath10 @xcite . in our analysis
, we will use the averages for these new results that were compiled by the `` heavy flavour averaging group '' ( hfag ) @xcite , but also make a number of refinements and generalizations , in particular : * we include the ew penguins of the @xmath42 system in our analysis . as anticipated in @xcite ,
this has a small impact on the numerics , but is a conceptual improvement .
* in view of the new @xmath43 data , we investigate the impact of certain hadronic effects , which can be constrained through the emerging experimental signal for @xmath10 decays .
the outline is as follows : in section [ sec : bpipi ] , we focus on the @xmath1 system and move on to the @xmath4 modes in section [ sec : bpik ] .
finally , we summarize our conclusions in section [ sec : concl ] .
the starting point of our analysis of the @xmath1 decays is given by the amplitudes @xmath44 = -[t+c]\label{b+pi+pi0}\\ a(b^0_d\to\pi^+\pi^-)&=&-[\tilde t + p]\label{bdpi+pi-}\\ \sqrt{2}a(b^0_d\to\pi^0\pi^0)&=&-[\tilde c - p],\label{bdpi0pi0}\end{aligned}\ ] ] which satisfy the following well - known isospin relation @xcite : @xmath45 the individual amplitudes of ( [ b+pi+pi0])([bdpi0pi0 ] ) can be expressed as @xmath46\label{t - tilde}\\ \tilde c & = & \lambda^3 a r_b e^{i\gamma}\left[{\cal c}+\left({\cal p}_{tu}- { \cal e}\right)\right],\label{c - tilde}\end{aligned}\ ] ] where @xmath47 are the usual parameters in the wolfenstein expansion of the cabibbo
maskawa ( ckm ) matrix @xcite , @xmath48 measures one side of the ut , the @xmath49 describe the strong amplitudes of qcd penguins with internal @xmath24-quark exchanges ( @xmath50 ) , including annihilation and exchange penguins , while @xmath51 and @xmath52 are the strong amplitudes of colour - allowed and colour - suppressed tree - diagram - like topologies , respectively , and @xmath53 denotes the strong amplitude of an exchange topology .
the amplitudes @xmath54 and @xmath55 differ from @xmath56 through the @xmath57 pieces , which may play an important rle @xcite .
note that these terms contain also the `` gim penguins '' with internal up - quark exchanges , whereas their `` charming penguin '' counterparts enter in @xmath58 through @xmath59 , as can be seen in ( [ p - def ] ) @xcite@xcite . in order to characterize the dynamics of the @xmath1 system , it is convenient to introduce the following hadronic parameters : @xmath60 @xmath61,\ ] ] where @xmath62 , @xmath63 and @xmath64 denote the cp - conserving strong phases of @xmath55 , @xmath54 and @xmath58 . in the @xmath1 system ,
the ew penguin contributions are expected to play a minor rle @xcite , and were therefore neglected in ( [ b+pi+pi0])([bdpi0pi0 ] ) . however , applying the isospin flavour symmetry of strong interactions , they can be included @xcite , yielding @xmath65\left[e^{i\gamma}+\tilde qe^{-i\beta}\right ]
\label{b+pi+pi0ewp}\\ a(b^0_d\to\pi^+\pi^-)&=&-|\tilde t|e^{i\delta_{\tilde t } } \left[e^{i\gamma}-d e^{i\theta}\right]\label{bdpi+pi - ewp}\\ \sqrt{2}a(b^0_d\to\pi^0\pi^0)&=&|p|e^{i\delta_p } \left[1+\frac{x}{d}e^{i\gamma}e^{i(\delta-\theta)}+\tilde q\left ( \frac{1+x e^{i\delta}}{d}\right)e^{-i\theta}e^{-i\beta } \right],\label{bdpi0pi0ewp}\end{aligned}\ ] ] where @xmath66 measures the strength of the sum of the colour - allowed and colour - suppressed ew penguin contributions with respect to the sum of the colour - allowed and colour - suppressed tree - diagram - like contributions
. it should be emphasized that ( [ qtilde ] ) was derived for the sm .
in contrast to our previous analysis @xcite , we shall also include the ew penguin contributions in the numerical analysis performed below .
although their impact is actually small , this is a conceptual improvement .
however , as soon as we consider np in the ew penguin sector , the @xmath1 analysis does no longer fully separate from that of the @xmath4 system , i.e. items i ) and ii ) of section [ sec : intro ] are no longer completely independent . however , their cross talk is actually very small .
following @xcite , we use the ratios @xmath67 \frac{\tau_{b^0_d}}{\tau_{b^+}}\label{rpm - def}\\ r_{00}^{\pi\pi}&\equiv&2\left[\frac{\mbox{br}(b_d^0\to\pi^0\pi^0)+ \mbox{br}(\bar b_d^0\to\pi^0\pi^0)}{\mbox{br}(b_d^0\to\pi^+\pi^-)+ \mbox{br}(\bar b_d^0\to\pi^+\pi^-)}\right]\label{r00-def}\end{aligned}\ ] ] of the cp - averaged @xmath1 branching ratios , and the cp - violating observables provided by the time - dependent rate asymmetry @xmath68 as the input for our @xmath1 analysis . concerning the former quantities , they can be written in the following generic form : @xmath69 on the other hand , the cp - violating @xmath20 observables involve , in addition to the angle @xmath19 of the ut , only the hadronic parameters @xmath70 ; the mixing - induced cp asymmetry depends , furthermore , on the @xmath71@xmath72 mixing phase @xmath73 , which equals @xmath74 in the sm .
consequently , we may write @xmath75 explicit expressions for the functions @xmath76 and @xmath77 can be found in @xcite .
.the current status of the @xmath1 input data for our strategy , with averages taken from @xcite , and comparison with the picture of our previous analysis @xcite . for the evaluation of @xmath78
, we have used the life - time ratio @xmath79 @xcite . [ cols="^,^,^",options="header " , ] the most important new experimental development in the @xmath4 sector is the observation of direct cp violation in @xmath5 decays , which could eventually be established this summer by the babar @xcite and belle @xcite collaborations .
these measurements complement the observation of direct cp violation in the neutral kaon system by the na48 ( cern ) @xcite and ktev ( fnal ) @xcite collaborations , where this phenomenon is described by the famous observable @xmath80 ; the world average taking the final results of these experiments @xcite into account is given by @xmath81 . for recent theoretical overviews of @xmath80 ,
see @xcite . in the case of @xmath5 , direct cp violation
is characterized by the asymmetry @xmath82 which is now measured by the babar and belle collaborations with the following results : @xmath83 we observe that these numbers are nicely consistent with each other .
they correspond to the following average @xcite : @xmath84 establishing the direct cp violation in @xmath5 decays at the @xmath85 level .
let us now follow the strategy developed in @xcite to confront the direct cp asymmetry of the @xmath86 channel with theoretical considerations .
the corresponding decay amplitude can be written as @xmath87,\ ] ] with @xmath88 and @xmath89,\ ] ] yielding @xmath90 the notation in ( [ p - prime ] ) and ( [ r - def ] ) is analogous to that used for the discussion of the @xmath1 modes in section [ sec : bpipi ] ; the primes remind us that we are dealing with @xmath91 transitions . in ( [ acpdir - r ] )
, we can see nicely that @xmath92 is induced through the interference between tree and qcd penguin topologies , with a cp - conserving strong phase difference @xmath93 and a cp - violating weak phase difference @xmath19 .
if we use now the working hypothesis given in subsection [ ssec : bpik - pre ] , we obtain @xcite @xmath94 this relation allows us to determine @xmath95 from the values of the hadronic @xmath1 parameters @xmath70 given in table [ tab : bpipi - hadr ] , with the following result : @xmath96 having these parameters at hand , which refer to the range for @xmath19 in ( [ gamma - range ] ) , we are in a position to calculate the direct cp asymmetry of @xmath5 : @xmath97 which should be compared with our previous prediction of @xmath98 @xcite . looking at ( [ acpdir - av ] ) , we observe that ( [ acp - pred ] ) is in nice agreement with the experimental result .
in fact , in our previous analysis , which was confronted with the experimental average of @xmath99 , we advocated that this cp asymmetry should go up , in full accordance with the babar result in ( [ acpdir - exps ] ) . despite the large value of @xmath93
, the rather small value of @xmath100 ensures that this cp asymmetry does not take a value that is much larger than the experimental ones . in the case of qcdf @xcite and pqcd @xcite ,
the following patterns arise : @xmath101 @xmath102 consequently , the qcdf picture is not in agreement with the experimental result ( [ acpdir - av ] ) , pointing in particular towards the _ opposite _ sign of the direct cp asymmetry . on the other hand ,
pqcd reproduces the sign correctly , but favours an asymmetry on the larger side .
these features can also be seen with the help of ( [ acpdir - r ] ) and ( [ r - det ] ) from the qcdf and pqcd predictions given in ( [ qcdf - pred ] ) and ( [ pqcd - pred ] ) , respectively . another direct confrontation of @xmath92 with theory is provided by the following sm relation , which can be derived with the help of assumptions i ) and ii ) specified in subsection [ ssec : bpik - pre ] @xcite : @xmath103}_{\mbox{0.38$\pm$0.04}}= \underbrace{-\left[\frac{{\cal a}_{\rm cp}^{\rm dir}(b_d\to\pi^\mp k^\pm)}{{\cal a}_{\rm cp}^{\rm dir}(b_d\to\pi^+\pi^- ) } \right]}_{\mbox{0.31$\pm$0.11}}.\ ] ] here , we have introduced the parameter @xmath104 and the ratio @xmath105 of the kaon and pion decay constants takes the factorizable @xmath21-breaking corrections into account . in ( [ cp - rel ] ) , we have also indicated the current experimental results , and observe that this relation is nicely satisfied within the current experimental uncertainties .
this feature give us further confidence in our working assumptions , in addition to the agreement between ( [ acpdir - av ] ) and ( [ acp - pred ] ) . following from the cp - violating @xmath20 observables and the data for
the @xmath5 decays in the @xmath106@xmath107 plane and comparison with the other constraints for the ut , as discussed in @xcite .
the shaded dark ellipse is the result of the ut fit , while the quadrangle corresponds to the second value of @xmath108 in ( [ gamma ] ) and the @xmath109 constraint.,width=491 ] the quantity @xmath110 introduced in ( [ cp - rel ] ) can be written as follows : @xmath111 if we now complement @xmath110 with the cp - violating observables @xmath112 and @xmath113 , which take the general form in ( [ cp - bpipi - gen ] ) , and use the experimental result for @xmath73 in ( [ phi - d - range ] ) , we are in a position to determine @xmath19 and @xmath70 @xcite .
in addition to the expression involving the cp - averaged @xmath20 and @xmath5 branching ratios in ( [ cp - rel ] ) , the corresponding direct cp asymmetries provide an alternative avenue for the determination of @xmath110 , which is theoretically more favourable as far as the @xmath21-breaking corrections are concerned , but is currently affected by larger experimental uncertainties .
the corresponding values of @xmath110 are given as follows : @xmath114 complementing them with the cp - violating @xmath20 asymmetries in table [ tab : bpipi - input ] , we obtain the following solutions for @xmath19 : @xmath115 as we discussed in @xcite , the twofold ambiguities arising in these determinations can be lifted by using additional experimental information for @xmath116 and the other @xmath1 decays , thereby leaving us with the values around @xmath117 . in fig .
[ fig : ut ] , we show the corresponding situation for the ut in the @xmath106@xmath107 plane of the generalized wolfenstein parameters @xcite , where we compare the values of @xmath19 obtained above with the ut fit performed in @xcite . using , in addition , the range for the ut side @xmath109 in ( [ rb - def ] ) , we may also determine @xmath35 and @xmath36 , with the following results : @xmath118 @xmath119 the results for @xmath35 are nicely consistent with those obtained from the most recent data for @xmath120 processes , as reviewed in @xcite ; a similar comment applies to the ranges for @xmath19 following from decays of the kind @xmath121 .
let us also emphasize that our results for @xmath36 are in excellent agreement with the sm relation @xmath122 . in this context , it is important to stress that actually @xmath73 and not @xmath36 itself enters our analysis as an input parameter
. enters is in ( [ b+pi+pi0ewp ] ) and ( [ bdpi0pi0ewp ] ) to describe the tiny ew penguin effects in @xmath123 and @xmath124 , respectively .
however , @xmath36 does _ not _ enter the @xmath125 amplitude ( [ bdpi+pi - ewp ] ) , which is the only relevant @xmath1 ingredient for our ut analysis . ] the determination of @xmath35 and @xmath36 in ( [ alpha - determ ] ) and ( [ beta - determ ] ) is therefore an important test of the consistency of our approach ( or of the sm relation @xmath122 ) .
more refined determinations of @xmath19 from the cp - violating @xmath20 observables are provided by the decay @xmath38 @xcite .
this channel is already accessible at run ii of the tevatron @xcite , and can be fully exploited at lhcb @xcite . in appendix
[ app : bskk ] , we collect the updated values of the sm predictions for the @xmath38 observables presented in @xcite .
the direct cp violation in @xmath5 decays provides valuable information and is perfectly consistent with the sm picture emerging from our strategy @xcite .
let us now also consider the cp - averaged @xmath5 rate . in order to analyse this quantity ,
it is useful to consider simultaneously @xmath9 decays @xcite , and to introduce @xmath126\frac{\tau_{b^+}}{\tau_{b^0_d}}.\ ] ] the common feature of the @xmath127 and @xmath128 decays is that ew penguins may only contribute to them in colour - suppressed form , and are hence expected to play a minor rle . as can be seen in table [ tab : bpik - input1 ] , the experimental average for @xmath8 went down sizeably with respect to the situation of our previous analysis .
this feature is essentially due to the most recent update of the cp - averaged @xmath9 branching ratio by the babar collaboration @xcite , taking certain radiative corrections into account ; similar effects are currently investigated by the belle collaboration @xcite .
consequently , the experimental picture is not yet settled ( see also @xcite ) .
the last observable provided by the @xmath5 , @xmath9 system is in addition to @xmath92 and @xmath8 the direct cp asymmetry of the @xmath9 modes : @xmath129 the current experimental average is given as follows @xcite : @xmath130 and does not indicate any cp - violating effects in this channel .
in order to complement the @xmath127 amplitude in ( [ bdpi - k+ ] ) , we write @xmath131,\ ] ] where @xmath132 was defined in ( [ p - prime ] ) , and @xmath133.\ ] ] here @xmath134 describes the penguins with internal up - quark exchanges contributing to the _ charged _ @xmath4 modes , and @xmath135 is an annihilation topology .
we arrive then straightforwardly at the following expression for @xmath8 : @xmath136 while the direct cp asymmetry of the @xmath9 modes is given by @xmath137.\ ] ] let us first assume that @xmath138 can be neglected , as is usually done for the analysis of the @xmath5 , @xmath9 system .
this approximation corresponds to a vanishing value of ( [ acp - b+pi+k - expr ] ) , which is in accordance with ( [ acp - b+pi+k - exp ] ) . in view of the rather small experimental value of @xmath8 , it is interesting ( see also @xcite ) to return to the bounds on @xmath19 that can be obtained with the help of @xmath139 provided @xmath8 is measured to be smaller than 1 @xcite . using the value of @xmath140 in table [ tab : bpik - input1 ]
, we obtain the upper bound @xmath141 which is basically identical with the range for @xmath19 in ( [ gamma - range ] ) . in analogy to the prediction in ( [ acp - pred ] ) ,
the hadronic parameters in ( [ r - del - det ] ) allow us also to calculate @xmath8 , with the following result : @xmath142 which is the update of @xmath143 given in @xcite . comparing with the new experimental result in table [ tab : bpik - input1 ] , we observe that it favours a smaller value . consequently , in the case of @xmath8 , we encounter now a sizeable deviation of our prediction from the experimental average , whereas we obtain excellent agreement with the @xmath15-factory data for the cp - violating @xmath5 asymmetry . moreover ,
also the bound on @xmath19 in ( [ fm - bound1 ] ) appears to be surprisingly close to the sm range .
since @xmath8 may be affected by @xmath138 , whereas @xmath92 does _ not _ involve this hadronic parameter , it is therefore suggested that @xmath138 has actually a non - negligible impact on the numerical analysis .
in addition to @xmath8 , the parameter @xmath138 enters also the direct cp asymmetry of the @xmath116 decays .
it is interesting to illustrate these effects in the @xmath8@xmath144 plane . to this end , we use ( [ r - expr ] ) with the central values of the hadronic parameters in ( [ r - del - det ] ) , and ( [ acp - b+pi+k - expr ] ) to calculate the contours for @xmath145 and @xmath146 shown in fig .
[ fig : r - adir ] .
we observe that for @xmath145 and @xmath147 the @xmath148 ranges of experiment and theory ( the contours only show the central value ) practically overlap , thereby resolving essentially the discrepancy between our theoretical prediction for @xmath8 and its most recent experimental value .
following appendix d.3 of @xcite , we have also included a second error bar for our theoretical prediction that indicates the variation of @xmath8 if colour - suppressed ew penguins are taken into account at a rather prominent level of @xmath149 , with @xmath150 $ ] .
we observe that , while the inclusion of these effects could also help to resolve the discrepancy , the impact of @xmath138 is significantly more important .
after this encouraging observation , let us have a closer look at the status of @xmath138 .
access to this parameter is provided by the decay @xmath151 , which is related to @xmath152 through the interchange of all down and strange quarks , i.e. through the @xmath153-spin flavour symmetry of strong interactions @xcite , which is a subgroup of @xmath154 . applying this symmetry ,
we may write ( for a detailed discussion , see @xcite ) @xmath155 \left[\frac{\mbox{br}(b^\pm\to \pi^\pm k)}{\mbox{br}(b^\pm\to k^\pm k ) } \right]=\frac{1 + 2\rho_{\rm c}\cos\theta_{\rm c}\cos\gamma+ \rho_{\rm c}^2}{\epsilon^2 - 2\epsilon\rho_{\rm c}\cos\theta_{\rm c}\cos\gamma+ \rho_{\rm c}^2},\ ] ] where @xmath156 describes @xmath21-breaking corrections . in factorization
, we obtain @xmath157 \left[\frac{f_{b\pi}(m_k^2;0^+)}{f_{bk}(m_k^2;0^+)}\right]=0.79,\ ] ] where the numerical value refers to the recent light - cone sum - rule analysis performed in @xcite .
the measurement of @xmath18 allows us to obtain the following allowed range for @xmath138 : @xmath158 using the most recent upper bound of @xmath159 ( 90% c.l . )
reported by the babar collaboration @xcite , and the measured value of @xmath160 in table [ tab : bpik - input1 ] , this relation implies @xmath161 @xmath144 plane .
we show contours for @xmath145 and @xmath162 , with @xmath163 $ ] .
the experimental ranges for @xmath8 and @xmath144 and our theoretical prediction are indicated in grey .
the second error bar beneath the theoretical prediction ( almost identical in size to the first one ) indicates the variation of @xmath8 if colour - suppressed ew penguins are taken into account.[fig : r - adir],width=434 ] @xmath138 plane following from the current data for the cp - averaged @xmath10 branching ratio ( parametrized by @xmath18 ) and the direct cp violation in @xmath9 decays as discussed in the text.,width=434 ] the neutral counterpart of @xmath164 , the @xmath165 channel , was observed this summer by the babar collaboration @xcite , with the cp - averaged branching ratio @xmath166 corresponding to a significance of @xmath167 .
this exciting measurement is the first _
direct _ experimental evidence for a @xmath41 penguin process .
interestingly , it is in accordance with the lower sm bounds derived in @xcite , which suggested that the discovery of this transition should actually be just ahead of us . concerning @xmath164
, there is an emerging signal at the @xmath168 level , which would correspond to @xmath169 in the sm , we expect a lower bound of @xmath170 @xcite .
inserting the range in ( [ br - b+k+k ] ) into ( [ rhoc - bounds ] ) yields @xmath171 if we use the sm value of @xmath19 in ( [ gamma - range ] ) , the measurement of @xmath172 provides even more information .
in fact , ( [ k - def ] ) allows us then to determine @xmath138 as a function of @xmath173 with the help of @xmath174 where @xmath175 \cos\gamma\cos\theta_{\rm c } , \quad \tilde b\equiv\frac{\epsilon^2 k-1}{k-1}.\ ] ] in analogy , the experimental result for @xmath176 allows us to fix another contour in the @xmath173@xmath138 plane . using ( [ acp - b+pi+k - expr ] ) , we obtain @xmath177 with @xmath178\cos\gamma\cos\theta_{\rm c}.\ ] ] in fig .
[ fig : rhoc - thetac - dat ] , we assume @xmath179 , and confront these considerations with the experimental results in ( [ acp - b+pi+k - exp ] ) and ( [ br - b+k+k ] ) , despite the fact that the latter branching ratio corresponds only to an emerging signal for the @xmath10 channel . since the corresponding lower @xmath180 and central values of @xmath172 would be smaller than the lower bound derived in @xcite , ( [ rhoc - cont1 ] ) would not have a physical solution for these results . however , for values of @xmath172 larger than this bound , we obtain an expanding allowed region in the @xmath173@xmath138 plane , as shown in fig . [
fig : rhoc - thetac - dat ] .
we also observe that @xmath181 has a rather small impact on the overall allowed parameter space .
it is interesting to note that the data favour strong phases @xmath173 around @xmath182 ( and not around @xmath183 ) , as is suggested by the general expression in ( [ rho - c - def ] ) . in the following
, we will use @xmath145 and @xmath184 , in agreement with ( [ rhoc - bounds - num ] ) and the allowed region in fig . [ fig : rhoc - thetac - dat ] ; a more rigorous analysis will have to wait until the data for the @xmath10 decays will have improved . with these values and @xmath179 , we obtain @xmath185 as can be seen in ( [ r - expr ] ) , this quantity describes the impact of @xmath186 on @xmath8 . in particular , the numerical value in ( [ wc - range ] ) shifts the central value @xmath187 in ( [ r - pred1 ] ) accordingly to 0.903 .
we observe that @xmath8 moves actually towards the experimental value through the impact of @xmath138 , thereby essentially resolving the discrepancy arising in subsection [ ssec : r - puzzle ] .
to conclude the discussion of the @xmath5 , @xmath9 system , let us emphasize that we can accommodate the corresponding data in the sm by using additional experimental information on @xmath10 decays , allowing us to take the hadronic parameter @xmath138 into account .
the remaining small numerical difference in the analysis of @xmath8 , if confirmed by future data , could be due to ( small ) effects of colour - suppressed ew penguins , which enter @xmath8 as well @xcite , and/or the limitations of our working hypothesis specified in subsection [ ssec : bpik - pre ] .
moreover , we would also not be surprised to see the experimental value of @xmath8 moving up in the future .
let us now turn to the decays @xmath188 and @xmath189 , where ew penguins enter in colour - allowed form . in order to analyse these transitions ,
it is particularly useful to introduce the following ratios : @xmath190\ ] ] @xmath191,\ ] ] i.e. to consider separately the charged and neutral @xmath4 modes @xcite .
the experimental situation of these quantities is summarized in table [ tab : bpik - input1 ] .
we observe that @xmath2 went down , thanks to the larger value of @xmath192 , and that @xmath3 moved marginally up .
furthermore , the decay @xmath188 offers a direct cp asymmetry , @xmath193 where we have also given the experimental average @xcite . in the case of @xmath194
decays , we have a final state with cp eigenvalue @xmath195 .
consequently , we may introduce a time - dependent rate asymmetry with the same structure as ( [ rate - asym ] ) , exhibiting direct _ and _ mixing - induced cp asymmetries .
the most recent values for these observables obtained by the babar @xcite and belle @xcite collaborations are consistent with each other , and correspond to the following new averages @xcite : @xmath196 the sm amplitudes for the decays @xmath127 and @xmath128 were already given in ( [ bdpi - k+ ] ) and ( [ b+pi+k0 ] ) , respectively . in the case of the @xmath188 and @xmath189 modes , the decay amplitudes can be written in the following form within the sm : @xmath197\label{b+pi0k+-sm}\\ \sqrt{2}a(b^0_d\to\pi^0k^0)&=&-p'\left[1+\rho_{\rm n}e^{i\theta_{\rm n } } e^{i\gamma}-qe^{i\omega}r_{\rm c}e^{i\delta_{\rm c}}\right ] .
\label{b0pi0k0-sm}\end{aligned}\ ] ] here the parameter @xmath24 , with the cp - conserving strong phase @xmath198 , measures the importance of the ew penguins with respect to the tree - diagram - like topologies . in the sm
, it can be determined with the help of the @xmath21 flavour symmetry of strong interactions @xcite , yielding @xmath199;\ ] ] for a detailed discussion of the colour - suppressed ew penguin contributions , which are neglected in ( [ b+pi0k+-sm ] ) and ( [ b0pi0k0-sm ] ) , see appendix d of @xcite .
moreover , we have @xmath200,\ ] ] as well as @xmath201,\ ] ] where the notation is analogous to the one introduced in subsection [ ssec : bpipi - ampls ] ; the primes remind us again that we have now turned to @xmath202 modes .
we observe that the hadronic parameters in ( [ r - def ] ) , ( [ rc - def ] ) and ( [ rho - n - def ] ) satisfy the following relations : @xmath203 @xmath204 with @xmath205 the values of @xmath95 following from the @xmath1 data can be found in ( [ r - del - det ] ) . in analogy , we may use @xmath206xe^{i\delta}\ ] ] to determine @xmath207 from their @xmath1 counterparts given in table [ tab : bpipi - hadr ] .
the factor @xmath208=1.05 \pm 0.18,\ ] ] where the numerical value refers to the light - cone sum - rule analysis of @xcite , describes factorizable @xmath21-breaking corrections .
since ( [ r - det ] ) is not affected by @xmath21-breaking effects within factorization , such a factor is not present in the case of this relation @xcite .
finally , we obtain then , with the help of ( [ rho - n - det ] ) , the numerical values @xmath209 and ( [ r - rho - rel ] ) yields @xmath210 alternatively , @xmath211 can be determined through the following well - known relation @xcite : @xmath212 which relies on the @xmath21 flavour symmetry and the neglect of the @xmath138 term in ( [ b+pi+k0 ] ) .
having the parameters in ( [ rhon - thetan - det ] ) and ( [ rc - deltac - det ] ) at hand , we may predict the values of @xmath2 and @xmath3 and of the cp - violating observables of @xmath9 , @xmath213 in the sm with the help of the formulae given in @xcite . in the case of the charged modes , we obtain @xmath214 @xmath215 where here ( and in the following ) the numbers with errors refer to @xmath216 , and the central value for the case @xmath145 , @xmath184 is given in brackets .
we observe that the impact of @xmath138 on @xmath2 is significantly weaker than in the case of @xmath8 discussed in subsection [ ssec : rhoc - analysis ] .
this is due to the feature that @xmath138 enters @xmath8 already at @xmath217 , whereas it affects @xmath2 through second order terms of @xmath218 and @xmath219 @xcite .
on the other hand , @xmath3 and the cp - violating observables of @xmath220 are _ not _ affected by @xmath138 , so that we obtain the following sm predictions : @xmath221 @xmath222 so far , we could accommodate all features of the @xmath15-factory data for the @xmath1 and @xmath4 modes in a satisfactory manner in the sm .
now we observe that this is _ not _ the case for @xmath3 and to a smaller extend for @xmath2 . as we have emphasized above ,
@xmath3 does _ not _ depend on @xmath138 , so this parameter can not be at the origin of this puzzle , in contrast to the case of @xmath8 , and has , moreover , a minor impact on @xmath2 .
moreover , as we discussed in @xcite , the colour - suppressed ew penguin topologies have _ no _ impact on @xmath3 in our sm analysis , as they can be absorbed in a certain manner , but could affect @xmath2 .
however , as we have seen in subsection [ ssec : rhoc - analysis ] , the analysis of @xmath8 disfavours anomalously large contributions of this kind , in contrast to the claims made in @xcite . concerning @xmath21-breaking corrections , the agreement between ( [ acpdir - av ] ) and our sm prediction ( [ acp - pred ] ) , the successful confrontation of ( [ cp - rel ] ) with the data , and the emerging picture of the ut in perfect accordance with the sm discussed in subsection [ ssec : ut - det ] do not indicate large corrections to ( [ r - det ] ) .
moreover , the agreement between ( [ rc - deltac - det ] ) and ( [ rc - alt - det ] ) indicates that the leading @xmath21-breaking effects are indeed described by the corresponding factors in ( [ x - rel - su3 ] ) and ( [ rc - alt - det ] ) .
so what could then be the origin of the puzzling pattern of the measured values of @xmath3 and @xmath2 ?
@xmath2 plane .
we show contours for values of @xmath223 , @xmath224 and @xmath225 , with @xmath226 $ ] .
the experimental ranges for @xmath2 and @xmath3 and those predicted in the sm are indicated in grey , the dashed lines serve as a reminder of the corresponding ranges in @xcite.[fig : rn - rc],width=453 ] since @xmath3 and @xmath2 are significantly affected by ew penguins , it is an attractive possibility to assume that np enters through these topologies @xcite . in this case
, the successful picture described above would not be disturbed . on the other hand ,
we may obtain full agreement between the theoretical values of @xmath3 and @xmath2 and the data .
following @xcite , we generalize the ew penguin parameter as @xmath227 where @xmath25 is a cp - violating weak phase that vanishes in the sm , i.e. arises from np .
we may then use the measured values of @xmath2 and @xmath3 to determine @xmath24 and @xmath25 , with the following results : @xmath228 @xmath229 where the numbers in brackets illustrate again the impact of @xmath145 , @xmath184 on the central values .
although these hadronic parameters are not at the origin of the @xmath230 puzzle , as we have seen above , they have of course some impact on the extracted values of @xmath24 and @xmath25 , though the are not changing the overall picture .
it is useful to consider the @xmath3@xmath2 plane , as we have done in fig .
[ fig : rn - rc ] .
there we show contours corresponding to different values of @xmath24 , and indicate the experimental and sm ranges .
following @xcite , we choose the values of @xmath223 , @xmath231 and @xmath232 , where the latter reproduced the central values of @xmath2 and @xmath3 in our previous analysis @xcite .
the central values for the sm prediction have hardly moved , while their uncertainties have been reduced a bit . on the other hand , the central experimental values of @xmath2 and
@xmath3 have moved in such a way that @xmath24 decreased , while the weak phase @xmath25 remains around @xmath233 .
moreover , we obtain the following cp asymmetries in our np scenario : @xmath234 @xmath235 although the central value of our prediction for @xmath236 has a sign different from the corresponding experimental number , this can not be considered as a problem because of the large current uncertainties . concerning the observables of the @xmath237 channel ,
our prediction for the direct cp asymmetry is rather close to the experimental number , while the current experimental result for mixing - induced asymmetry is somewhat on the lower side .
however , the uncertainties of these very challenging measurements are still too large to draw conclusions . instead of using the value of @xmath238 in our analysis , which follows from the @xmath21 flavour symmetry
, we could alternatively determine this strong phase , together with @xmath24 and @xmath25 , from a combined analysis of @xmath3 , @xmath2 and the direct cp asymmetry of the @xmath239 modes . in our previous analysis @xmath240 ,
this led to small values of @xmath198 in perfect agreement with the picture following from the @xmath21 flavour symmetry . using the most recent data , we obtain @xmath241 where the values of @xmath24 and @xmath25 are practically unchanged from the numbers given in ( [ q - det ] ) and ( [ phi - det ] ) , respectively .
the updated value in ( [ omega - det ] ) does still not favour dramatic @xmath21-breaking effects .
finally , we would like to comment briefly on the direct cp asymmetry of the @xmath6 decays .
it was argued in the recent literature ( see , for instance , @xcite ) that the discrepancy between the experimental values in ( [ acpdir - av ] ) and ( [ acp - bpi0k - def ] ) was very puzzling .
however , our analysis shows nicely that this is actually not the case .
in particular , we have the following expression @xcite : @xmath242,\ ] ] where the @xmath138 terms are neglected for simplicity .
consequently , the small value in ( [ acp - bpi0k - def ] ) follows simply from the small strong phase @xmath243 in ( [ rc - deltac - det ] ) . on the other hand , the hadronic parameters @xmath100 and @xmath93 governing @xmath92 take very different values , as we have seen in subsection [ acp - dir - theo1 ] . the difference between the cp - violating @xmath239 and @xmath5 asymmetries can therefore be straightforwardly explained through hadronic effects within the sm , i.e. does not require np .
in this paper we have confronted our strategy for describing and correlating the @xmath1 , @xmath4 decays and rare @xmath18 and @xmath15 decays with the new data on @xmath1 , @xmath4 from babar and belle . within a simple np scenario of enhanced cp - violating ew penguins considered by us , the np contributions enter significantly only @xmath4 decays and rare @xmath18 and @xmath15 decays , while the @xmath1 system is practically unaffected by these contributions and can be described within the sm .
consequently the pattern of relations between various observables is in our strategy very transparent in that * the relations between @xmath1 and @xmath4 decays are strictly connected with the _ long - distance _ physics allowing us to calculate the hadronic parameters of the @xmath4 from the @xmath244 ones without the intervention of np contributions that enter at much shorter scales . * the relations between @xmath4 decays and rare
@xmath18 and @xmath15 decays are strictly connected with the _ short - distance _ physics allowing us to predict several spectacular departures from the sm expectations for rare @xmath18 and @xmath15 decays from the corresponding significant departures from the sm observed in the @xmath4 data .
the main messages from this new analysis are as follows : * the present data for those observables in the @xmath1 and @xmath4 systems that are essentially unaffected by np in the ew penguins are not only in accordance with our approach , but a number of predictions made by us in @xcite have been confirmed by the new data within theoretical and experimental uncertainties .
this is in particular the case of the direct cp asymmetry in @xmath245 but also in @xmath7 . for convenience of the reader
, we collect all cp - violating quantities involved in our analysis in table [ asymtab ] and show the comparison of the experimental values with our predictions . * the observed decrease of the ratio @xmath8 below our expectations in @xcite can be partially attributed to certain hadronic effects , represented by the non - vanishing value of @xmath138 , that could be tested in @xmath10 decays once these are experimentally better known .
in particular the sign of @xmath138 , which is more solid than its magnitude , points towards the decrease of @xmath8 relative to our previous estimate .
however , our present understanding of these effects allows us to expect that future more accurate measurements will find @xmath8 higher than its present central value . *
the decrease in the difference @xmath246 observed in the recent data of babar has been predicted by us on the basis of branching ratios for rare decays @xcite .
this is explicitly seen in table 2 of @xcite . in this manner
the overall description of @xmath247 , @xmath230 and rare decays within our approach has improved with respect to our previous analysis . *
the picture of rare decays presented by us in @xcite remains unchanged , since the values of @xmath24 and @xmath25 obtained are still slightly above the bound from @xmath248 used in @xcite .
in particular , the spectacular enhancement of @xmath249 as well as the enhancement of several other rare decays remain .
further implications for rare decays in this scenario can be found in @xcite * last but certainly not least the obtained value of @xmath19 and the ut are in full agreement with the usual ckm fits .
finally , we would like to comment on analyses using only the @xmath43 data .
it has been claimed in @xcite that the puzzle concerning the @xmath250 system is significantly reduced or not even present .
we would like to emphasize that a study of the @xmath4 decays alone is not very much constrained and consequently has a rather low resolution in search for np effects .
such an analysis is moreover not satisfactory as it ignores the information on long distance dynamics that we have already from other non - leptonic decays , in particular from @xmath1 decays that are connected with the @xmath4 system through @xmath21 flavour symmetry .
one should also not forget that , for a confrontation with the sm , the use of the @xmath21 flavour symmetry , which allows us to determine the ew penguin parameters @xmath24 and @xmath198 through ( [ q - sm ] ) , can not be avoided .
consequently one may ask why the @xmath21 flavour symmetry should be used to find @xmath24 and not for getting the full input from @xmath1 and in the future from @xmath251 decays . as demonstrated in @xcite and here , the use of the full information from the @xmath42 and @xmath43 systems allows us to uncover possible signals of np effects in @xmath43 decays that in turn change significantly the sm pattern of rare decay branching ratios . in this respect , we disagree with a statement made in @xcite that `` the data seem to disfavour np explanations , according to which np primarily modifies electroweak penguin contributions '' .
@xmath252 & @xmath253 ' '' '' & @xmath254 + @xmath255 & @xmath256 ' '' '' & @xmath257 + @xmath258 & @xmath259 ' '' '' & @xmath260 + @xmath261 & @xmath262 ' '' '' & @xmath263 + @xmath264 & @xmath265 ' '' '' & @xmath266 + @xmath267 & @xmath268 ' '' '' & @xmath269 + it will be exciting to follow the experimental progress on @xmath1 and @xmath4 decays and the corresponding efforts in rare decays . in particular new messages from babar and belle that the present central values of @xmath2 and @xmath3 have been confirmed at a high confidence level , a slight increase of @xmath8 and a message from kek @xcite in the next two years that the decay @xmath31 has been observed would give a strong support to the np scenario considered here .
* acknowledgments * + the work presented here was supported in part by the german bundesministerium fr bildung und forschung under the contract 05ht4woa/3 and the dfg project bu .
706/1 - 2 .
the decay @xmath38 is related to @xmath20 through the interchange of all down and strange quarks , i.e. through the @xmath153-spin flavour symmetry of strong interactions .
consequently , this symmetry allows us to determine the hadronic @xmath38 parameters @xmath270 through the values of their @xmath20 counterparts @xmath70 given in table [ tab : bpipi - hadr ] . using then the range of @xmath19 in ( [ gamma - range ] ) , and the sm value @xmath271 for the @xmath272@xmath273 mixing phase
, we arrive at the following sm predictions , updating those given in @xcite : @xmath274 concerning the cp - averaged @xmath38 branching ratio , which is of more immediate experimental interest , we have to take a certain @xmath21-breaking factor into account that has recently been calculated through qcd sum rules @xcite .
following @xcite , we obtain the updated value @xmath275 from the @xmath1 data .
alternatively , we may calculate @xmath276 with the help of the cp - averaged @xmath245 branching ratio , which requires , however , the additional assumption that penguin annihilation and exchange topologies play a minor rle ( see item ii in subsection [ ssec : bpik - pre ] ) . following this avenue yields @xmath277 in nice agreement with ( [ br - bskk ] ) .
let us note that the difference between ( [ acp - dir - bskk])([br - bskk ] ) and the corresponding numbers in @xcite is very small , whereas ( [ br - bskk - alt ] ) did not change at all .
the cdf collaboration has recently reported the first measurements of the cp - averaged @xmath38 branching ratio @xcite , corresponding to the preliminary result @xmath278 the agreement with our theoretical sm predictions is very impressive , giving further support to our strategy .
we look forward to better data and hope that also first measurements of the cp - violating @xmath38 observables will be available in the near future . here
@xmath279 would be particularly exciting , since this asymmetry may well be affected by np contributions to @xmath272@xmath273 mixing , which would manifest themselves then as a discrepancy to ( [ acp - mix - bskk ] ) . by the time this measurement will be available , the uncertainty of the sm prediction given there should be further reduced thanks to better @xmath20 input data .
r. fleischer , _ phys .
_ * 370 * ( 2002 ) 537 .
b. aubert _
et al . _
[ babar collaboration ] , _ phys .
lett . _ * 91 * ( 2003 ) 241801
et al . _ [ belle collaboration ] , _ phys .
_ * 91 * ( 2003 ) 261801 .
m. beneke , g. buchalla , m. neubert and c.t .
sachrajda , _ phys .
lett . _ * 83 * ( 1999 ) 1914 .
m. beneke and m. neubert , _ nucl . phys .
_ * b675 * ( 2003 ) 333 .
d. cronin - hennessy _ et al .
[ cleo collaboration ] , _ phys . rev . lett . _ * 85 * ( 2000 ) 515 .
buras and r. fleischer , _ eur .
j. _ * c16 * ( 2000 ) 97 .
t. yoshikawa , _ phys .
_ * d68 * ( 2003 ) 054023 .
m. gronau and j.l .
rosner , _ phys .
_ * b572 * ( 2003 ) 43 .
buras , r. fleischer , s. recksiegel and f. schwab , _ eur .
j. _ * c32 * ( 2003 ) 45 .
buras and r. fleischer , _ eur .
j. _ * c11 * ( 1999 ) 93 .
m. neubert , _ jhep _ * 9902 * ( 1999 ) 014 .
r. fleischer and t. mannel , ttp-97 - 22 [ hep - ph/9706261 ] .
y. grossman , m. neubert and a. kagan , _ jhep _ * 9910 * ( 1999 ) 029 .
buras , r. fleischer , s. recksiegel and f. schwab , _ phys .
lett . _ * 92 * ( 2004 ) 101804 .
buras , r. fleischer , s. recksiegel and f. schwab , _ nucl .
_ * b697 * ( 2004 ) 133 .
m. battaglia _ et al .
_ , hep - ph/0304132 .
j. charles _ et al .
_ [ ckmfitter group collaboration ] , hep - ph/0406184 . m. bona _ et al . _ [ utfit collaboration ] , hep - ph/0408079
. a.j .
buras , g. colangelo , g. isidori , a. romanino and l. silvestrini , _ nucl .
_ * b566 * ( 2000 ) 3 .
g. buchalla , g. hiller and g. isidori , _ phys .
_ * d63 * ( 2001 ) 014015 ; + d. atwood and g. hiller , lmu-09 - 03 [ hep - ph/0307251 ] .
v. barger , c. w. chiang , p. langacker and h. s. lee , phys .
b * 598 * , 218 ( 2004 ) b. aubert _
et al . _
[ babar collaboration ] , babar - conf-04/30 [ hep - ex/0408062 ] .
b. aubert _
et al . _
[ babar collaboration ] , babar - conf-04/044 [ hep - ex/0408080 ] .
b. aubert _
et al . _ [ babar collaboration ] , babar - conf-04/035 [ hep - ex/0408081 ] .
k. abe _ et al . _ [ belle collaboration ] , belle - conf-0406 [ hep - ex/0408101 ] .
b. aubert _
et al . _
[ babar collaboration ] , babar - conf-04/047 [ hep - ex/0408089 ] .
et al . _ [ belle collaboration ] , _ phys .
lett . _ * 93 * ( 2004 ) 021601
. y. chao _
et al . _
[ belle collaboration ] , belle preprint 2004 - 20 [ hep - ex/0407025 ] .
k. abe _ et al . _ [ belle collaboration ] , belle - conf-0435 [ hep - ex/0409049 ] .
b. aubert _
et al . _
[ babar collaboration ] , hep - ex/0407057 .
et al . _
[ belle collaboration ] , hep - ex/0408100 .
y. l. wu and y. f. zhou , hep - ph/0409221 .
x. g. he , c. s. li and l. l. yang , hep - ph/0409338
. a.j .
buras and r. fleischer , _ phys .
_ * b341 * ( 1995 ) 379 .
m. ciuchini , e. franco , g. martinelli , l. silvestrini , _ nucl .
_ * b501 * ( 1997 ) 271 ; + c. isola , m. ladisa , g. nardulli , t.n .
pham and p. santorelli , _ phys .
_ * d64 * ( 2001 ) 014029 and * d65 * ( 2002 ) 094005 ; + m. ciuchini , e. franco , g. martinelli , m. pierini and l. silvestrini , _ phys .
_ * b515 * ( 2001 ) 33 .
buras , r. fleischer and t. mannel , _ nucl .
_ * b533 * ( 1998 ) 3 .
bauer , d. pirjol , i.z . rothstein and i.w .
stewart , hep - ph/0401188 .
m. gronau , o.f .
hernandez , d. london and j.l .
rosner , _ phys .
_ * d52 * ( 1995 ) 6374 .
r. fleischer , _ phys .
_ * b365 * ( 1996 ) 399 .
m. gronau , d. pirjol and t.m .
yan , _ phys .
_ * d60 * ( 1999 ) 034021 [ e : * d69 * ( 2004 ) 119901 ] .
s. eidelman _
et al . _
[ particle data group ] , _ phys .
_ * b592 * ( 1994 ) 1 .
b. aubert _
_ [ babar collaboration ] , babar - conf-04/38 [ hep - ex/0408127 ] .
et al . _ [ belle collaboration ] , belle - conf-0436 [ hep - ex/0408111 ] .
r. fleischer and j. matias , _ phys .
_ * d66 * ( 2002 ) 054009 .
r. fleischer , g. isidori and j. matias , _ jhep _ * 0305 * ( 2003 ) 053 .
r. fleischer , _ nucl .
_ * b671 * ( 2003 ) 459 .
m. verderi [ for the babar collaboration ] , babar - talk-04 - 011 [ hep - ex/0406082 ] .
et al . _
[ belle collaboration ] , belle - conf-0438 [ hep - ex/0408104 ] .
a. ali , e. lunghi and a.y .
parkhomenko , _ eur .
j. _ * c36 * ( 2004 ) 183 .
chiang , m. gronau , j.l .
rosner and d.a .
suprun , _ phys .
rev . _ * d70 * ( 2004 ) 034020 .
x. g. he and b. h. j. mckellar , hep - ph/0410098 .
r. fleischer and s. recksiegel , cern - ph - th/2004 - 146 [ hep - ph/0408016 ] , to appear in _ eur .
j. _ * c*. g. buchalla and a.s .
safir , lmu-24 - 03 [ hep - ph/0406016 ] .
c.h . chang and h.n .
li , _ phys .
_ * d55 * ( 1997 ) 5577 ; + t.w . yeh and h.n .
li , _ phys .
_ * d56 * ( 1997 ) 1615 ; + h.y .
cheng , h.n .
li and k.c .
yang , _ phys .
_ * d60 * ( 1999 ) 094005 ; + y.y .
keum , h.n .
li and a.i .
sanda , _ phys .
_ * b504 * ( 2001 ) 6 , and references therein .
keum and a.i .
sanda , _ phys .
_ * d67 * ( 2003 ) 054009 .
bauer , d. pirjol and i.w .
stewart , _ phys .
_ * d65 * ( 2002 ) 054022 ; + c.w .
bauer , s. fleming , d. pirjol , i.z .
rothstein and i.w .
stewart , _ phys .
_ * d66 * ( 2002 ) 014017 ; + c.w .
bauer , d. pirjol and i.w .
stewart , _ phys .
_ * d66 * ( 2002 ) 054005 ; + i.w .
stewart , hep - ph/0308185 , and references therein .
t. feldmann and t. hurth , hep - ph/0408188 .
r. fleischer , _ eur .
j. _ * c16 * ( 2000 ) 87 .
et al . _
[ na48 collaboration ] , _ phys .
_ * b465 * ( 1999 ) 335 .
a. alavi - harati _ et al . _ [ ktev collaboration ] , _ phys . rev
. lett . _ * 83 * ( 1999 ) 22 .
et al . _
[ na48 collaboration ] , _ phys .
_ * b544 * ( 2002 ) 97 .
a. alavi - harati _ et al .
_ [ ktev collaboration ] , _ phys . rev . _ * d67 * ( 2003 ) 012005 .
buras and m. jamin , _ jhep _ * 0401 * ( 2004 ) 048 .
a. pich , ific/04 - 57 [ hep - ph/0410215 ] .
m. beneke , g. buchalla , m. neubert and c.t .
sachrajda , _ nucl .
_ * b606 * ( 2001 ) 245 .
buras , f. schwab and s. uhlig , tum - hep-547 [ hep - ph/0405132 ] . z. ligeti , lbnl-55944 [ hep - ph/0408267 ]
. k. anikeev _ et al .
_ , fermilab - pub-01/197 [ hep - ph/0201071 ] .
; + g. punzi [ for the cdf collaboration ] , talk at ichep 2004 , beijing , china , 1622 august 2004 , http://ichep04.ihep.ac.cn/. a.f .
falk , a.l .
kagan , y. nir and a.a .
petrov , _ phys .
_ * d57 * ( 1998 ) 4290 .
p. ball and r. zwicky , ippp-04 - 23 [ hep - ph/0406232 ] .
r. fleischer and s. recksiegel , cern - ph - th/2004 - 171 [ hep - ph/0409137 ] .
m. neubert and j.l .
rosner , _ phys .
_ * b441 * ( 1998 ) 403 ; _ phys .
lett . _ * 81 * ( 1998 ) 5076 .
m. gronau , j.l . rosner and d. london , _
_ * 73 * ( 1994 ) 21 .
s. rai choudhury , n. gaur and a. s. cornell , hep - ph/0402273 . g. isidori , c. smith and r. unterdorfer , _ eur .
j. _ * c36 * ( 2004 ) 57 . | recently , we developed a strategy to analyse the @xmath0 data .
we found that the @xmath1 measurements can be accommodated in the standard model ( sm ) through large non - factorizable effects . on the other hand , our analysis of the ratios @xmath2 and @xmath3 of the cp - averaged branching ratios of the charged and neutral @xmath4 modes , respectively , suggested new physics ( np ) in the electroweak penguin sector , which may have a powerful interplay with rare decays . in this paper , we confront our strategy with recent experimental developments , addressing also the direct cp violation in @xmath5 , which is now an established effect , the relation to its counterpart in @xmath6 , and the first results for the direct cp asymmetry of @xmath7 that turn out to be in agreement with our prediction .
we obtain hadronic @xmath0 parameters which are almost unchanged and arrive at an allowed region for the unitarity triangle in perfect accordance with the sm .
the `` @xmath4 puzzle '' persists , and can still be explained through np , as in our previous analysis .
in fact , the recently observed shifts in the experimental values of @xmath2 and @xmath3 have been predicted in our framework on the basis of constraints from rare decays .
conversely , we obtain a moderate deviation of the ratio @xmath8 of the cp - averaged @xmath5 and @xmath9 rates from the current experimental value . however , using the emerging signals for @xmath10 modes , this effect can be attributed to certain hadronic effects , which have a minor impact on @xmath2 and do not at all affect @xmath3 .
our results for rare decays remain unchanged .
cern - ph - th/2004 - 216 + tum - hep-561/04 + mpp-2004 - 141 + hep - ph/0410407 * andrzej j. buras,@xmath11 robert fleischer,@xmath12 stefan recksiegel@xmath11 and felix schwab@xmath13 * @xmath11 _ physik department , technische universitt mnchen , d-85748 garching , germany _
@xmath12 _ theory division , department of physics , cern , ch-1211 geneva 23 , switzerland _
@xmath14 _ max - planck - institut fr physik werner - heisenberg - institut , d-80805 munich , germany
_ october 2004 |
double white dwarf ( dwd ) binaries are considered to be one of the most promising sources of gravitational waves for lisa , the proposed laser interferometer space antenna @xcite . if , as has been predicted @xcite , close dwd pairs are the end product of the thermonuclear evolution of a sizeable fraction of all binary systems , then dwd binaries must be quite common in our galaxy and the gravitational waves ( gw ) emitted from these systems
may be a dominant source of background noise for lisa in its lower frequency band , @xmath1 @xcite .
dwd binaries are also believed to be ( one of the likely ) progenitors of type ia supernovae @xcite in situations where the accreting white dwarf exceeds the chandrasekhar mass limit , collapses toward nuclear densities , then explodes . because its instruments will have sufficient sensitivity to detect gw radiation from close dwd binaries throughout the volume of our galaxy , lisa will provide us with an unprecedented opportunity to study this important tracer of stellar populations and it will provide us with a much better understanding of the formation and evolution of close binary systems in general .
clearly , a considerable amount of astrophysical insight will be gained from studying the dwd population as a guaranteed source for lisa . broadly speaking , our galaxy s dwd binary population should be dominated by systems that are in two distinctly different evolutionary phases : an `` inspiral '' phase , during which both stars are detached from their respective roche lobes ; and a semi - detached , `` stable mass - transfer '' phase during which the less massive star fills its roche lobe and is slowly transferring mass to its more massive companion . while dwd binaries may encounter other interesting evolutionary phases for example , a phase of so - called common envelope evolution , or a phase of rapid , unstable mass transfer the inspiral and stable mass - transfer phases are expected to dominate the population because they are especially long - lived .
it should be noted that we already have a handle on the size of the galactic population of dwd binaries from optical , uv , and x - ray observations . in the immediate solar neighborhood
, there are 18 systems @xcite known to be undergoing a phase of stable mass transfer ( am cvn being the prototype ) and the eso sn ia progenitor survey ( spy ) has detected nearly 100 detached dwd systems @xcite . at present , orbital periods and the component masses for 24 detached dwd systems have been determined ( see table 3 of @xcite and references therein ) , five of which come from the spy survey . during both of these
relatively long - lived evolutionary phases , a dwd system s orbital period ( and corresponding gw frequency ) changes on a time scale that is governed by the rate at which angular momentum is being lost from the system due to gravitational radiation , that is , the so - called `` chirp '' time scale ( see the discussion associated with eq .
[ tau_chirp ] , below ) .
while the system is detached , the orbital separation slowly decreases so the system should emit a gw signal with a characteristic `` chirp '' signature , that is , the frequency and amplitude of the gw signal should monotonically increase with time . during a phase of stable mass transfer
, however , the orbital separation steadily increases so the gw signal should exhibit an inverse - chirp character where by its frequency and amplitude should steadily decrease with time .
the primary objective of our present study is to analyze the imprint that both of these relatively long - lived phases of evolution will have on our galaxy s dwd binary population , as viewed by lisa .
lisa s capabilities as a gw detector are usually discussed in the context of the @xmath2 diagram , where the gw frequency @xmath0 is measured in hertz , and the gw amplitude @xmath3 is a dimensionless `` strain '' ( generally quoted per @xmath4 , reflecting the frequency resolution of the data stream ) .
a useful analogy can be drawn between this amplitude - frequency diagram and the astronomy community s familiar color - magnitude ( cm ) diagram . directly from photometric measurements , astronomers can produce a cm diagram that is based on the apparent brightness ( the apparent magnitude @xmath5 ) of various sources .
however , a determination of the intrinsic brightness of each source must await the determination of the distance @xmath6 to each source and the corresponding conversion of each measured value of @xmath5 to an absolute magnitude @xmath7 .
lisa s measurement of @xmath8 for a given astrophysical source is analogous to a measurement of @xmath5 ; it only tells us how bright the gw source appears to be on the sky .
a determination of the intrinsic brightness of each gw source must await a determination of the source distance @xmath6 and the corresponding conversion of each measured value of ( the apparent brightness ) @xmath3 to a quantity that represents the `` absolute '' brightness of the source .
for lisa sources , the relevant quantity ( analogous to @xmath7 ) is @xmath9 .
astronomers realize that the underlying physical properties of stars , their evolution , and their relationship to one another in the context of stellar populations can only be ascertained from a cm diagram if @xmath7 , rather than @xmath5 , is used to quantify stellar magnitudes . by analogy
, it should be clear that the underlying physical properties of dwd systems , their evolution , and their relationship to one another in the context of stellar populations can be ascertained only if the observational properties of such systems are displayed in a @xmath10 diagram , rather than in a plot of @xmath8 versus @xmath11 .
for this reason , our discussion of dwd systems will be presented in the context of this more fundamental , but rather under - utilized , `` absolute '' amplitude - frequency domain .
as we investigate the evolution of dwd systems across lisa s `` absolute '' amplitude - frequency domain , we will utilize a simplified description of the two long - lived evolutionary phases mentioned above .
we will assume ( 1 ) all orbits are circular ; ( 2 ) the orbital frequency is related to the orbital separation via kepler s third law ; ( 3 ) the spin of both stars can be ignored so that each system s total angular momentum is given by the point - mass expression for orbital angular momentum ; ( 4 ) the total mass of each system is conserved ; and ( 5 ) angular momentum is lost from each system only via the radiation of gravitational waves and that the rate of angular momentum loss is correctly described by a quadrupole radiation formula .
simplification # 4 means , for example , that after the low - mass white dwarf comes into contact with its roche lobe , we will assume that each dwd system evolves along a `` conservative '' mass transfer ( cmt ) trajectory , and simplification # 5 means that we will be ignoring effects that might arise due to direct - impact accretion @xcite . a more thorough analysis that removes some or all of these simplifications is likely to provide additional valuable insight into the evolution of dwd populations ; @xcite , for example ,
have expressed concern that the time - rate - of - change of the gw frequency for many of lisa s most interesting sources will not be correctly interpreted without a proper treatment of tides . in this context , our analysis should be viewed as an important first step in what is likely to be a long - term study of the evolution of the dwd binary population across lisa s `` absolute '' amplitude - frequency domain . at the outset
, we acknowledge that our understanding of this subject has benefitted significantly from the insight of others whose work has preceded ours ; most notably , we recognize the insightful publications by @xcite , @xcite , @xcite , @xcite , @xcite , and @xcite .
in the quadrupole approximation @xcite , the time - dependent gravitational - wave strain , @xmath12 , generated by a point mass binary system in circular orbit has two polarization states .
the plus and cross polarizations of @xmath12 generically take the respective forms , , we will assume that the binary system is being viewed `` face on '' so that the measured peak - to - peak amplitude of the two polarization states are equal and at their maximum value , given by @xmath13 .
if the orbit is inclined to our line of sight , the inclination angle can be determined as long as a measurement is obtained of both polarization states as shown , for example , by @xcite . because our discussion focuses on galactic dwd binaries , we ignore the effects of cosmological expansion . ]
@xmath14 $ ] and @xmath15 $ ] , where the time - dependent phase angle , @xmath16 where @xmath17 is the phase at time @xmath18 , @xmath19 is the frequency of the gravitational wave measured in hz , @xmath20 is the angular velocity of the binary orbit given in radians per second , and the characteristic amplitude of the wave , @xmath21 where @xmath22 is the gravitational constant , @xmath23 is the speed of light , @xmath6 is the distance to the source , @xmath24 and @xmath25 are the masses of the two stars , and @xmath26 is the distance between the stars . if the principal parameters of the binary system do not change with time , then @xmath0 and @xmath13 will both be constants and the phase angle @xmath27 will vary only linearly in time , so the source will emit ( monochromatic ) `` continuous - wave '' radiation . if , however , any of the binary parameters @xmath24 , @xmath25 , @xmath26 , or @xmath20 vary with time , then @xmath13 and/or @xmath0 will also vary with time in accordance with the physical process that causes the variation . here
we will only be considering the evolution of dwd systems in which the basic system parameters vary on a timescale that is long compared to @xmath28 .
we know from the mass - radius relationship for white dwarfs ( see the discussion associated with eq . [ mass_radius ] , below ) that the less massive star in a dwd binary will always have the larger radius . therefore , in a dwd system that is undergoing mass transfer , we can be certain that the less massive star is the component that is filling its roche lobe and is transferring ( donating ) mass to its companion ( the more massive , accretor ) . with this in mind ,
throughout the remainder of our discussion we will identify the two stars by the subscripts @xmath29 ( for donor ) and @xmath26 ( for accretor ) , rather than by the less descript subscripts @xmath30 and @xmath31 , and will always recognize that the subscript @xmath29 identifies the less massive star in the dwd system .
this notation will be used even during evolutionary phases ( such as a gr - driven inspiral phase ) when the two stars are detached and therefore no mass - transfer is taking place .
furthermore , we will frequently refer to the total mass of the system , @xmath32 and the mass ratio , @xmath33 which will necessarily be confined to the range @xmath34 because @xmath35 .
also , it will be understood that the limiting mass for either white dwarf is the chandrasekhar mass , @xmath36 .
as mentioned above , throughout this investigation we will assume that kepler s third law provides a fundamental relationship between the angular velocity and the separation of dwd binaries , that is , @xmath37 relation ( [ kepler ] ) allows us to replace either @xmath20 or @xmath26 in favor of the other parameter in eq .
( [ hnorm ] ) .
furthermore , we will find it useful to interchange one or both of these parameters with the binary system s orbital angular momentum @xmath38 where , @xmath39 is the ratio of the system s reduced mass to its total mass . for our future discussion , it is useful to express the gravitational wave amplitude @xmath13 and frequency @xmath0 in terms of @xmath40 and @xmath41 .
so , @xmath42 we note as well that the so - called `` chirp mass '' @xmath43 of a given system @xcite is obtained from @xmath44 and @xmath41 via the relation , @xmath45
[ theory_section . ] detached dwd binaries slowly inspiral toward one another as they lose orbital angular momentum due to gravitational radiation .
assuming that @xmath44 and @xmath46 remain constant during this phase of evolution , eqs.([hnorm_j_q ] ) and ( [ f_j_q ] ) can be combined to give , @xmath47^{1/3 } = 5.38 ~[k^5 f^2]^{1/3}~\mathrm{meters } \ , , \label{h_f_relationship}\end{aligned}\ ] ] where the dimensionless mass parameter , @xmath48 has been defined such that it acquires a maximum value of unity in the limiting case where @xmath49 ; otherwise , @xmath50 .
( we note that in the limiting case of @xmath51 , the chirp mass of the system is @xmath52 . ) from expression ( [ h_f_relationship ] ) , we see that the trajectory of an inspiraling , detached dwd binary in the amplitude - frequency diagram can be determined without specifying precisely the rate at which angular momentum is lost from the system . specifically , because @xmath53 , trajectories of inspiraling dwd binaries will be straight lines with slope @xmath54 in a plot of @xmath55 versus @xmath56 . by way of illustration ,
inspiral trajectories for systems having three different total masses ( @xmath57 , @xmath58 , and @xmath59 ) but all having the same @xmath60 mass ratio are displayed in the top panel of figure [ logh_logf_q0.66 ] .
the detached inspiral phase of the evolution of a dwd binary will terminate when the binary separation @xmath26 first becomes small enough that the less massive white dwarf fills its roche lobe . from eggleton
s mass - radius relationship for zero - temperature white dwarfs , as quoted by @xcite and also in @xcite , we know that the radius of the donor @xmath61 is , @xmath62^{1/2 } \biggl[1 + 3.5\biggl(\frac{m_d}{m_\mathrm{p}}\biggr)^{-2/3}+ \biggl(\frac{m_d}{m_\mathrm{p}}\biggr)^{-1}\biggr]^{-2/3 } \ , , \label{mass_radius}\end{aligned}\ ] ] where @xmath63 .
furthermore , from @xcite we find that the roche - lobe radius @xmath64 is , @xmath65 .
% = \frac{j_\mathrm{orb}^2}{gm_\mathrm{tot}^3}\frac{(1+q)^4}{q^2 } % \biggl [ \frac{0.49 \ q^{2/3}}{0.6 \
q^{2/3 } + \ln(1 + % q^{1/3})}\biggr ] \ , .
\label{roche_radius}\end{aligned}\ ] ] the orbital separation and the corresponding gw amplitude @xmath66 and frequency @xmath0 at which the inspiral phase terminates can therefore be determined uniquely for a given donor mass @xmath67 and system mass ratio @xmath46 by setting @xmath68 and combining expressions ( [ mass_radius ] ) and ( [ roche_radius ] ) accordingly .
the termination points of the three inspiral trajectories marked by plus symbols in the top panel of figure [ logh_logf_q0.66 ] have been calculated in this manner .
the curve connecting the sequence of plus symbols in figure [ logh_logf_q0.66 ] traces out the locus of points that define the termination points of the detached inspiral phase of numerous other dwd systems that have mass ratios @xmath60 but that have values of @xmath44 ranging from @xmath57 to @xmath69 .
as a dwd system fills its roche lobe and starts transferring mass to its companion , it evolves to lower frequencies and amplitudes . without knowing the precise rate
at which this phase of mass transfer proceeds , we can map out the evolutionary trajectory of various sytems in the @xmath70 diagram if we assume that the system s total mass is conserved and the donor s radius is marginally in contact with its roche lobe . by way of illustration ,
the bottom panel of figure [ logh_logf_q0.66 ] shows two stable , conservative mass - transfer ( cmt ) trajectories : the ( blue ) dashed trajectory is for a system of mass @xmath71 ; the ( pink ) dotted trajectory is for a system of mass @xmath72 .
we have assumed that both of these systems began the mass - transfer phase of their evolution with an initial mass ratio @xmath73 .
hence , the starting point of both trajectories lies on the termination boundary for inspiralling systems having mass ratios of @xmath74 . for systems with @xmath75 , the mass of the accretor will exceed @xmath76 when @xmath46 drops below the value , @xmath77 with the expectation that something catastrophic ( _ e.g. _ , a type ia supernova explosion ) will occur when this happens , it is reasonable to assume that cmt trajectories with @xmath78 will terminate at a point in the amplitude - frequency diagram that is marked by @xmath79 .
the locus of points that is defined by the termination points of these trajectories defines another interesting astrophysical boundary in lisa s `` absolute '' amplitude - frequency diagram .
this termination boundary has been drawn as a thick , ( green ) dashed curve in the bottom panel of figure [ logh_logf_q0.66 ] . the inspiral trajectory drawn for @xmath80 ( @xmath81 ) and the curve marking the termination of various inspiral trajectories in the top panel of figure [ logh_logf_q0.66 ]
define boundaries in the amplitude - frequency domain outside of which no dwd system should exist if it has a mass ratio @xmath82 .
analogous domain boundaries can be constructed readily for other values of @xmath46 . for each value of @xmath46 ,
the shapes of the bounding curves are roughly the same as shown in the top panel of figure [ logh_logf_q0.66 ] , but for higher ( lower ) values of @xmath46 the right - hand termination boundary shifts to higher ( lower ) frequencies and the limiting inspiral trajectory ( set by the parameter @xmath83 ) shifts to higher ( lower ) strain amplitudes . given our present understanding of the structure of white dwarfs
, it seems extremely unlikely that any dwd binary systems can exist outside of the domain that is defined by the bounding curves for systems with @xmath84 ( see , for example , figure [ conclusion_figure ] ) .
[ logh_logf_q0.66 ] up to this point , we have described key features of dwd evolutionary trajectories in lisa s `` absolute '' amplitude - frequency diagram without referring to the rate at which the evolution of any given system proceeds . here
we investigate the time scales on which significant changes in various system parameters and , as a consequence , the rate at which measurable changes in the gw signature occur . using eqs.([hnorm_j_q ] ) and ( [ f_j_q ] ) , and assuming @xmath44 as constant , we can write the time - rate - of - change of the amplitude and frequency , for both the inspiral and cmt phase , as follows : @xmath85 during the inspiral phase of dwd binary evolution @xmath86 , so the evolution is driven entirely by the loss of angular momentum due to gravitational radiation . according to @xcite ( see also @xcite ) , the time - dependent behavior of @xmath40 is described by the relation , @xmath87 where the inspiral evolutionary time scale is , @xmath88 = \frac { 5}{64 \pi^2 } \biggl ( \frac{c}{rh f^2 } \biggr ) \ , .
\label{tau_chirp}\end{aligned}\ ] ] hence , @xmath89 for the cmt phase , @xmath90 .
hence both the terms on the right - hand side of eq .
( [ dlnf_mt ] ) affect the evolution . from the work of @xcite and @xcite
, we deduce that during a phase of stable cmt ( see @xcite for a detailed derivation ) , @xmath91 where @xmath92 is a parameter that is of order unity .
the quantities @xmath93 and @xmath94 represent the change in the donor s radius and roche lobe radius , respectively , as a function of its mass @xcite to represent the change in the radius of the donor with respect to its mass , whereas @xcite use @xmath95 indicating the donor as the secondary star .
also , they use a slightly different definition for @xmath94 . the relation between their s ( @xmath96 ) and
our s ( @xmath94 ) is : @xmath97 . ] .
it should be emphasized that the timescale on which dwd systems evolve during both the inspiral and cmt phases is @xmath98 , as indicated by eqs .
( [ djorb_dt ] ) and ( [ tau_mt ] ) .
it is for this reason that we have drawn various `` chirp isochrones '' in both panels of figure [ logh_logf_q0.66 ] ; as can be ascertained from eq.([tau_chirp ] ) , each isochrone depends only on the product of @xmath99 and @xmath100 and , hence , has a slope of @xmath101 in the figure panels . in practice ,
for a given source , lisa will be unable to measure changes in the strain amplitude @xmath3 at the levels predicted by expression ( [ dlnf_mt ] ) because variations in @xmath3 do not accumulate secularly over time and , in particular , will not contribute to the observed phase of the signal .
since , as indicated in eq.([phasedefinition ] ) , the phase depends on the time - rate - of - change of the gw frequency we will concentrate on the @xmath102 expression in eq .
( [ dlnf_mt ] ) from here onwards .
combining eqs .
( [ dlnf_mt ] ) , ( [ djorb_dt ] ) and ( [ tau_mt ] ) , a concise form of this expression can be written to indicate its behavior during both the inspiral @xcite and cmt @xcite phases as : @xmath103 \label{fdot_eq}\ ] ] where , @xmath104 since @xmath105 depends on both @xmath44 and @xmath46 , there exists a critical @xmath106 for which @xmath107 . for systems with @xmath108 ,
a phase of unstable mass transfer ensues and our present analysis becomes invalid in that regime .
hence @xmath109 represents the limiting mass ratio for a system to be in stable cmt phase .
as we have discussed , the physical processes that drive the evolution of dwd binaries operate on a `` chirp '' timescale , and @xmath110 is typically much longer than the operational time for lisa ( assumed one year here ) . hence ,
the time - variation of a given system s gw frequency @xmath111 can be well approximated by a truncated taylor series expansion in time and , using eq .
( [ phasedefinition ] ) , the observed phase of the gw signal @xmath112 can be written in the form @xmath113 \ , , \ ] ] where @xmath114 is the signal frequency at time @xmath18 , @xmath115 , @xmath116 , and so on . if we truncate the taylor series at @xmath117 and this observed signal ( o ) is compared to a computed template ( c ) that assumes a continuous - wave signal and therefore has a phase that increases only linearly with time , @xmath118 , the amount of time for the o - c phase difference to reach @xmath119 will be , @xmath120 substituting for @xmath117 from eq .
( [ fdot_eq ] ) , we obtain @xmath121^{1/2 } \ , .\label{to - cexpression}\ ] ] if the function @xmath122 in eq.([to - cexpression ] ) is independent of @xmath3 and @xmath0 as is the case for the inspiral phase of dwd evolutions then curves of constant @xmath123 in the amplitude - frequency diagram will be straight lines having a slope of @xmath124 .
an example of this curve , assuming @xmath125 year , is shown in figure [ conclusion_figure ] joining the two low frequency points .
any inspiral system that lies to the right of this constant @xmath123 line will lose phase coherence in less than one year of observation if one assumes that it emits continous - wave radiation .
an analogous one - year demarcation boundary can be drawn for dwd binaries that are undergoing a phase of stable , cmt by evaluating eq .
( [ to - cexpression ] ) using the function @xmath126 . because this function generally is of order unity , the one - year demarcation boundary for mass - transferring systems is generally well - approximated by the line segment that marks the one - year demarcation boundary for inspiral systems .
lisa will be able to measure frequency evolution for dwd systems that lie to the right of the @xmath125 year line and , as will be discussed in the following section , it will then be possible to determine the distances and masses for these systems .
the downside is that millions of dwd systems will lie to the left of the one year @xmath123 curve in the low frequency region , for which frequency evolution can not be measured .
an analysis of a one - year - long lisa data stream that utilizes a proper set of frequency - varying strain templates should be able to determine the rate at which the gw frequency is changing in both inspiral and mass - transfering dwd systems .
when used in conjunction with the measurement of @xmath13 and @xmath0 , an accurate measurement of @xmath117 for any source should permit a determination of the distance to the source @xmath6 and should give information about the chirp mass @xmath43 or the individual component masses of the binary system , as follows .
equation ( [ h_f_relationship ] ) provides a relation between the three unknown binary system parameters @xmath127 and @xmath46 , and the experimentally measurable parameters @xmath0 and @xmath13 , namely , @xmath128 ^ 3 & = & \frac{\mathcal{m}^5}{r^3 } = \frac{c^{12}}{2 ^ 6 \pi^2 g^5 } \biggl [ \frac{h_\mathrm{norm}^3 } { f^2 } \biggr ] \ , .\label{unknowns1}\end{aligned}\ ] ] a second relation between the unknown astrophysical parameters and measurable ones is provided by combining the expression for @xmath117 in eq .
( [ fdot_eq ] ) with the definition of @xmath110 given in eq .
( [ tau_chirp ] ) .
specifically , we obtain , @xmath129 \ , , \label{unknowns2}\ ] ] with only two equations , of course , it is not possible to uniquely determine all three of the binary s primary system parameters .
however , in the inspiral phase @xmath130 , so @xmath43 and @xmath6 can be determined as was shown in @xcite . during the cmt phase of an evolution , the function @xmath131 is nonzero so eq . ( [ unknowns2 ] )
does not provide an explicit determination of @xmath6 .
however , the requirement that @xmath68 provides an important additional physical relationship between the unknown astrophysical parameters and measurable ones . specifically , by setting @xmath61 from eq .
( [ mass_radius ] ) equal to @xmath64 from eq .
( [ roche_radius ] ) and using kepler s law to write @xmath26 in terms of @xmath0 , we obtain , @xmath132^{1/2 } f & = & \biggl [ \pi^2 ( 0.0114)^3 \frac{m_\mathrm{ch}}{m_\odot } \biggr]^{-1/2 } \frac{m_\mathrm{tot}}{m_\mathrm{\odot } } \biggl ( \frac{q}{1+q } \biggr)h(m_d , q ) \ , , \label{unknowns3}\end{aligned}\ ] ] where , @xmath133^{3/2 } \biggl[1 - \biggl(\frac{m_d}{m_\mathrm{ch } } \biggr)^{4/3}\biggr]^{-3/4 } \nonumber \\ & & \times \biggl[1 + 3.5\biggl(\frac{m_d}{m_\mathrm{p}}\biggr)^{-2/3}+ \biggl(\frac{m_d}{m_\mathrm{p}}\biggr)^{-1}\biggr ] \ , .\label{define_h}\end{aligned}\ ] ] hence , taken together , eqs .
( [ unknowns1])-([unknowns3 ] ) can be used to determine all three primary system parameters
@xmath6 , @xmath44 , and @xmath46 from the three measured quantities , @xmath134 , and @xmath117 .
we are unable to solve this set of equations analytically due to the complexity of the functions @xmath131 and @xmath135 .
however , the formulae that @xcite adopted for @xmath136 and @xmath137 lead to much simpler expressions for both of these functions , namely , @xmath138 $ ] and @xmath139 . in this case eqs .
( [ unknowns1])-([unknowns3 ] ) can be combined to give @xmath46 in terms of @xmath0 and @xmath117 : @xmath140\frac{f^{16}}{[-\dot f]^3 } \
, , \label{pac_q_manuscript}\ ] ] where @xmath141 .
once @xmath46 is known , @xmath6 can be obtained using eq.([unknowns2 ] ) in conjunction with paczyski s @xmath142 relation ; then @xmath44 can be obtained from eq.([unknowns1 ] ) .
@xmath143 ( 2 - 3q ) \ , ; \label{pac_r_manuscript } \\
m_\mathrm{tot } & = & \biggl [ \frac{5 ^ 3 c^{15}}{2^{15}\cdot 3 ^ 3 \pi^8 g^5 } \biggr]^{1/5 } \biggl\ { \frac{(1+q)^{6}(2 - 3q)^3}{q^{3 } } \cdot \frac{[-\dot f]^{3}}{f^{11 } } \biggr\}^{1/5 } \ , .
\label{pac_m_manuscript}\end{aligned}\ ] ] for any @xmath144 , these three equations are valid for mass ratios over the range @xmath145 because , for paczyski s model , @xmath146 independent of @xmath44 .
the solid curves in figure [ fig : ffdot_pac ] illustrate results obtained numerically from a self - consistent solution of eqs .
( [ unknowns1])-([unknowns3 ] ) ; the dashed curves illustrate results obtained analytically from expressions ( [ pac_q_manuscript ] ) and ( [ pac_m_manuscript ] ) . across the parameter domain defined by the two observables @xmath11 and @xmath147 where @xmath148 ^ 3/f^{16}\}^{1/10 } \ , , \label{define_gamma}\end{aligned}\ ] ] is measured in seconds each curve traces a constant @xmath44 `` trajectory '' with @xmath46 varying along each curve , as indicated . at high frequencies ,
each curve begins at a value of @xmath46 that is slightly below @xmath149 ; at low frequencies , the curves have been extended down to @xmath150 , unless @xmath75 , in which case the curve has been terminated at the value @xmath151 , as given by eq .
( [ qchandra ] ) .
the general behavior of these curves can best be understood by analyzing analytic expression ( [ pac_q_manuscript ] ) . over the relevant range of mass ratios @xmath152 , the analytic function , @xmath153^{-1/10 } ~~\mathrm{seconds } \ , , \end{aligned}\ ] ] reaches a minimum value ( @xmath154 ) when @xmath155 , where @xmath156 moving from high frequency to low frequency along each @xmath44 `` trajectory , '' the function @xmath157 steadily drops until @xmath155 and @xmath158 .
( this behavior holds for the solid curves as well as the dashed curves , although the precise values of @xmath159 and @xmath160 are different for each solid curve . )
when @xmath46 drops below @xmath160 [ based on the function @xmath79 , this will only happen along curves for which @xmath161 , each curve climbs back above @xmath159 , reflecting the fact that eq .
( [ pac_q_manuscript ] ) admits two solutions over the relevant range of mass ratios .
this , in turn , implies that for mass - transferring dwd systems that have @xmath162 , a measurement of @xmath117 will generate two possible solutions rather than a unique solution for the pair of key physical parameters @xmath163 .
once lisa has measured @xmath117 as well as @xmath0 for a given dwd system , figure [ fig : ffdot_pac ] provides a graphical means of determining the values of @xmath44 and @xmath46 for the system , assuming it is undergoing a phase of stable cmt .
we do not expect that lisa will probe the entire parameter space depicted in this figure , however . as discussed above
, we expect that lisa will only be able to detect frequency changes in systems for which @xmath164 .
using expression ( [ to - csimple ] ) , this means that lisa will only be able to measure @xmath117 for systems that have , @xmath165 the dashed black line in figure [ fig : ffdot_pac ] with a slope of @xmath166 that is labeled `` @xmath167 '' shows this boundary ; the parameter regime that can be effectively probed by lisa lies above and to the right of this line .
once the distance has been determined to individual lisa sources , as was outlined in [ q_and_mtot ] , it will be possible to place them in an `` absolute '' amplitude - frequency diagram , that is , in a @xmath168 diagram .
the location of individual sources in such a diagram should help us understand a great deal about our galaxy s dwd population .
our expectation is that systems in different evolutionary states will fall into several distinct sub - domains of lisa s `` absolute '' amplitude - frequency diagram , and the diagram will exhibit natural zones of avoidance as well .
our models of inspiral and cmt systems permit us to predict where the boundaries will lie between these various population sub - domains .
as depicted in figure [ conclusion_figure ] , the principal sub - domains can be identified as follows .
* _ zones of avoidance _ : no dwd systems will be found in the region above the boundary line that is defined by an inspiral trajectory with @xmath51 ; according to eq .
( [ h_f_relationship ] ) , this boundary is defined by the expression , @xmath169 also , no dwd systems will be found in the region below the bounding curve that is defined by the inspiral termination boundary for systems with @xmath84 ; over the region of parameter space shown in figure [ conclusion_figure ] this bounding curve is given approximately by the expression , @xmath170 both of these population boundaries have been drawn as solid red curves in figure [ conclusion_figure ] .
detached , inspiraling dwd systems will be found throughout the domain whose upper and lower borders are defined , respectively , by boundary curves * a * and * b*. this region of parameter space can be subdivided into the following two principal population sub - domains .
* _ region i _ : only inspiraling dwd systems will be found in the region of parameter space that is bounded above by curve * a * and below by the locus of points that identifies semi - detached systems for which @xmath171 ( the green solid curve in figure [ conclusion_figure ] ) ; over the region of parameter space shown in figure [ conclusion_figure ] , this curve is given approximately by the expression , @xmath172 * _ region ii _ : mass - transferring dwd systems will only be found in the region of parameter space that is bounded above by curve * c * and below by curve * b*. * _ region iii _ : an inspiraling dwd system that has @xmath108 value will encounter an unstable rather than a stable phase of mass - transfer when the less massive star initially makes contact with its roche lobe .
the region of parameter space in which these systems will be found at the onset of mass - transfer is bounded below by curve * b * and above by two intersecting curve segments : at @xmath173 , curve * c * defines the upper boundary ; for @xmath174 , the relevant upper boundary has been drawn as a blue solid curve in figure [ conclusion_figure ] and is given approximately by the expression , @xmath175 * _ region iv _ : dwd systems that enter a phase of stable cmt with a total system mass @xmath75 will be found in a region of parameter space that is bounded above by curve * c * and below by two intersecting curve segments : for @xmath176 , curve * d * defines the lower boundary ; for @xmath177 , the relevant lower boundary has been drawn as a pink solid curve in figure [ conclusion_figure ] and is given approximately by the expression , @xmath178 boundary curve * e * is defined by the stable cmt trajectory for a system with @xmath179 . an additional demarcation line has been drawn in figure [ conclusion_figure ] that is associated with the @xmath180 yr boundary .
this `` time boundary '' line * t * is defined by setting @xmath181 and @xmath125 yr in eq .
( [ to - cexpression ] ) , that is , it is given by the expression , @xmath182 lisa will be unable to determine distances to dwd systems that lie below and to the left of this demarcation line because their orbital periods and associated gw frequencies are changing so slowly that lisa will be unable to measure with confidence the value of @xmath117 for these systems .
studies of dwd populations will therefore benefit most from the data that lisa collects on systems that lie above and to the right of boundary line * t*.
the segment of this line that bounds _ region ii _ has been drawn as a dashed line to emphasize that it is only an approximate one year boundary for mass - transferring systems .
the particular boundaries * a*-*e * of the population domains that are depicted in figure [ conclusion_figure ] arise as a consequence of the specific mass - radius relationship ( eq . [ mass_radius ] ) and , hence , the equation of state that we have chosen to use to describe the structure of individual white dwarfs . because it is generally believed that the mass - radius relationship given in eq.([mass_radius ] ) represents the properties of white dwarfs quite well , we expect that the population sub - domains and zones of avoidance shown in figure [ conclusion_figure ] will map well onto lisa s observationally determined `` absolute '' amplitude - frequency diagram .
systems that are found on the `` wrong '' side of a given boundary for example , any system that lies in one of the zones of avoidance , or cmt systems that are found in region i will be of particular interest because they may provide evidence that the equation of state that we adopted is not general enough to properly describe this stellar population two examples suffice to illustrate this point .
first , in the presence of tidal stresses , the donor stars in mass - transferring dwd systems may be hotter than assumed here @xcite , which will affect the frequency at which the donor fills the roche lobe .
consequently , the inspiral termination boundary * b * in figure 3 will change .
second , because individual neutron stars obey an entirely different mass - radius relationship and generally seem to have masses @xmath183 , inspiraling double neutron - star systems will likely be distinguishable from dwd systems because they will lie in the zone of avoidance above boundary curve * a * , as defined above .
dwd systems that are undergoing a phase of stable , cmt and that are found to reside in region iv of the `` absolute '' amplitude - frequency diagram can be identified as progenitors of type ia supernovae .
efforts to better understand the origin of supernova explosions in old stellar populations will especially benefit from follow - up studies that identify the optical ( or uv or x - ray ) counterparts to these region iv systems . inspiraling systems that are found to reside in region iii may prove to be equally interesting candidates for follow - up studies .
detached dwd systems in region iii are destined to enter a phase of unstable mass transfer ( likely accompanied by super - eddington accretion , see @xcite ) that will significantly transform the system s properties on a dynamical , rather than a chirp , timescale .
such rapid mass - transfer events may lead to merger of the binary components , perhaps followed by an explosion . as stated above , we are confident that the population sub - domains and zones of avoidance shown in figure [ conclusion_figure ] will map well onto lisa s observationally determined `` absolute '' amplitude - frequency diagram .
we are less confident about the degree to which lisa s measurements of @xmath117 will match the values that are predicted by our simplified model of the slow , orbit - averaged time - evolution of dwd systems ( as displayed , for example , in figure [ fig : ffdot_pac ] ) .
mass - transferring binary systems , in particular , are notoriously messy laboratories .
for example , significant and unexplained variations in the mass - transfer rate can arise in an individual system over times that are much shorter than the gr - driven evolutionary time scale ; this can introduce significant short - term variations in @xmath117 .
also , magnetic fields can be effective at carrying away mass and angular momentum from a system , thereby violating our assumption of conservative mass transfer .
v407 vul and rx j0806 + 1527 @xcite provide perhaps the best examples of the type of confusion that is likely to arise when attempts are made to extract measurements of @xmath117 from a one - year - long lisa data stream .
these are optically identified , mass - transferring binaries that are thought to be dwd systems because their orbital periods are less than ten minutes .
the best available measurements of period variation in both of these systems indicate that their orbits are slowly shrinking , rather than slowing growing larger as would be predicted by our model of stable cmt .
although , in the mean , the long - term evolutionary behavior of these systems is likely to agree with the predictions of our model , fluctuations about the mean that occur on a timescale that is short compared to @xmath110 may totally confound our ability to interpret lisa s measurements of @xmath117 . on the bright side , if a significant number of lisa sources exhibit noticeable deviations away from the mean behavior predicted by our simplified model , in the end we are likely to gain a more complete understanding of the evolution of such systems .
we acknowledge numerous very useful discussions that we have had with j. frank , v. gokhale , p. motl , and x. peng about the evolutionary behavior of dwd systems .
we appreciate the encouragement that we received from s. larson and m. benacquista to push this work to completion after they reviewed a preliminary presentation of our population boundary diagram .
we also appreciate the advice of an anonymous referee , who urged us to shorten the original manuscript making our key point more succinctly .
this work has been supported , in part , by nasa grant nag5 - 13430 and by nsf grants ast-0407070 and phy-0326311 .
anderson , s. f. , haggard , d. , homer , l. , joshi , n. r. , margon , b. , silvestri , n. m. , szkody , p. , wolfe , m. a. , agol , e. , becker , a. c. , henden , a. , hall , p. b. , knapp , g. r. , richmond , m. w. , schneider , d. p. , stinson , g. , barentine , j. c. , brewington , h. j. , brinkman , j. , harvanek , m. , kleinman , s. j. , krzesinski , j. , long , d. , neilsen , e. h. jr . , nitta , a. , & snedden , s. a. 2005 , , 130 , 2230 napiwotzki , r. , yungelson , l. , nelemans , g. , marsh , t. r. , leibundgut , b. , renzini , r. , homeier , d. , koester , d. , moehler , s. , christlieb , n. , reimers , d. , drechsel , h. , heber , u. , karl , c. , & pauli , e .-
2004 , aspc , 318 , 402 | detached , inspiraling and semi - detached , mass - transferring double white dwarf ( dwd ) binary systems are both expected to be important sources for the proposed space - based gravitational - wave detector , lisa . the mass - radius relationship of individual white dwarf stars in combination with the constraints imposed by roche geometries permit us to identify population boundaries for dwd systems in lisa s `` absolute '' amplitude - frequency diagram . with five key population boundaries in place , we are able to identify four principal population sub - domains , including one sub - domain that identifies where progenitors of type ia supernovae will reside
. given one full year of uninterrupted operation , lisa should be able to measure the rate at which the gravitational - wave frequency @xmath0 and , hence , the orbital period is changing in the highest frequency subpopulation of our galaxy s dwd systems .
we provide a formula by which the distance to each dwd system in this subpopulation can be determined ; in addition , we show how the masses of the individual white dwarf stars in mass - transferring systems may be calculated . |
the space @xmath8 of ordinary differential operators of the form @xmath9 has two poisson structures , called the first(second resp . )
adler - gelfand - dickey poissson structure @xcite @xcite .
these two poisson structures are realized by drinfeld - sokolov reductions @xcite via the hamiltonian reductions of the dual space to the affine kac - moody algebra @xmath10 . as a consequence ,
these two poisson structures are compatible and provide a bihamiltonian system .
we view the configuration space @xmath1 of @xmath2-twisted polygons in @xmath3 as the discrete version of the space @xmath8 . for @xmath11 , r. schwartz , v. ovsienko and s. tabachnikov
@xcite introduced the poisson structure @xmath12 corresponding to the first adler - gelfand - dickey poissson structure defined on the coordinate function ring of @xmath13 .
they show that the pentagram map relative to this poisson structure is completely integrable .
they conjecture that there exists a poisson structure @xmath14 on the coordinate function ring of @xmath13 corresponding to the second adler - gelfand - dickey poisson structure(it is also mentioned by b. khsein , f. soloviev in @xcite for @xmath15 in general and many others ) .
moreover , they conjecture that @xmath14 is compatible with their poisson structure .
it is very interesting to consider these two conjectures for @xmath15 in general .
but the poisson structure corresponding to the first adler - gelfand - dickey poissson bracket on @xmath1 for @xmath16 is still missing .
the latter conjecture is not clear even for @xmath4 .
the swapping algebra @xcite , introduced by f. labourie , is used to characterize the atiyah - bott - goldman poisson structure @xcite@xcite for the hitchin component @xcite and the second adler - gelfand - dickey poisson structure for the space of real opers . taken linearity relations into consideration , the quotient algebra rank @xmath15 swapping algebra @xcite is used to characterize the fock
goncharov poisson structure @xcite on cluster @xmath17 moduli space in @xcite . in this paper , for @xmath18 , we define the poisson structures on the coordinate function ring of @xmath1 induced from the swapping algebra @xcite , the induced poisson structures are corresponding to the second adler - gelfand - dickey poisson structures . for @xmath4 ,
the poisson algebra induced from the swapping algebra is the same as takhtajan - faddeev - volkov algebra , which is studied by l. a. takhtajan , l. d. faddeev , a. yu .
volkov in 90s @xcite @xcite @xcite @xcite and many others in lattice virasoro algebra .
we consider a poisson bracket @xmath5 for @xmath6 , which is a `` degenerate '' version of the r. schwartz et al.s poisson bracket @xmath12 for @xmath13 , induced from the discrete version of the first adler - gelfand - dickey poissson bracket by miura transformation . by studying the asymptotic behavior of these two poisson structures , we have we relate asymptotically the dual of the poisson structures @xmath7 and @xmath5 to the virasoro algebra . with these evidences ,
we compare these two poisson structures , we have our main theorem .
the poisson structures @xmath7 and @xmath5 for @xmath6 are compatible .
we calculate some examples for @xmath11 , the two poisson structures @xmath19 and @xmath12 are not compatible . in our further study
, we will modify the definition of the induced poisson bracket @xmath19 to make it work .
we hope that this paper helps to understand the conjectures in general cases .
in this section , we induce a poisson algebra on the configuration space of @xmath2-twisted polygons in @xmath3 from the swapping algebra . we give explicit formulas for @xmath20 . for @xmath4 ,
the induced poisson algebra coincides with takhtajan - faddeev - volkov algebra . for @xmath11 ,
the induced poisson algebra coincides with a lattice @xmath21 algebra . in this subsection ,
we recall briefly some definitions about the swapping algebra introduced by f. labourie .
our definitions here are based on section 2 of @xcite .
let @xmath22 be a quadruple of four points in @xmath23 .
the _ linking number _ between @xmath24 and @xmath25
is @xmath26 such that for any @xmath27 , we define @xmath28 as follows .
remove any point @xmath29 different from @xmath30 in @xmath23 in order to get an interval @xmath310,1[$ ] .
then the points @xmath32 correspond to the real numbers in @xmath310,1[$ ] , @xmath33 whenever @xmath34 ; @xmath35 ; @xmath36 respectively . and @xmath25.,width=384 ] in fact , the value of @xmath37 belongs to @xmath38 , depends on the corresponding positions of @xmath30 , and does not depend on the choice of the point @xmath29 . in figure [ swapswap1 ] , we describe five possible values of @xmath37 .
let @xmath39 be a finite subset of the circle @xmath23 provided with cyclic order .
@xmath40 is a characteristic zero field .
we represent an ordered pair @xmath41 of @xmath39 by the expression @xmath24 . _
the swapping ring of @xmath39 _ is the ring @xmath42/\{xx| \forall x \in \mathcal{p}\}\ ] ] over @xmath40 , where @xmath43 are variables with values in @xmath40 .
notably , @xmath44
if @xmath45 in @xmath46 . then we equip @xmath46 with a poisson bracket defined by f. labourie in section 2 of @xcite .
[ defn2.2 ] _ the swapping bracket _ over @xmath46 is defined by extending the following formula on generators to @xmath46 by _ leibniz s rule _ :
@xmath47 ( here is the case for @xmath48 in section 2 of @xcite . )
[ swappoisson ] the swapping bracket as above verifies the jacobi identity .
so the swapping bracket defines a poisson structure on @xmath46 . _
the swapping algebra of @xmath39 _ is @xmath46 equipped with the swapping bracket .
the _ swapping fraction algebra of @xmath39 _ is fraction ring @xmath49 of @xmath46 equipped with the swapping bracket .
let @xmath50 belong to @xmath39 so that @xmath51 and @xmath52 .
the _ cross fraction _ determined by @xmath53 is the element of @xmath49 : @xmath54 : = \frac{xz}{xt } \cdot \frac{yt}{yz}.\ ] ] let @xmath55 \in \mathcal{q}(\mathcal{p } ) \ ; |\ ; \forall x , y , z , t \in \mathcal{p } , \ ; x \neq t , y \neq z\}$ ] be the set of all the cross - fractions in @xmath49 .
let @xmath56 be the subring of @xmath49 generated by @xmath57 . _
the swapping multifraction algebra of @xmath39 _ is @xmath56 equipped with the swapping bracket .
let @xmath58 be a map from @xmath59 to @xmath3 such that for any @xmath60 in @xmath61 , the images @xmath62 are in general position in @xmath3 .
the configuration space @xmath63 is the space of all these maps up to projective transformations . for @xmath64 , a _
@xmath2-twisted polygon in @xmath3 _ is a map @xmath58 from @xmath59 to @xmath3 such that for any @xmath65 , we have @xmath66 where the _ monodromy _
@xmath67 belongs to @xmath68 .
we say that @xmath58 is _ in general position _ if for any @xmath69 , the images @xmath62 in @xmath3 are in general position .
the _ configuration space of @xmath2-twisted polygons in @xmath3 _ , denoted by @xmath1 , is the space of the @xmath2-twisted polygons in general position in @xmath3 up to projective transformations . without loss of ambiguity ,
let @xmath70 , the swapping multifraction algebra of @xmath39 is @xmath71 .
for each configuration @xmath58 in general position , we associate @xmath58 to a map @xmath72 , where @xmath73 for any @xmath65 .
we lift @xmath58 ( @xmath72 resp . ) to the map @xmath74 ( @xmath75 resp . ) with images in @xmath76 ( @xmath77 resp . ) .
let @xmath78 be a volume form of @xmath76 .
let @xmath79 be a subfraction ring of @xmath49 generated by @xmath80 , where @xmath80 is the set of the cross fractions such that for each factor @xmath81 of the cross fraction we have @xmath82 .
we define a homomorphism @xmath83 from @xmath79 to @xmath84 by extending the following formula on generators : @xmath85 which does not depend on the lifts and the volume form @xmath78 . then the _ induced poisson bracket @xmath86 on @xmath87 _ is @xmath88 for any @xmath89 . 1
. we can modify @xmath72 such that @xmath90 , where @xmath91 in @xmath59 , to get different induced poisson brackets .
the kernel of @xmath92 is characterized by the @xmath93 determinant relations in @xcite .
the natural embedding @xmath94 from @xmath1 to @xmath95 induces a map @xmath94 from @xmath84 to @xmath96 , the _ induced poisson bracket @xmath97 on @xmath98 _ is @xmath99 for any @xmath100 .
let @xmath101:=\frac{a - c}{a - d}\cdot \frac{b - d}{b - c},\ ] ] @xmath102:=\frac{\{x , y\}}{xy}.\ ] ] for any @xmath103 , suppose @xmath104 \in \mathbb{rp}^1 $ ] , let @xmath105.\ ] ] we have @xmath106 is a coordinate system of @xmath6 .
[ propc2 ] we have @xmath107 @xmath108 @xmath109 @xmath110 for @xmath111 with the convention @xmath112 . for the other cases @xmath113 by calculating the swapping bracket , we have the following non - trivial equations @xmath114
\\&= -1 + \frac{r_{k-1}r_{k+2}\cdot r_{k}r_{k+1}}{r_{k-1}r_{k+1 } \cdot r_{k}r_{k+2 } } + \frac{r_{k+2}r_{k+1}\cdot r_{k+3}r_{k}}{r_{k+2}r_{k } \cdot r_{k+3}r_{k+1 } } , \end{aligned}\ ] ] @xmath115 \\&= -\frac{r_{k+2}r_{k+3}\cdot r_{k+1}r_{k}}{r_{k+2}r_{k } \cdot r_{k+1}r_{k+3}}. \end{aligned}\ ] ] since @xmath116 @xmath117 @xmath118 @xmath119 we conclude that we verify the formulas in the proposition . faddeev - takhtajan - volkov algebra coincides with @xmath7 , but the original paper @xcite works on the coordinate system @xmath120 of @xmath6 . for any @xmath121 , let @xmath122\ ] ] @xmath123.\ ] ] by r. schwartz et al.@xcite , @xmath124 is a coordinate system of @xmath13 .
[ propc3 ] we have @xmath125 @xmath126 @xmath127 @xmath128 @xmath129 @xmath130 @xmath131 @xmath132 @xmath133 @xmath134 for @xmath111 with the convention @xmath112 . except the above brackets and their symmetry ones by @xmath135 , all the other brackets between two generators are zero . by observing that @xmath136 @xmath137
we calculate explicitly the swapping brackets between the @xmath138 terms above . by more complicated calculations and the same argument as in proposition [ propc2 ] , we have the above formulas .
we notice that the above formulas coincide with equations ( 2.2.5 ) of @xcite when @xmath139 is replaced by @xmath140 and @xmath141 is replaced by @xmath142 , where @xmath143 are the coordinates for the lattice @xmath21 algebra arising from the study of the quantum invariant ring .
there are two poisson algebras on @xmath6 , one is the faddeev - takhtajan - volkov algebra , another is induced from the discrete version of the first adler - gelfand - dickey poissson bracket by miura transformation . in this section ,
we relate asymptotically the dual of these poisson algebras to the virasoro algebra .
[ defndh2 ] let @xmath144 be an integer .
given a periodic sequence @xmath145 in @xmath146 where @xmath147 for any @xmath65 .
the _ discrete hill s equation _ is the difference equation in @xmath148 : @xmath149 or equivalently @xmath150 for any @xmath151 belongs to @xmath59 .
we define @xmath152 to be a _ discrete hill s operator _ , and @xmath148 is the _ solution _ to the discrete hill s equation . given a discrete hill s operator , by equation [ equhill ] ,
the series @xmath148 is fixed when two initial values @xmath153 , @xmath154 are given . since the difference equation is homogeneous , so up to scalar and @xmath155 projective transformation , there are exactly two linear independent solutions of the discrete hill s operator , which are corresponding to one point of @xmath6 .
conversely , when @xmath2 is odd , given a point of @xmath6 , there is one unique hill s operator corresponding it by the similar argument of proposition 4.1 of @xcite .
[ corhb ] let @xmath156 be odd , let @xmath157 , @xmath158 \in \mathbb{rp}^1 $ ] .
there exists one unique discrete hill s equation [ equhill ] such that @xmath159 and @xmath160 are two linear independent solutions of [ equhill ] and @xmath161= [ f_i:1]$ ] .
let @xmath162 .
let @xmath156 be odd , @xmath163 and @xmath164 are two coordinate systems of @xmath6 .
the following proposition explains the relation between the discrete hill s operator and cross ratio coordinate @xmath106 .
[ calh ] let @xmath159 and @xmath160 be two linear independent solutions of [ equhill ] and @xmath161= [ f_i:1]$ ] .
for any @xmath165 , we have @xmath105 = b_k \cdot b_{k+1}.\ ] ] since @xmath159 and @xmath160 are linear independent , we have @xmath166 for any @xmath167 , we have @xmath168 @xmath169 @xmath170 thus we have @xmath171 for any @xmath172 . by the similar argument , for @xmath173 , we have @xmath174 . we conclude that @xmath174 for any @xmath65 .
the cross ratio @xmath175 can be understood as schwarzian derivative @xcite . by comparing with proposition 2.3 of @xcite page 67
, we define a discrte version of virasoro algebra .
let @xmath176 and @xmath177 , the on @xmath178 is defined to be : for @xmath179, ... ,\left[\frac{n}{2}\right]$ ] , 1 . when @xmath180 , we have @xmath181 with the convention @xmath182 ; 2 .
when @xmath183 , we have @xmath184 notice that @xmath185-virasoro bracket is asymptotic to the poisson bracket associated to the 2-cocycle with @xmath186 , @xmath187 as in proposition 2.3 of @xcite page 67 when @xmath2 converges to infinite , but it is not a poisson bracket .
very specifical values of @xmath188 and @xmath189 correspond to virasoro algebra .
when @xmath188 is fixed , @xmath189 varies , they correspond to same element in the cohomology group @xmath190 .
different @xmath188 corresponds to different element in the one dimensional space @xmath190 .
let @xmath106 be the cross ratio coordinates of @xmath6 .
let @xmath191 . the _ discrete fourier transformation _
@xmath192 of @xmath193 is defined to be @xmath194 our main proposition of this subsection is [ propswapvir ] let @xmath156 . for @xmath195, ... ,\left[\frac{n}{2}\right]$ ] ,
let @xmath196 we have @xmath197 for @xmath179, ... ,\left[\frac{n}{2}\right]$ ] , we have @xmath198 by @xmath199 we have @xmath200 we have the above formula equals to @xmath201 \\ & = \sum_{k=1}^n e^{\frac{-2 ( p+q ) k \pi i}{n } } \cdot \left[\frac{32 h_k ( p - q)\pi i}{n^3 } - \frac{16 \pi^2 ( p^2-q^2)}{n^2 } + \frac { 32\pi^3 ( p^3-q^3 ) i } { n^3 } + o\left(\frac{1}{n^3}\right ) \right ] .
\end{aligned}\ ] ] when @xmath180 , since @xmath202 , by @xmath203 , the above formula equals to @xmath204 when @xmath183 , the above formula equals to @xmath205 replacing @xmath206 by @xmath207 we obtain that : for @xmath180 , @xmath208 for @xmath183 @xmath209 we conclude that @xmath210 the discrete version of the miura transformation @xmath211
@xcite @xcite : @xmath212 is a poisson map with respect to the poisson bracket : @xmath213 corresponding to the first adler - gelfand - dickey poissson bracket .
let us consider the miura transformation as a map @xmath211 from @xmath214 to @xmath215 .
the map @xmath211 induce another poisson bracket for @xmath6 by : @xmath216 thus we have @xmath217 we have similar result as proposition [ propswapvir ] for @xmath218 .
[ propschvir ] let @xmath156 . for @xmath195,
... ,\left[\frac{n}{2}\right]$ ] , let @xmath219 we have @xmath220 for @xmath179, ... ,\left[\frac{n}{2}\right]$ ] , we have @xmath221 by @xmath222 we have @xmath200 we have the above formula equals to @xmath223 .
\end{aligned}\ ] ] when @xmath180 , since @xmath202 , by @xmath224 , the above formula equals to @xmath225 when @xmath183 , since @xmath226 the above formula equals to @xmath227 thus we have : 1 . for @xmath180 @xmath228 2 . for @xmath183 @xmath229
we conclude that @xmath230
with the evidences shown in proposition [ propswapvir ] [ propschvir ] , we prove that @xmath231 and @xmath232 are compatible in this section .
firstly , let us recall the traditional definition of the bihamiltonian system .
two poisson brackets @xmath233 and @xmath234 for a manifold @xmath235 are said to be compatible if and only if for any @xmath236 , @xmath237 is poisson . +
a dynamic system @xmath238 over @xmath235 is bihamiltonian if its vector field @xmath239 is hamiltonian with respect to these two poisson brackets @xmath233 and @xmath234 .
let @xmath247 by definition , we have to check that @xmath248 for any @xmath249 .
since @xmath250 when some indexes coincide , for example @xmath251 , we have @xmath252 let @xmath253 be the permutation of the @xmath2 indexes such that @xmath254 . the permutation @xmath253 induce a ring automorphism @xmath255 of @xmath215 such that @xmath256 for @xmath257 .
moreover , we have @xmath258 and @xmath259 let @xmath260 be the permutation of the @xmath2 indexes such that @xmath261 for @xmath257 .
the permutation @xmath260 induce a ring automorphism @xmath262 of @xmath215 such that @xmath263 moreover , we have @xmath264 @xmath265 by the above symmetry , we suppose that @xmath266 and @xmath267 let @xmath268 we suppose that @xmath269 , we have to verify the following cases : 1 .
when @xmath270 , we have @xmath271 since @xmath272 is a polynomial of @xmath273 , we have @xmath274 similarly , we have @xmath275 and @xmath276 we conclude that @xmath277 2 . when @xmath278 , @xmath279 , we have @xmath280 3 . when @xmath281 , @xmath282 and @xmath283 , we have @xmath284 4 . when @xmath281 , @xmath282 and @xmath285 , we have @xmath286 the author is very grateful to franois labourie for suggesting the subject and for the guidance .
he thanks vladimir fock for very interesting conversations and useful comments on this work .
he is grateful to university of paris - sud , max planck institute for mathematics and yau mathematical sciences center for their hospitality . | we induce a poisson algebra @xmath0 on the configuration space @xmath1 of @xmath2 twisted polygons in @xmath3 from the swapping algebra @xcite , which is found coincide with faddeev - takhtajan - volkov algebra for @xmath4 .
there is another poisson algebra @xmath5 on @xmath6 induced from the first adler - gelfand - dickey poissson algebra by miura transformation . by observing that these two poisson algebras are asymptotically related to the dual to the virasoro algebra ,
finally , we prove that @xmath7 and @xmath5 are schouten commute . |
the cylindrical bessel can be considered generalizations of the trigonometric functions , while the associated modified forms are an extension of the relevant hyperbolic counterparts @xcite .
+ such an academic identification is non - particularly deep and might be useful for pedagogical reasons or as a guiding element to study their properties as e.g. those relevant to the asymptotic forms . we must however underline that bessel and trigonometric / hyperbolic functions share some resemblances only , but they do not display any full correspondence . +
the search for functions which are `` true '' generalizations of the trigonometric ( t- ) or hyperbolic ( h- ) forms is however recurrent in the mathematical literature .
the attempts in this direction can be ascribed to different strategies , roughly speaking the geometrical @xcite and the analytical @xcite point of views .
+ the first is based on definitions extending to higher powers the pythagorean identity of ordinary trigonometric functions , such a program identifies new trigonometries , with their own geometrical interpretation on elliptic curves and with different numbers playing the role of @xmath0 @xcite . + the second invokes the analogy with series expansions , differential equations and the theory of special functions .
+ the generalized @xmath1 functions , defined within these two contexts , are different ; in particular those belonging to the geometric strategy can be recognized as elliptic functions , including jacobi and weierstrass forms . in this article
we develop a systematic procedure within the framework of the analytical point of view .
+ we look for `` true '' generalizations , in the sense that the functions we define allow a one to one mapping onto the properties of the elementary @xmath1 functions , like addition or duplication theorems . to this aim
we exploit the insightful point of view offered by the recent understanding of bessel functions as umbral manifestation of gauss or of exponential functions @xcite .
these conceptual tools , as well as the ideas developed by cholewinsky and reneke in ref .
@xcite , provide the elements underlying the formalism of this paper , aimed at exploring in depth the identification of trigonometric functions associated with bessel functions , by getting the proper algebraic environment to establish the relevant properties . + our starting point is the equation @xmath2 where @xmath3 is the laguerre @xmath4 , introduced in the past to deal with the monomiality properties of laguerre polynomials @xcite .
+ it is easily checked that the solution of eq .
, when @xmath5 , can be cast in the form @xmath6 where @xmath7 is an example of hybrid polynomial , introduced in ref .
+ for reasons which will be clear in the following , we introduce the notation , borrowed from ref .
@xcite , @xmath8 to define a composition rule , which will be defined laguerre binomial sum .
+ it is evident that such a notion is an extension of the newton binomial , which can be generated by the action of the shifting exponential operator on an ordinary monomial , namely @xmath9 an analogous rule for the generation of lbs can be achieved by replacing the exponential function with the @xmath10 @xmath11 and the ordinary derivative with the @xmath12-derivative , satisfying the identity @xmath13 accordingly we find @xmath14^{2 } } x^{n - r } = ( x\oplus _ { l } y)^{n } \end{split}\ ] ] the function @xmath15 is a @xmath16-order bessel tricomi function @xcite - order modified bessel through the identity @xmath17 . ] and satisfies the `` laguerre''-eigenvalue equation @xmath18=\lambda \left[{}_{l } e(\lambda \ , x)\right]\ ] ] which corroborates the interpretation of its role as that of an @xmath12-exponential .
+ according to the previous identities we can also state that @xmath19 allowing the derivation of the following `` semi - group '' property of the @xmath12-exponential @xmath20 in full analogy with the ordinary euler formula we introduce the l - trigonometric @xmath21 functions through the identity @xmath22 where @xmath21 _ cosine _ and _ sine _ functions are specified by the series functions are defined by the corresponding series expansion @xmath23^{2 } } , } \\ { { } _ { l } sh(x)=\sum _ { r=0}^{\infty } \frac{x^{2r+1 } } { \left[(2\ , r+1)!\right]^{2 } } } \end{array}$ ] . ]
@xmath24 it is easily checked that they satisfy the identities @xmath25=-\alpha \ , { } _ { l } s(\alpha \ , x ) , \\ & { } _ { l } \partial _ { x } \left[{}_{l } s(\alpha \ , x)\right]=\alpha \ , { } _ { l } c(\alpha \ , x ) \end{split}\ ] ] and therefore the `` harmonic '' equation @xmath26=-\alpha ^{2 } { } _ { l } c(\alpha \ , x ) , \\ & \left({}_{l } \partial _ { x } \right)^{2 } \left[{}_{l } s(\alpha \ , x)\right]=-\alpha ^{2 } { } _ { l } s(\alpha \ , x ) \end{split}\ ] ] it is worth noting that laguerre and ordinary trigonometric functions are linked by the borel like transforms @xcite @xmath27 the use of the dilatation operator identity @xcite @xmath28 allows to cast transforms of the type in the operatorial form @xmath29 thus yielding the following identifications
@xmath30 we note that @xmath31 the use of the properties of the @xmath12-exponential function allows the derivation of the following addition theorems for the functions in eqs .
@xmath32 the proof of the last identity is given below , by noting that @xmath33\ , \left[{}_{l } c(y)+i\ , { } _ { l } s(y)\right]= \\ & = \left[{}_{l } c(x){}_{l } c(y)-{}_{l } s(x)\ , { } _ { l } s(y)\right]+i\ , \left[{}_{l } c(x){}_{l } s(y)+{}_{l } s(x)\ , { } _ { l } c(y)\right ] \end{split}\ ] ] and since @xmath34 we can equate real and imaginary parts to infer the identities .
+ it is evident that , according to the procedure we have proposed , the properties of ordinary trigonometric functions are extended to their @xmath12-counterparts , provided that the ordinary sum is replaced by the composition rule specified in eq . .
+ the formalism allows the derivation of the corresponding duplication formulae , which can be stated by defining the following product rule @xmath35 which along with eq .
yields @xmath36 furthermore the sum can be iterated as @xmath37 and , accordingly , we can state the following extension of the de moivre formula @xmath38^{n } = { } _ { l } c\left(n\otimes_{l } x\right)+i\ , { } _ { l } s\left(n\otimes_{l } x\right)\ ] ] the lissajous plot of the @xmath39 functions is reported below and will be further commented in the concluding remarks .
+ the composition rule is one of the pivotal points of this paper and will be further exploited in the following .
it is therefore worth stressing some of its properties reported below @xmath40 \dfrac{i^{\;n}\;n ! } { \left ( \left(\frac{n}{2 } \right)!\right ) ^{2 } } , & n=2k,\;\;\ ; k\in \varmathbb{n } \end{array}\right . \\ & d ) \;\ ; ( i\oplus _ { l } ( -i))^{n}=\dfrac{(-1)^{n}\;n ! } { \left ( \left(\frac{n}{2 } \right)!\right ) ^{2 } } \dfrac{\left ( 1+(-1)^{n}\right)}{2}= \left\lbrace \begin{array}{ll } = 0 , & n = 2k+1,\ ; k\in \varmathbb{n}\\[1.1ex ] = \dfrac{(-1)^{n}\;n ! } { \left ( \left(\frac{n}{2 } \right)!\right ) ^{2 } } , & n=2k,\;\;\ ; k\in \varmathbb{n } \end{array}\right .
\end{split}\ ] ] it is furthermore worth noting that @xmath41 and that the quantity @xmath42 , a kind of napier - laguerre number , is presumably trascendent .
+ the previous identities , albeit trivial , are important to appreciate the structural differences with respect to their ordinary counterparts , for example eq .
implies that @xmath43 and therefore that @xmath44^{2 } + \left[{}_{l } s(x)\right]^{2 } \ne 1\ ] ] in the forthcoming sections we will go deeper into the theory of these families of functions and we will be able to better appreciate the similitudes and the differences with the ordinary forms .
functions , @xmath45 vs @xmath46 . ]
the bessel functions are characterized by a continuous variable and by a real or complex index . a fairly natural extension of the function defined by eqs .
is therefore provided by the @xmath21 functions , associated with the @xmath47-order tricomi- bessel function @xcite @xmath48 which is an eigenfunction of the operator @xmath49 we can now proceed as in the previous section , by noting that the polynomials @xmath50 are solutions of the equation @xmath51 we further define the composition rule @xmath52 and prove that @xmath53 finally , by a slight extension of the discussion of the introductory section , we define the @xmath21 functions of @xmath47-order as @xmath54\gamma ( \alpha + 2r+1 ) } , \\ & { } _ { ( \alpha , \ , l ) } s(x)=\sum _ { r=0}^{\infty } \dfrac{(-1)^{r } x^{2r+1 } } { \left[(2\ , r+1)!\right]\gamma ( \alpha + 2r+2 ) } \end{split}\ ] ] which are shown to satisfy the addition theorems for the composition rule .
+ the procedure can be further extended by the use of humbert bessel like functions , which are defined by the series @xcite @xmath55 they satisfy the differential equation @xmath56 { } _ { l } e_{\alpha , \beta } ( \eta ) = { } _ { l } e_{\alpha , \;\beta } ( \eta ) \ ] ] and are therefore eigenfunctions of the operator @xmath57 whose @xmath58-order derivative is @xmath59 by following the same procedure as before , we introduce the composition rule @xmath60 so that the associated @xmath21 functions , defined as , @xmath61 are straightforwardly shown to satisfy the differential equations @xmath62=-\lambda \left[{}_{\left(\alpha , \;\beta , \ , l\right ) } s(\lambda \ , x)\right]\ , , \\ & { } _ { \left(\alpha , \;\beta , \ ,
l\right ) } \partial _ { \eta } \left[{}_{\left(\alpha , \;\beta , \ , l\right ) } s(\lambda \ , x)\right]=\lambda \left[{}_{\left(\alpha , \;\beta , \ , l\right ) } c(\lambda \ , x)\right ] \end{split}\ ] ] and the addition theorems based on the extension of the definition of sum specified in eq . and .
+ we have shown that the concept of @xmath39 function is a fairly natural consequence of the notion of laguerre derivative , of its extensions and of the associated eigenfunctions , which belong to bessel like forms . in the following
we will show how to frame the cholewinsky - reneke @xmath63 functions within the present framework . before entering this specific aspect of the problem
we introduce some consequences of the previous formalism on the theory of diffusion equation associated to the laguerre derivative and to its generalization .
this short section , in which we will discuss some evolutive equations based on the operators introduced in the previous sections , is an apparent detour from the main stream of the paper .
+ laguerre type diffusive equations like @xcite @xmath64 can be formally solved as @xmath65 to make the above solution meaningful it is necessary to specify how to calculate the action of the exponential operator containing the laguerre derivative on the function @xmath66 .
we will discuss therefore , as introductory example , the case in which @xmath67 , and proceed as it follows : 1 .
we note that the exponential can be written as an integral transform of the tricomi function + @xmath68 2 .
we use the properties to end up with @xmath69 more in general , whenever @xmath70 the solution of the problem can be cast in the form @xmath71 and @xmath72 is the @xmath12-transform of the function @xmath66 . +
before going further with the above formalism , we note that the equation @xmath73 can be solved in an analogous way provided that we replace @xmath74 with @xmath75 in eq . .
+ in the case in which we consider equations of the type @xmath76 the solution of the problem can be obtained as @xmath77 further comments on the previous statements will be provided in the forthcoming parts of the paper .
as already stressed the study of generalized forms of trigonometric and of hyperbolic functions is an old `` leit motiv '' in the mathematical literature . on the eve of the seventies of the last century ricci introduced @xcite a family of pseudo hyperbolic functions ( phf ) , which will be proven of noticeable importance for the topics we are discussing .
+ according to ref .
@xcite the phf of order @xmath78 are defined by the series @xmath85 } \hat{t}(y)\ , x^{3n } = \frac{1}{3 } \left[e^{\hat{\omega } _ { 0 } \
, y\ , \partial _ { x } } + e^{\hat{\omega } _ { 1 } \ , y\ , \partial _ { x } } + e^{\hat{\omega } _ { 2 } \
, y\ , \partial _ { x } } \right]\ , x^{3n } = \\ & = \frac{1}{3 } \sum _ { \alpha = 0}^{2}(x+\hat{\omega } _ { \alpha } y)^{3n } = ( x\oplus_{\left[0,3\right ] } y)^{3n } \end{split}\ ] ] by direct use of the series expansion definition of the function @xmath86 } e\left(x\right)$ ] , we end up with } 1)^{3n } = \dfrac{1}{3 } \left(2^{3n}+\left(1+e^{\frac{2i\pi}{3 } } \right)^{3n}+\left(1+e^{\frac{4i\pi}{3 } } \right)^{3n } \right ) = \dfrac{1}{3 } \left(2^{3\ , n } + ( -1)^{n } 2\right)$ ] ] @xmath87 } e\left(y\ , \partial _ { x } \right)x^{3\ , n } = \sum _ { r=0}^{\infty } \frac{y^{3r } } { ( 3r ) ! } \ , \partial _ { x}^{3r } x^{3n } = \\ & = \sum _ { r=0}^{n}\binom{3n}{3r } \ ; y^{3r } x^{3\ , \left(n - r\right ) } = ( x\oplus_{\left[0,3\right ] } y)^{3n } \end{split}\ ] ] it is therefore evident that the following further identities can be stated @xmath88 } e\left(y\ , \partial _ { x } \right)\ , { } _ { [ 0,3 ] } e(x)={}_{[0,3 ] } e(x\oplus_{0 } y ) , \\ & { } _ { [ 0,3 ] } e\left(y\ , \partial _ { x } \right)\ , { } _ { [ 0,3 ] } e(x)={}_{[0,3 ] } e(y)\ , { } _ { [ 0,3 ] } e(x ) \end{split}\ ] ] which , once merged , yield @xmath89 } e\left(x\right)\ , { } _ { [ 0,3 ] } e(y)={}_{[0,3 ] } e(x\oplus_{[0,3 ] } y)\ ] ] in this way we have obtained a result allowing the introduction of @xmath1 like functions according to the paradigm developed so far .
+ by a straightforward generalization of the discussion developed in the introductory remarks we introduce the generalized @xmath90-functions defined as @xmath91}ch\left(x\right)=\frac{1}{2 } \left ( { } _ { [ 0,3]}e(x)+ { } _ { [ 0,3]}e\left(-x\right)\right ) = \sum _ { r=0}^{\infty } \dfrac{x^{\;6\ , r } } { ( 6\ , r ) ! } \\ & { } _ { [ 0,3]}sh\left(x\right)=\frac{1}{2 } \left ( { } _ { [ 0,3 ] } e(x)- { } _ { [ 0,3]}e\left(-x\right)\right ) = \sum _ { r=0}^{\infty } \dfrac{x^{\;6\ , r+3 } } { ( 6\ , r+3 ) ! } \end{split}\ ] ] @xmath92}ch\left ( \alpha \oplus_{[0,3 ] } \beta \right ) = { } _ { [ 0,3]}ch(\alpha ) { } _ { [ 0,3]}ch(\beta ) + { } _ { [ 0,3 ] } sh(\alpha ) { } _ { [ 0,3]}sh(\beta ) , \\ & { } _ { [ 0,3 ] } sh\left ( \alpha \oplus_{[0,3 ] } \beta \right ) = { } _ { [ 0,3]}ch(\alpha ) { } _ { [ 0,3]}sh(\beta ) + { } _ { [ 0,3]}sh(\alpha ) { } _ { [ 0,3]}ch(\beta ) \end{split}\ ] ] furthermore , as discussed later in the paper , analogous conclusions can be reached for @xmath93 } e(x),\,\ ; k=0,\dots , m-1 $ ] .
+ to obtain the link with the topics discussed so far , we consider a particular case of the function defined in eq . , namely studied in ref .
the function in eq . is equivalent , apart from an unessential normalizing factor , to the function exploited by cholewinsky and reneke @xcite to study the solution of eq . which will be linked , in particular , to the following expression forming the generalized lbs for the addition theorem of the relevant @xmath104 functions .
+ a natural extension of the previously developed formalism allows the introduction of the l / h - functions ; their definitions and properties , apart from some computational complications , does not produce any significant conceptual progress .
+ in this paper we have seen how a `` wise '' combination of different methods borrowed from special function theory , operational and umbral calculus and integral transforms , opens new interesting possibilities for the introduction and systematic study of new families of trigonometric functions .
+ in addition , the method yields , as byproduct , the opportunity of getting natural solutions of a large family of pde belonging to the family of generalized heat equation , whose links with the radial heat equation have been touched on in ref . @xcite and will be more carefully discussed elsewhere . +
before closing the paper we like to stress two points just touched on in the previous part of the paper .
+ we have noted in eq . that the laguerre exponential can be obtained through a limit procedure analogous to that involving the napier number .
the same method can e.g. be used to state that which is recognized as the asymptotic limit of laguerre polynomials @xcite .
+ the second point we want to emphasize is the geometrical interpretation of the @xmath39 functions from the geometrical point of view .
+ such an interpretation is provided in fig .
[ fig1 ] , where we have considered the lissajous curves plotting @xmath12-sin vs @xmath12-cos .
[ fig2 ] it is evident that the curves are open since no periodic behaviour is envisaged .
however a kind of self - similarity can be noted when the amplitude of the oscillations increase with increasing @xmath106 .
+ the last figure provides the explicit correspondence of the @xmath12-sinus and @xmath12-cosinus along with the relevant `` @xmath12-angle '' , intended as the area intercepted by the @xmath12-curve and the segment forming the angle in the positive abscissa direction .
this paper has provided the formalism for a wider understanding of concepts associated with the laguerre derivative , the underlying algebraic structures and possible extensions to new forms of trigonometry . in a forthcoming investigation
we will provide an account of application of the method to problems in physics and applied mathematics .
+ * references * g. dattoli , e. di palma , e. sabia , k. grska , a. horzela , k. a. penson , _ `` operational versus umbral methods and the borel transform '' _
, international journal of applied and computational mathematics , doi 10.1007/s40819 - 017 - 0315 - 7 , feb 2017 .
cocolicchio , g. dattoli , h.m .
srivastava , `` proceedings of the workshop advanced special functions and applications '' , melfi school on advanced topics in mathematics and physics , melfi ( pz ) , italy , 9 - 12 may 1959 , aracne . | we develop a new point of view to introduce families of functions , which can be identified as generalization of the ordinary trigonometric or hyperbolic functions .
they are defined using a procedure based on umbral methods , inspired to the bessel calculus of bochner , cholewinsky and haimo .
we propose further extensions of the method and of the relevant concepts as well and obtain new families of integral transform allowing the framing of the previous concepts within the context of generalized borel tranforms .
trigonometric and hyperbolic functions , laguerre polynomials , airy forms , umbral calculus , integral transforms , operational methods . |
the dust attenuation curve is a principal tool for inferring the absorption and scattering properties of interstellar dust grains , the spatial distribution of that dust relative to stars , and the intrinsic stellar populations of galaxies .
observations of nearby galaxies have yielded detailed information on both the normalization and shape of the dust curve ( e.g. , @xcite ) .
in contrast , our knowledge of the dust curve ( and how the curve varies ) at high redshift has only recently improved with large multi - wavelength and spectroscopic surveys ( e.g. , @xcite ) . despite these advances , the shape of the attenuation curve at far - ultraviolet ( far - uv ) wavelengths , in particular at @xmath12 ,
is unconstrained at high redshift .
establishing the shape of this curve at these wavelengths is imperative for several reasons .
first , dust obscuration of starlight reaches its maximum in the far - uv , such that even small differences in the continuum color excess , @xmath4 , may lead to large variations in the inferred attenuation of far - uv light .
hence , inferences of the intrinsic formation rate of the very massive stars that dominate the far - uv requires knowledge of the shape of the dust curve at these wavelengths .
second , the translation between reddening and far - uv color depends on the shape of the far - uv attenuation curve .
therefore , different shapes of the far - uv curve will result in different color criteria to select galaxies at intermediate redshifts ( @xmath13 ) where the ly@xmath14 forest is sufficiently thin and where the selection may include photometric bands that lie blueward of ly@xmath14 .
stated another way , fixed far - uv color selection criteria will select galaxies at slightly different redshifts depending on the shape of the uv dust attenuation curve , and may affect the inferred selection volumes that are necessary for computing uv luminosity functions ( e.g. , @xcite ) .
third , the utility of the ly@xmath14 emission line as a probe of the physical conditions ( e.g. , dust / gas distribution , gas kinematics ) in high - redshift galaxies requires some knowledge of dust grain properties and the distribution of the dust in the ism ( e.g. , @xcite ) , both of which are reflected in the shape of the attenuation curve .
fourth , there has been renewed attention in the modeling of massive stars @xcite given evidence of rest - optical line ratios @xcite and uv spectral features of @xmath15 galaxies ( steidel et al . , in prep ) that can not be modeled simultaneously with standard stellar population synthesis models .
the modeling of the far - uv continuum of high - redshift galaxies is sensitive to the assumed dust attenuation curve .
lastly , knowledge of the attenuation curve at or near the lyman break is needed to deduce the effect of dust obscuration on the depletion of hydrogen - ionizing photons , a potentially important ingredient in assessing the contribution of galaxies to reionization and translating recombination line luminosities to star - formation rates . given the relevance of the far - uv shape of the dust attenuation curve , we sought to use the existing large spectroscopic samples of lyman break galaxies at @xmath1 to provide the first direct constraints on the shape of this curve at high redshift . in section
[ sec : sample ] , we discuss the sample of galaxies and the procedure to construct far - uv spectral composites .
the calculation of the far - uv dust attenuation curve is presented in section [ sec : faruvcurve ] . in section [ sec :
specfitting ] , we assess whether the dust curve can be used in conjuction with spectral models to reproduce the composite uv spectra .
the implications of our results are discussed in section [ sec : discussion ] , and we summarize in section [ sec : summary ] .
all wavelengths are in vacuum .
all magnitudes are expressed in the ab system .
we adopt a cosmology with @xmath16kms@xmath17mpc@xmath17 , @xmath18 , and @xmath19 .
the galaxies used in our analysis were drawn from the lyman break selected and spectroscopically confirmed sample described in @xcite . in brief , these galaxies were selected as @xmath20 candidates based on their rest - frame uv ( @xmath21 and @xmath22 ) colors as measured from ground - based imaging of 17 fields . of the 2,347 @xmath23 lyman break galaxy ( lbg )
candidates , 1,293 were observed spectroscopically with the original single channel configuration of the low resolution imaging spectrograph ( lris ; @xcite ) on the keck telescopes , with a typical integration time of @xmath24hr .
the spectroscopy was accomplished using the 300linesmm@xmath17 grating blazed at 5000 giving a typical observed wavelength coverage of @xmath25 with 2.47pix@xmath17 dispersion .
the spectral resolution , including the effects of seeing , is @xmath26 in the observed frame .
redshifts were determined from either interstellar absorption lines at @xmath27 , or from ly@xmath14 emission . in some cases
, the detection of a single emission line with a continuum break shortward of that line was used to assign a secure redshift .
our sample includes only those galaxies with secure emission and/or absorption line redshifts that were vetted by at least two people and published in @xcite .
further details of the photometry , candidate selection , followup spectroscopy , and data reduction are provided in @xcite .
subsequent deep ( @xmath28hrs ) spectroscopy was obtained for a sample of 121 @xmath29 galaxies with existing spectroscopic redshifts from the @xcite samples , with the aim of obtaining very sensitive observations of the lyman continuum ( lyc ) emission blueward of the lyman limit ( @xmath30)we refer to the set of objects with deeper spectroscopy as the `` lyc sample '' .
this spectroscopy was obtained using the blue channel of the upgraded double spectrograph mode of lris @xcite , which presents a factor of @xmath31 improvement in throughput at @xmath32 relative to the red - side channel .
the new configuration of lris includes an atmospheric dispersion corrector ( adc ) that eliminates wavelength dependent slit losses .
the deep spectroscopic observations were obtained using the d500 dichroic with the 400linesmm@xmath17 grism blazed at 3400 , yielding a ( binned ) dispersion of 2.18pix@xmath17 . including the effect of seeing , the typical observed - frame resolution is @xmath33 for the blue channel spectra .
redshifts were determined using the same procedure as for the @xcite sample ( see above ) .
the depth of the spectroscopic observations for the lyc sample allowed us to assign absorption line redshifts for @xmath34 of the objects .
the 121 galaxies of the lyc sample excludes all objects ( @xmath35 of the sample ) where the deep spectra indicated a blend between the target lbg and a foreground galaxy .
details of the data reduction and spectroscopy are provided in steidel et al .
( in prep ) .
we further excluded agn from the samples based on the presence of strong uv emission lines ( e.g. , ly@xmath14 and ) .
the agn fraction in the @xmath1 sample ( to the magnitude limit of @xmath36 ) based on strong uv emission lines is @xmath37 ( e.g. , @xcite ) .
we did not have the requisite multi - wavelength ( infrared , x - ray ) data for the majority of the fields to evaluate agn using other methods .
however , we note that optically - identified agn in uv - selected samples comprise @xmath38 of agn identified either from x - ray emission , or from a mid - infrared excess ( see @xcite ) , and we do not see evidence of agn either in the remaining individual spectra or in the composite spectra
. two important advantages of the @xmath1 sample that facilitate spectroscopy in the far - uv region is the lower sky background in the observed optical relative to the near - infrared , and the fact that the ly@xmath14 forest is still relatively thin at these redshifts , allowing for the transmission of sufficient flux blueward of ly@xmath14 .
figure [ fig : zhist ] shows the spectroscopic redshift distribution of our final sample , which includes 933 galaxies . throughout our analysis , we employed composite spectra which were computed by combining the individual spectra of galaxies .
we first made separate composites of the galaxies in the @xmath1 lbg and lyc samples , and then combined the two resulting composites to obtain a final composite .
this procedure allows us to take advantage of the greater spectroscopic depth of the lyc objects in the far - uv range @xmath39 . for each galaxy in the @xmath1 lbg sample
, we determined the redshift using either emission ( e.g. , ly@xmath14 ) and/or interstellar absorption lines .
as these features are typically either redshifted ( in the case of emission ) or blueshifted ( in the case of absorption ) due to bulk flows of gas , we computed a systemic redshift from the emission and/or absorption line redshifts using the relations in @xcite .
we then shifted each galaxy s spectrum into the rest frame using the systemic redshift , and interpolated the spectrum to a grid with a wavelength spacing of @xmath40 .
the error spectrum for each galaxy was shifted and interpolated accordingly , and then multiplied by @xmath41 to account for wavelength resampling from the original dispersion of the reduced spectra of 1pix@xmath17 .
each spectrum was scaled according to the median flux density in the range @xmath42 .
the individual galaxy spectra were then averaged , rejecting the upper and lower three flux density values in each @xmath43 interval .
the lyc composite was constructed using a similar procedure , with two exceptions .
first , we did not scale the lyc spectra as the spectroscopic setup ( with the d500 dichroic ) precluded coverage of @xmath42 within the blue channel of lris .
second , we combined the lyc spectra using a weighted average , where the weights are @xmath44 , where @xmath45 is the error spectrum .
we still rejected the upper and lower three weighted flux density values in each @xmath43 interval .
the lyc composite was scaled to have the same median flux density in the range @xmath46 as that of the @xmath1 lbg composite .
the final composite results from combining the scaled lyc composite at @xmath47 and the @xmath1 lbg composite at @xmath48 .
we confirmed that this combination does not introduce a systematic bias in the level of the stellar continuum of the composite below and above @xmath49 .
the mean redshifts of the objects contributing to the composite spectrum at the blue ( @xmath50 ) and red ( @xmath51 ) ends of the composite spectrum are @xmath52 and @xmath53 , respectively .
given this small difference in redshift , corresponding to @xmath54myr , we assumed that the lbg spectral properties do not vary over this short time interval .
the final composite is shown in figure [ fig : compall ] , normalized such that the median flux density in the range @xmath42 is unity .
the error spectrum corresponding to the composite was computed as the weighted combination of the error spectra of the individual objects contributing to the composite .
the error in the mean flux of the composite is @xmath55@xmath56jy for the scaling and full wavelength range shown in figure [ fig : compall ] .
the method of computing the far - uv attenuation curve , @xmath57 , is based on taking the ratios of composite spectra in bins of @xmath4 .
the attenuation curve is defined as @xmath58 , \label{eq : attcurve}\ ] ] where @xmath59 and @xmath60 refer to the average flux density and average @xmath4 , respectively , for objects in the _
i_th bin of @xmath4 .
our procedure involves computing the @xmath4 for each galaxy based on its rest - uv ( @xmath61 ) color , and an assumed intrinsic template and attenuation curve .
the template is a combination of the @xcite models and the 2010 version of starburst99 spectral synthesis models in the far - uv @xcite .
the models were computed assuming a stellar metallicity of @xmath62 , based on current solar abundances @xcite , with a constant star - formation history and age of 100myr .
this particular combination of parameters was found to best reproduce the uv stellar wind lines blueward of ly@xmath14 ( e.g. , @xmath631176 ) , as well as the @xmath631240 and @xmath631549 p - cygni profiles .
the star - formation history and age of the model ( @xmath64myr ) were chosen to be consistent with the results from stellar population modeling of @xmath1 lbgs presented in @xcite and @xcite .
on average , such galaxies exhibit uv - to - near - ir seds and bolometric sfrs that are well - described by either a rising or constant star - formation history @xcite .
for these star - formation histories , the uv spectral shape remains constant for ages @xmath65myr as the relative contribution to the uv light from o and b stars stabilizes after this point .
moreoever , there are no significant correlations between reddening and either star - formation history or age for galaxies in our sample @xcite .
thus , adopting a different model star - formation history and age will serve to systematically shift the computed @xmath4 , but the value of the attenuation curve , which depends on the _ difference _ in the average @xmath4 from bin - to - bin ( equation [ eq : attcurve ] ) , will be unaffected .
nebular continuum emission was added to the stellar template to produce a final model that we subsquently call the `` intrinsic template . ''
that are systematically smaller ( because the spectrum of the template is redder ) , typically by @xmath66 , than a model that includes only the stellar light . ]
we generated a grid of @xmath61 colors by reddening the intrinsic template assuming @xmath67 , with increments of @xmath68 , and two different attenuation curves : the @xcite and @xcite curves .
the former curve is the most commonly used attenuation curve for high - redshift galaxies .
the latter curve is our most recent determination using a sample of @xmath69 star - forming galaxies at redshifts @xmath70 from the mosfire deep evolution field ( mosdef ; @xcite ) survey .
while lower in redshift than most of the galaxies examined here , the mosdef sample covers roughly the same range in stellar mass and star - formation rate as the latter .
the reddened templates were redshifted and attenuated by the average igm opacity appropriate for that redshift .
the average opacity was determined by randomly drawing line - of - sight absorber column densities from the distribution measured in the keck baryonic structure survey ( kbss ; @xcite ) , and then taking an average over many lines - of - sight .
as the attenuation curve depends only on the ratio of the composite spectra and the difference in their @xmath4 , adopting a different prescription for the igm opacity will not affect the shape of the attenuation curve .
the models were then multiplied by the _
g_- and @xmath71-band transmission filters to produce a model - generated @xmath61 color associated with each @xmath4 .
the relationship between @xmath61 and @xmath4 , and the distribution of @xmath4 for the 933 galaxies in our sample , are shown in figure [ fig : colordist ] .
we developed an iterative derivation of the attenuation curve which has 7 main steps : ( 1 ) compute the @xmath4 of each object in the sample ; ( 2 ) bin the objects by @xmath4 and construct the composite for each bin ; ( 3 ) determine the average @xmath4 that best fits the composite ; ( 4 ) renormalize the composite to be referenced to a dust - corrected spectrum with a @xmath72 flux density equal to unity ; ( 5 ) compute the @xmath57 curve from the ratio of the composite spectra in different @xmath4 bins , and fit a curve to it ; ( 6 ) compute the variation between the new @xmath57 and that of the previous iteration ; and ( 7 ) repeat steps ( 1 ) through ( 6 ) until the shape of the curve converges .
a flowchart of the procedure is depicted in figure [ fig : flowchart ] .
our calculation presents several advantages .
first , performing this analysis using the composite spectra of many galaxies allows us to average over stochastic variations in the ly@xmath14 forest so that the resulting composites can be used reliably to measure the far - uv shape of the dust attenuation curve .
second , as noted above , the average stellar populations of galaxies contributing to each bin of @xmath4 are similar . as the far - uv portion of the spectrum
is dominated by only the most massive stars , the shape of the spectrum remains constant after @xmath73myr and is insensitive to the presence of older stellar populations .
third , the stellar mass distribution of @xmath1 lbgs in our sample , which spans the range from a few @xmath74@xmath75 to @xmath76@xmath75 , implies an @xmath77dex difference in mean oxygen abundance from the lowest to highest stellar mass galaxies in our sample , based on the stellar mass - metallicity relation at @xmath1 @xcite . assuming a similar range of stellar metallicities for a fixed star - formation history and imf implies a @xmath78 difference in the intrinsic uv slope across the range of galaxies considered here .
thus , variations in the composite spectral shapes are driven primarily by changes in dust attenuation , rather than by differences in stellar metallicity .
the details of our dust attenuation curve calculation are presented below . in step 1 , the observed @xmath61 color of each galaxy was corrected for ly@xmath14 emission or absorption as measured from the rest - uv spectrum .
this line - corrected color was compared to the model - generated @xmath61 colors for each attenuation curve in order to find the appropriate @xmath4 . in step 2 , the galaxies were sorted and divided into two bins of @xmath4 , a `` low '' ( `` blue '' ) and `` high '' ( `` red '' ) @xmath4 bin , with roughly an equal number of galaxies in each bin ( 466 and 467 objects , respectively ) . the composite spectrum was computed for the objects in each bin of @xmath4 , resulting in a `` blue '' and `` red '' @xmath4 composite , following the procedure specified in section [ sec : compspec ] . for step 3 ,
the best - fit @xmath4 for each composite spectrum was determined by taking the same intrinsic template specified above and reddening it according to the @xcite and @xcite attenuation curves using the same range of @xmath4 specified above .
these models were then redshifted to the mean redshift of objects contributing to the composite , and attenuated according to the igm opacity appropriate for that redshift . the model that best fits the composite spectrum then determines the @xmath4 .
we considered the largely line - free wavelength windows listed under `` composite @xmath4 '' in table [ tab : waves ] when comparing the model spectrum to the composite .
the best - fit @xmath4 was found by @xmath79 minimization of the model with respect to the composite spectrum .
the @xmath4 computed in this way agreed well with the value inferred from the average @xmath61 color of objects contributing to the composite .
lc composite @xmath4 & 1268 - 1293 + & 1311 - 1330 + & 1337 - 1385 + & 1406 - 1521 + & 1535 - 1543 + & 1556 - 1602 + & 1617 - 1633
+ + @xmath57 & 950 - 970 + & 980 - 985 + & 994 - 1017 + & 1040 - 1182 + & 1268 - 1500 + + @xmath80 & 967 - 974 + & 1021 - 1029 + & 1200 - 1203 + & 1222 - 1248 + + @xmath81 & 941 - 944 + & 956 - 960 + & 979 - 986 + & 992 - 1020 [ tab : waves ] in step 4 , the composite was divided by the factor @xmath82 , where @xmath83 and @xmath84 are the flux density and attenuation curve , respectively , at @xmath85 .
this renormalization forces the dust - corrected spectrum ( i.e. , the spectrum obtained when de - reddening the composite according to its @xmath4 for a given attenuation curve ) to have @xmath86 .
the new shape of the attenuation curve is computed in step 5 using equation [ eq : attcurve ] .
we fit the new @xmath57 with a polynomial of the form @xmath87 , where @xmath63 is the wavelength in @xmath56 m . in fitting @xmath57 , we only considered windows in the far - uv ( @xmath88 ) that are free of @xmath5 lines , and which are listed in table [ tab : waves ] . while the windows do include some metal lines , we confirmed by excluding the wavelength regions corresponding to the weaker interstellar and stellar absorption lines that the presence of such lines did not significantly alter the fit to the attenuation curve .
in addition , we used the errors in the attenuation curve ( see below ) to determine the best - fit polynomial using @xmath79 minimization . in step 6
, we determined if the new @xmath57 computed in step 5 differed from the assumed @xmath57 ( i.e. , the one from the previous iteration ) by more than @xmath89 at any given wavelength .
if so , the new @xmath57 curve blueward of @xmath72 was spliced onto the @xmath57 from the previous iteration ( note that the curve redward of @xmath72 remains fixed throughout the iterative procedure ) .
this new version of @xmath57 was then used to repeat steps 1 through 6 .
convergence in the shape of the attenuation curve defined such that the curve varies by no more than @xmath89 at any wavelength from that of the previous iteration was reached within just two or three iterations .
the curves obtained when initially assuming the @xcite and @xcite attenuation curves are shown in figure [ fig : attcurve ] .
the best - fit polynomial to the far - uv curve with the @xcite normalization at @xmath72 , and which takes into account the random errors discussed below , is @xmath90 as shown in figure [ fig : attcurve ] , this polynomial is essentially identical to the fit presented in @xcite for @xmath91 , but deviates from the former curve at shorter wavelengths ( @xmath92 ) . similarly , the functional fit to the curve with the @xcite normalization at @xmath72 is @xmath93 defined over the same wavelength range specified in equation [ eq : kcalzfit ] .
there are several sources of random error in our calculation of the far - uv attenuation curve .
first are the measurement errors in the composite spectra ( section [ sec : compspec ] ) .
second are measurement uncertainties in the @xmath61 colors , which can scatter galaxies from one @xmath4 bin to another .
third are deviations away from the attenuation curve from one galaxy to another ( e.g. , due to differences in dust / stars geometries ) , which will affect the @xmath4 inferred for a given @xmath61 color .
fourth are random variations in the igm opacity when averaged over the 466 and 467 sightlines ( 70 and 60 sightlines for @xmath94 ) that contribute to the two @xmath4 bins .
the combined effect of these sources of random error was quantified by calculating @xmath57 many times using the same procedure described above , where each time we ( a ) varied the @xmath61 color of each galaxy according to its measurement uncertainty ; ( b ) perturbed the @xmath4 associated with the @xmath61 color for a given attenuation curve by @xmath95 based on the intrinsic scatter of galaxies used to calibrate the @xcite and @xcite relations , but ensuring that @xmath96 ; ( c ) randomly drew 466 and 467 sightlines ( and 70 and 60 sightlines ) based on the column density distribution measured from kbss ; and ( d ) perturbed the composite spectra according to the corresponding error spectra .
we then computed the dispersion in the fits of the attenuation curve obtained at each wavelength from these simulations to estimate the random error in the attenuation curve . the total random error obtained in this way
is indicated by the shaded regions around the attenuation curves shown in figure [ fig : attcurve ] .
note that while the random error increases towards bluer wavelengths at @xmath97 , this error is still a factor @xmath98 smaller than the systematic differences between our new determinations of @xmath57 and the extrapolations of the original @xcite and @xcite curves .
thus , the differences between the new far - uv curves derived here and the extrapolations of the original @xcite and @xcite curves at @xmath99 are statistically significant .
the largest source of systematic error in the far - uv curve is the attenuation curve initially assumed for the calculation ( section [ sec : compspec ] ) , as demonstrated in figure [ fig : attcurve ] .
the far - uv shape of @xmath57 is similar regardless of whether we have assumed the @xcite or @xcite curves when initially calculating the @xmath4 for galaxies in our sample , but because of the way in which we calculate @xmath57 , the normalization is fixed to that of the initially assumed attenuation curve at 1500 .
thus , the normalization of @xmath57 is unconstrained in our analysis .
a second source of systematic error arises from the fact that the gas covering fractions and/or column densities of and @xmath6 , as well as ism metal absorption , might be expected to increase with @xmath4 .
given that we have computed the attenuation curve in the far - uv where such lines become prevalent ( figure [ fig : compall ] ) , we must assess whether differences in the average absorption present in the composites for the low and high @xmath4 bins have a significant impact on our calculation of @xmath57such differences would cause us to over - estimate the value of @xmath57 .
this issue is addressed in section [ sec : specfitting ] .
other sources of systematic uncertainty include our choice of the intrinsic template and the extrapolation of the column density distributions measured in kbss to @xmath100 . both of these have a negligible effect on the derived attenuation curve as systematic changes in the intrinsic template or the extrapolation used to model the ly@xmath14 forest at @xmath100 will tend to systematically shift the @xmath4 of individual objects either up or down , but the ordering of the galaxies in @xmath4 will be preserved .
our calculation marks the first direct measurement of the far - uv attenuation curve at high redshift using spectroscopic data extending to the lyman break at @xmath50 .
as can be seen from figure [ fig : attcurve ] , the dust curve derived here rises less rapidly at @xmath101 than the polynomial extrapolations of the @xcite and @xcite attenuation curves ; the latter were empirically calibrated for @xmath102 .
the far - uv attenuation curve derived here agrees well with that of @xcite for local star - forming galaxies . for comparison ,
figure [ fig : attcurve ] also includes the @xcite average extinction curve measured over 75 individual galactic sightlines , and the extinction curve obtained from the dust grain model of @xcite . in general , the far - uv attenuation curve has a shallower dependence on wavelength than the galactic extinction curve , though the comparison is complicated by the fact that the curve derived here is an _ attenuation _ curve which represents an average reddening across many sightlines .
thus , the differences observed between the various curves in the far - uv may be tied to variations in the spatial geometry of the dust and stars , as well as to changes in the dust grain size distribution and scattering and absorption cross - sections
. the new far - uv measurements of the attenuation curves derived here are subsequently referred to as the `` updated '' versions of the @xcite and @xcite curves .
we assessed whether the uv attenuation curve can adequately describe the composite spectra of galaxies in our sample and , in the process , determined the magnitude of any systematic variations in the @xmath5 and @xmath6 column densities , and metal absorption , between the two bins of @xmath4 used to calculate the curve . to accomplish this , we reddened the intrinsic template by various amounts assuming the updated @xcite attenuation curve and a unity covering fraction of dust ( @xmath103 ) , and determined which @xmath4 results in the best fit to the composite spectrum . enforcing @xmath103
was motivated by the fact that the attenuation curve is defined assuming that the entire galaxy is reddened by some average amount .
the fitting was accomplished using the `` composite @xmath4 '' windows specified in table [ tab : waves ] .
we refer to these as the `` @xmath4 only '' fits ( i.e. , fits that do not include the lines or lyman - werner absorption from @xmath6 ) , and they are shown for the two @xmath4 bins in figure [ fig : hih2fit ] .
the uncertainty in @xmath4 is @xmath104 based on perturbing the composite spectra according to their errors and refitting them many times . while these models provide a good fit to the composite spectra at @xmath105 , they clearly overpredict the observed continuum level at bluer wavelengths , implying that additional absorption must be present at these wavelengths . at first glance , the discrepancy between the reddened templates and the composite spectra in the far - uv may appear to be at odds with the fact that the attenuation curve was derived from the composite spectra themselves . however , recall that the attenuation curve is calculated from the _ ratio _ of the composite spectra .
systematic absorption in the far - uv that is present in the individual composite spectra will largely `` divide out '' in the calculation of the attenuation curve . in the next iteration of the spectral fitting
, we accounted for absorption by neutral hydrogen within the galaxies as follows .
we added @xmath5 lyman series absorption to the intrinsic template assuming a voigt profile with a doppler parameter @xmath106kms@xmath17 and varying column densities @xmath80 .
the wings of the line profile are insensitive to the particular choice of @xmath107 .
we also blueshifted the @xmath5 absorption lines relative to the model by @xmath108kms@xmath17 to provide the best match to the observed wavelengths of the lines in the composite .
this blueshift is generally attributed to galactic - scale outflows of gas seen in absorption against the background stellar continuum .
the composite spectra were compared to the model spectra , where the latter were computed as : @xmath109 , \label{eq : hifit}\end{aligned}\ ] ] where @xmath110 is the covering fraction of neutral hydrogen ; @xmath111 and @xmath112 are the intrinsic templates with and without @xmath5 absorption , respectively ; and @xmath4 is the value found above ( from the @xmath4 only fitting ) .
note that the reddening was applied to the entire spectrum ; i.e. , @xmath103 .
these final models were computed for a range of @xmath80 and @xmath110 and modified by the igm opacity appropriate for the average redshift of each composite .
the final models were compared to the composite spectra in the `` @xmath80 '' wavelength windows specified in table [ tab : waves ] to find the best - fitting combination of @xmath80 and @xmath110 .
the wavelength windows were chosen to avoid regions of strong metal line absorption ( e.g. , @xmath63977 and @xmath631036 ) and strong ly@xmath14 emission , and include the core and as much of the wings of ly@xmath113 and ly@xmath114 , and the wings of ly@xmath14 , particularly the region between ly@xmath14 and @xmath631238 .
the higher lyman series lines ( ly@xmath115 and greater ) were not used due to the larger random errors at these wavelengths and because these lines begin to overlap .
the covering fraction is constrained primarily by the depths of the lines , whereas the column density is most sensitive to the wings .
the fitting suggests column densities of @xmath116}\approx 20.8 - 21.3 $ ] and covering fractions of @xmath117 for the two bins of @xmath4 ( figure [ fig : hih2fit ] ) . by perturbing the composite spectra and re - fitting the data many times
, we estimated the ( measurement ) errors in @xmath116}$ ] and @xmath110 to be @xmath118})\approx 0.20$]dex and @xmath119 .
the systematic error in @xmath116}$ ] from changing the doppler parameter by a factor of two from our assumed value is @xmath118})\approx 0.10$]dex .
note that the values of @xmath110 derived here are physically meaningful only if the @xmath5 lines are optically thick _ and _ the residual flux observed at the @xmath5 line cores is not an artifact of limited spectral resolution , redshift imprecision , or foreground contamination . in the first case ,
a high @xmath80 is required to reproduce the damping wings of ly@xmath14 , ly@xmath113 , and ly@xmath114 , implying optically thick gas . in the second case , as shown in detail in reddy et al .
( 2016b ) , redshift and spectral resolution uncertainties are not sufficient to account for the residual flux seen at the @xmath5 line centers .
briefly , we executed simulations where we added @xmath5 of varying column densities to the intrinsic spectrum , replicated many versions of this @xmath5-absorbed spectrum with slightly different redshifts based on the typical redshift uncertainty for galaxies in our sample , smoothed these versions with various values of the spectral resolution , and then produced a composite of all these versions .
this exercise revealed that we would have had to systematically underestimate the rest - frame spectral resolution by more than @xmath120 in order to reproduce the amount of residual flux at the @xmath5 line centers .
this @xmath120 is much larger than the rms uncertainties in determining the spectral resolution from the widths of night sky lines .
lastly , we ruled out foreground contamination as a source of the residual flux based on the exclusion of spectroscopic blends from our sample and the fact that this flux inversely correlates with reddening , a result that would not be expected if the flux was dominated by foreground contaminants ( see reddy et al .
2016b for more discussion ) .
thus , we conclude that the spectra can be modeled with a partial covering fraction of optically thick @xmath5 . in the final iteration of the fitting
, we included @xmath6 lyman - werner absorption in the models as follows : @xmath121,\end{aligned}\ ] ] where @xmath122 is the covering fraction of molecular hydrogen , @xmath123 is the model including both @xmath5 and @xmath6 absorption , and the remaining variables have the same definitions as in equation [ eq : hifit ] .
this calculation assumes that the @xmath6 is entirely embedded in @xmath5 clouds .
@xmath6 absorption was modeled using the @xcite templates , assuming a doppler parameter @xmath124kms@xmath17 , and all vibrational transitions from the first vibrational mode , and the @xmath125 transitions .
the templates were smoothed to match the resolution of the composite spectra . in fitting the models to the composite spectra , the @xmath4 , @xmath80 , and @xmath110
were fixed to the values found above , and @xmath81 and @xmath122 were allowed to vary .
the `` best - fit '' @xmath81 and @xmath126 were found by minimizing the differences in the model and composite spectra in the four wavelength windows indicated by the `` @xmath81 '' entry in table [ tab : waves ] .
these windows were chosen because they are free from @xmath5 absorption and strong metal line absorption as identified in far ultraviolet spectroscopic explorer ( fuse ) spectra of the lmc by @xcite .
the best - fitting models including @xmath6 are shown in figure [ fig : hih2fit ] , and we find @xmath127 and @xmath128}\approx 20.9 $ ] . as the @xmath6 lines are blended and unresolved , we can not accurately determine the errors in the fitted parameters .
nevertheless , we note that the deficit in flux of the composites relative to the models which _ do not _ include the lyman - werner bands in the four @xmath6 windows ( table [ tab : waves ] ) implies that @xmath6 absorption must be present , albeit with a low covering fraction or low column density ( figure [ fig : h2fit ] ) .
is expected to be weak relative to the absorption arising from transitions from the ground vibrational state for the typical conditions in the ism .
transitions from the ground vibrational state ( for the six rotational transitions considered here , @xmath129 ) completely blanket the far - uv continuum given the resolution of our spectra . thus , adding absorption corresponding to upper level vibrational transitions will simply remove `` power '' from the considered ( ground state vibrational ) mode and the column densities and covering fractions derived here would be upper limits . ]
the apparently low covering fraction of @xmath6 relative to that of the @xmath5 and dust implies that most of the dust in the ism of typical galaxies at @xmath1 is unrelated to the catalysis of @xmath6 molecules , and is associated with other phases of the ism ( e.g. , ionized and neutral gas ) .
this result is robust against the assumption of a non - unity covering fraction of dust . in other words ,
as shown in reddy et al .
( 2016b ) , assuming a _ partial _ covering fraction of dust yields dust covering fractions that are @xmath130 , still much larger than the covering fraction deduced for @xmath6 .
the aforementioned fitting demonstrates that the updated @xcite curve provides a reasonable fit to the composite spectra over the entire far - uv through near - uv range once @xmath5 and @xmath6 absorption are accounted for .
we found similar results when considering the updated @xcite attenuation curve .
moreover , the column density and covering fraction of @xmath6 indicate a similar degree of absorption of the far - uv continuum between the blue and red @xmath4 bins , implying that systematic differences in this absorption are negligible and thus unlikely to bias our estimate of the far - uv shape of the dust attenuation curve . in comparison ,
the original versions of these dust curves imply a far - uv rise that is too steep to simultaneously model the mid - uv range used to compute @xmath4 ( @xmath131 ) and the far - uv at @xmath132 ( figure [ fig : hih2fit ] ) .
a primary result of our analysis is that the far - uv attenuation curve rises less rapidly toward bluer wavelengths ( @xmath133 ) relative to other commonly assumed attenuation curves .
a consequence of this less - rapid far - uv rise is a lower attenuation of lyc photons by dust . in particular
, the average @xmath4 for galaxies in our @xmath1 sample is @xmath134 and @xmath135 if we assume the @xcite and @xcite reddening curves , respectively . for these values ,
the attenuation of lyc photons at @xmath136 is a factor of @xmath137 lower than that inferred using polynomial extrapolations of the @xcite and @xcite curves at @xmath138 .
clearly , this result has implications for the fraction of hydrogen - ionizing photons that escape galaxies before being absorbed by hydrogen in the igm or cgm , i.e. , the ionizing escape fraction . at face value
, this result implies a larger escape fraction of ionizing photons for galaxies of a given dust reddening .
however , as discussed in reddy et al .
( 2016b ) , the actual covering fraction of neutral gas likely plays a larger role in modulating the escape of such photons than attenuation by dust . specifically , the dominant mode by which ionizing photons escape @xmath9 galaxies at @xmath1 is through channels that have been cleared of gas and dust , while other sightlines are characterized by optically - thick @xmath5 that extinguishes ionizing photons through photoelectric absorption ( reddy et al .
regardless , our new determinations of the far - uv dust curve have direct applications to assessing the attenuation of ionizing photons either internal to regions ( e.g. , @xcite ) or in the context of high - redshift galaxies in the ionized galactic - scale ism or in low column density neutral gas ( i.e. , @xmath116}\la 17.2 $ ] ) , where dust absorption begins to dominate the depletion of ionizing photons .
the revised shape of the far - uv curve also has consquences for the reddening and ages of galaxies derived from stellar population ( i.e. , sed ) modeling . to illustrate this
, we considered 1,874 galaxies from kbss that have ground - based @xmath139 , near - ir intermediate - band ( @xmath140 , @xmath141 , @xmath142 , @xmath143 ) , and broad - band ( @xmath144 , @xmath145 ) photometry , along with _ hubble space telescope _
wfc3/ir f160w and _ spitzer space telescope _ irac imaging ( strom et al .
, in prep ) .
we fit @xcite @xmath146 stellar population models to the photometry , assuming a constant star - formation history , @xmath4 in the range 0.0 - 0.6 , and ages from @xmath147myr to the age of the universe at the redshift of each galaxy ( a full description of the sed - fitting procedure is provided in @xcite ) .
the modeling was performed assuming the original form of the @xcite attenuation curve as well as the updated version ( i.e. , which incorporates the new shape of the far - uv curve found here ) .
figure [ fig : sedcomparison ] shows the comparison between the reddening and ages derived for galaxies in kbss as function of redshift , for the two versions of the attenuation curve .
a less rapid rise of the far - uv attenuation curve implies that a larger reddening is required to reproduce the rest - uv color , which in turn makes the galaxies appear younger .
thus , the updated attenuation curve results in @xmath4 that are on average @xmath148mag redder , and ages that are on average @xmath149myr younger , than those derived with the original @xcite curve . the redder @xmath4 result in sfrs that are about @xmath150 larger , while the stellar masses do not change appreciably .
of course , the magnitude of the effect of changing the far - uv attenuation curve on the derived stellar population parameters depends on which bands are being fit . for galaxies at @xmath15 , where the ly@xmath14 forest is thin and the @xmath151 and @xmath152-bands probe the rest - frame far - uv
, we find systematic offsets in both the reddening and ages of galaxies .
to summarize , we have used a large sample of 933 spectroscopically confirmed @xmath153 lbgs at @xmath1 , 121 of which have very deep spectroscopic coverage in the range @xmath154 , in order to deduce the shape of the far - uv ( @xmath0 ) dust attenuation curve .
we used an iterative procedure to construct composite spectra of the @xmath1 lbgs in bins of continuum color excess , @xmath4 , and , by taking ratios of these spectra , we calculated the far - uv wavelength dependence of the dust attenuation curve .
the curve found here implies a lower attenuation at @xmath12 relative to that inferred from simple polynomial extrapolations of the @xcite and @xcite attenuation curves .
our new determination of the far - uv shape of the dust curve enables more accurate estimates of the effect of dust on the attenuation of hydrogen - ionizing ( lyc ) photons .
specifically , our results imply a factor of @xmath8 lower attenuation of lyc photons for typical @xmath4 relative to that deduced from simple extrapolations of the most widely used attenuation curves at high redshift .
our results may be used to assess the degree of lyc attenuation internal to regions , in the ionized media of high - redshift galaxies , or in low column density ism ( @xmath116}\la 17.2 $ ] ) .
the composite spectra of @xmath1 galaxies can be described well by a model that assumes our new attenuation curve with a unity covering fraction of dust , an @xmath155 covering fraction of @xmath5 , and additional absorption associated with the lyman - werner bands of @xmath6 .
the inferred covering fraction of @xmath6 is very low ( @xmath156 ) and , when contrasted with the high covering fraction of dust , suggests that most of the dust in the ism of high - redshift galaxies is not associated with the catalysis of @xmath6 .
the dust is likely associated with other phases of the ism such as the ionized and/or neutral @xmath5 components .
nar is supported by an alfred p. sloan research fellowship , and acknowledges the visitors program at the institute of astronomy in cambridge , uk , where part of this research was conducted .
ccs acknowledges nsf grants ast-0908805 and ast-1313472 .
mb acknowledges support of the serbian mestd through grant on176021 .
we thank allison strom for providing the photometry for the kbss galaxies .
we are grateful to the anonymous referee whose comments led to significant improvements in the presentation of the analysis .
we wish to extend special thanks to those of hawaiian ancestry on whose sacred mountain we are privileged to be guests . without their generous hospitality , the observations presented herein would not have been possible . | we present the first measurements of the shape of the far - ultraviolet ( far - uv ; @xmath0 ) dust attenuation curve at high redshift ( @xmath1 ) .
our analysis employs rest - frame uv spectra of 933 galaxies at @xmath1 , 121 of which have very deep spectroscopic observations ( @xmath2hrs ) at @xmath3 , with the low resolution imaging spectrograph on the keck telescope . by using an iterative approach in which we calculate the ratios of composite spectra in different bins of continuum color excess , @xmath4 ,
we derive a dust curve that implies a lower attenuation in the far - uv for a given @xmath4 than those obtained with standard attenuation curves .
we demonstrate that the uv composite spectra of @xmath1 galaxies can be modeled well by assuming our new attenuation curve , a high covering fraction of @xmath5 , and absorption from the lyman - werner bands of @xmath6 with a small ( @xmath7 ) covering fraction .
the low covering fraction of @xmath6 relative to that of the @xmath5 and dust suggests that most of the dust in the ism of typical galaxies at @xmath1 is unrelated to the catalysis of @xmath6 , and is associated with other phases of the ism ( i.e. , the ionized and neutral gas ) .
the far - uv dust curve implies a factor of @xmath8 lower dust attenuation of lyman continuum ( ionizing ) photons relative to those inferred from the most commonly assumed attenuation curves for @xmath9 galaxies at @xmath1 .
our results may be utilized to assess the degree to which ionizing photons are attenuated in regions or , more generally , in the ionized or low column density ( @xmath10@xmath11 ) neutral ism of high - redshift galaxies . |
there are different possible measures of cyclicity " of a graph @xmath6 .
one simple such traditional fundamental measure is the cyclomatic number @xmath1 ( also called the first betti number , the nullity or the cycle rank ) which is defined for a connected graph @xmath2 with @xmath3 vertices and @xmath4 edges as @xmath7 motivated from electrical network theory , klein and ivanciuc proposed a new cyclicity measure .
this new cyclicity measure is established on the basis of the novel concept of resistance distance proposed in the excellent paper by klein and randic @xcite . as an intrinsic graph metric , the _ resistance distance _ on a connected graph
@xmath2 is a distance function @xmath8 which may be defined @xcite for @xmath9 as the net effective resistance between vertices @xmath10 @xmath11 @xmath12 when unit resistors are identified with each edge of @xmath2 . comparing to the traditional ( shortest path )
distance function @xmath13 , it is well known that @xmath14 equals the length @xmath15 of the shortest path between @xmath10 and @xmath12 iff there is a unique single path between @xmath10 and @xmath12 , while if there is more than one path , then @xmath14 is strictly less than @xmath15 .
thence the conductance excess @xmath16 indicate in some manner the presence of cyclicity in the portion of the graph interconnecting @xmath10 and @xmath12 , where @xmath17 is known as the effective conductance between @xmath10 and @xmath12 . to measure the cyclicity of a graph @xmath2 , klein and ivanciuc @xcite proposed the _ global cyclicity index _ @xmath0 as @xmath18,\ ] ] where @xmath19 means @xmath10
@xmath11 @xmath12 are adjacent and the sum is over all edges of @xmath2 . since @xmath20 for @xmath19 , @xmath0 can also be written as @xmath21=\sum_{i\sim j}\sigma_g(i , j)-|e(g)|.\ ] ] as a new measure of cyclicity of graphs , the global cyclicity index has less degeneracy than the standard cyclomatic number and has some intuitively appealing features .
since the idea of cyclicity is intimately related to measures of connectivity or complexity @xcite and characterization of cyclicity " is an aspect of key importance in the study of molecular graphs @xcite , it is worth studying the global cyclicity index from both mathematical and chemical points of view . in @xcite , klein and ivanciuc established a number of theorems for the global cyclicity index of graphs (
even not connected ) . in @xcite ,
the present author obtained bounds for the global cyclicity index of fullerene graphs . in this paper
, we proceed to study this interesting cyclicity measure . in view of the fact that the global cyclicity index of a disconnected graph is the sum of global cyclicity indices of its components ( * ? ? ?
* theorem f ) , we focus ourselves only on connected graphs . by utilizing techniques from graph theory , electrical network theory and real analysis
, we obtain some further results on the global cyclicity number , including the strictly monotone increasing property , some lower and upper bounds , and some nordhuas - gaddum - type results .
in particular , we establish a relationship between @xmath0 and the cyclomatic number @xmath1 of a connected graph @xmath2 with @xmath3 vertices and @xmath4 edges : @xmath5
in this section , we will prove the _ strictly monotone increasing property _
( smi - property for short ) of the global cycicity index . for this purpose , we first show if a new edge is added to a graph , then the global cyclicity index will be strictly increased .
it turns out that the following recursion formula for resistance distances @xcite plays an essential roll in proving this point .
[ recur ] ( * ? ? ?
* theorem 2.1 ) let @xmath22 and @xmath23 be resistance distance functions for edge - weighted connected graphs @xmath2 and @xmath24 which are the same except for the weights @xmath25 and @xmath26 on an edge @xmath27 with end vertices @xmath10 and @xmath12 . then for any @xmath28 , @xmath29 ^ 2}{4[1+\delta\cdot\omega(i , j)]},\ ] ] where @xmath30 . theorem [ recur ] is applicable to more general edge - weighted graphs and the weight on each edge represents its conductance ( the reciprocal of resistance ) . in this paper , we only consider simple graphs of unit weight on each edge . especially , it should be pointed out the that if two vertices @xmath10 and @xmath12 are not adjacent , then it is equivalent to say that they are connected by an edge of weight 0 . applying theorem [ recur ] to the global cyclicity index , we have the following result .
[ diff ] let @xmath2 be a connected graph such that @xmath10 and @xmath12 are not adjacent in @xmath2 . denote by @xmath31 the graph obtained from @xmath2 by adding a new edge @xmath32
. then @xmath33 ^ 2}{4\omega^2_g(p , q)[1+\omega_g(i , j)]-\omega_g(p , q)[\omega_g(p , i)+\omega_g(q , j)-\omega_g(p , j)-\omega_g(q , i)]^2}\end{aligned}\ ] ] @xmath34-\sum\limits_{pq\in e(g)}\big[\frac{1}{\omega_{g}(p , q)}-1\big]\notag\\ & = \frac{1}{\omega_{g+ij}(i , j)}-1+\sum\limits_{pq\in e(g)}\big[\frac{1}{\omega_{g+ij}(p , q)}-\frac{1}{\omega_{g}(p , q)}\big].\end{aligned}\ ] ] noticing that the weights on edge @xmath35 in @xmath31 and @xmath2 are @xmath36 and 0 , respectively , hence @xmath37 and by theorem [ recur ] , we have @xmath38 @xmath39 ^ 2}{4[1+\omega_g(i , j)]}.\ ] ] then theorem [ diff ] is obtained by substituting eqs .
( [ e1 ] ) and ( [ e2 ] ) into eq .
( [ e0 ] ) .
furthermore , the difference @xmath40 may be bounded in terms of @xmath14 as given in the following proposition . to obtain the proposition
, we will use the famous foster s ( first ) formula @xcite which demonstrates that the sum of resistance distances between all pairs of adjacent vertices of a connected graph @xmath2 with @xmath3 vertices is equal to @xmath41 , that is , @xmath42 let @xmath2 be a connected graph with @xmath4 edges . if @xmath43 are not adjacent in @xmath2 , then @xmath44 with equality if and only if @xmath2 is the path graph @xmath45 with @xmath10 and @xmath12 being its two end vertices .
it is obvious that @xmath46 , with equality if and only if @xmath47 holds for each @xmath48 .
however , on one hand , by eq .
( [ e2 ] ) , we have @xmath49 , on the other hand , by foster s first formula , @xmath50 hence we can see that there must exist some @xmath48 such that @xmath51 , and consequently , @xmath52 according to theorem [ diff ] , one could easily obtain the strictly monotone increasing property . [ sm](smi - property ) let @xmath2 be a connected graph with @xmath3 vertices and let @xmath64 be a connected spanning subgraph of @xmath2 such that @xmath65
. then @xmath66
for convenience , we first compute the global cyclicity index of some special graphs . let @xmath67 , @xmath68 , @xmath69 and @xmath70 denote a tree with @xmath3 vertices , the cycle graph with @xmath3 vertices , the complete graph with @xmath3 vertices , and the complete bipartite graph with partitions of size @xmath71 and @xmath72 , respectively .
it has been shown in @xcite that @xmath73 and @xmath74 . for @xmath69 , since the resistance distance between every adjacent vertices is @xmath75 , it is easily obtained that @xmath76 for @xmath70 , as it has been shown that the resistance distance between adjacent vertices is @xmath77 , simple calculation leads to @xmath78 by the strictly monotone increasing property , we could establish bounds on the global cyclicity number of some classes of graphs . for a connected graph @xmath2 with @xmath3 vertices
, we have @xmath79 the first equality holds if and only if @xmath2 is a tree , and the second does if and only if @xmath80 . for a connected bipartite graph @xmath2 with partitions of size @xmath71 and @xmath72 , we have @xmath81 the first equality holds if and only if @xmath2 is a tree , and the second does if and only if @xmath82 .
now we turn to another interesting class of graphs circulant graphs .
a circulant graph is a graph of @xmath3 vertices in which the @xmath10th vertex is adjacent to the @xmath83th and @xmath84th vertices for each @xmath12 in a list @xmath85 .
since any circulant graph of order @xmath3 has a hamilton cycle @xmath68 @xcite and it is a subgraph of @xmath69 , by proposition [ sm ] , we have for a connected circulant graph @xmath2 with @xmath3 vertices , we have @xmath86 the first equality holds if and only if @xmath87 , and the second does if and only if @xmath80 .
we first give a lower bound for resistance distances between adjacent vertices . to obtain the lower bound
, we will use a classical result in electrical network theory
rayleigh s short - cut method @xcite : shorting certain sets of vertices together can only decrease the resistance distance of the network between two given vertices .
cutting certain edges can only increase the resistance distance between two given vertices .
let @xmath2 be a connected graph
. for @xmath88 , denote the degree of @xmath89 by @xmath90 .
then [ ouv ] the resistance distance between any two adjacent vertices @xmath91 satisfies : @xmath92 with equality if and only if @xmath93 and there are @xmath94 vertices that are adjacent both to @xmath95 and to @xmath89 .
denote by @xmath96 the graph obtained from @xmath2 by deleting the edge @xmath97 . in @xmath96 ,
short all vertices of @xmath98 together , obtaining a new graph @xmath99 . then by rayleigh s short - cut method , @xmath100 . by serial and parallel connections , @xmath101 hence @xmath102 .
noticing that @xmath103 by the parallel connection , we obtain ( [ e3 ] ) . to have the equality in ( [ e3 ] )
, it requires that @xmath104 .
now we show that if @xmath104 , then @xmath105 .
suppose to the contrary that @xmath106 .
without loss of generality , suppose that @xmath107 .
choose a vertex which is adjacent to @xmath89 but not to @xmath95 , say @xmath25 . in @xmath96 ,
short all vertices of @xmath108 together , obtaining a graph @xmath109 .
then it is not difficult to observe that @xmath110 .
thus @xmath111 , a contradiction .
hence the equality in ( [ e3 ] ) holds only if @xmath105 .
if @xmath105 , then it is obvious that @xmath112 . besides , it has been shown shown in @xcite that @xmath113 if and only if @xmath114 , @xmath93 and there are @xmath94 vertices that are adjacent both to @xmath95 and to @xmath89 . hence equality in ( [ e3 ] ) holds if and only if @xmath93 and there are @xmath94 vertices that are adjacent both to @xmath95 and to @xmath89 .
by lemma [ ouv ] , we have [ cd ] let @xmath2 be a connected graph .
then @xmath115 with equality if and only if @xmath2 is complete . by ( [ e3 ] ) , for any adjacent vertices @xmath10 and
@xmath12 , we have @xmath116 which yields ( [ e4 ] ) .
suppose that @xmath2 satisfies the equality in ( [ e4 ] ) .
then for any two adjacent vertices @xmath117 , we have @xmath118 .
this indicates that @xmath2 is regular since @xmath2 is connected .
suppose that @xmath2 is @xmath119-regular
. then any two adjacent vertices of @xmath2 have @xmath120 common neighbors .
choose any two adjacent vertices @xmath43 and suppose that the @xmath120 common neighbors of them are @xmath121 .
for any vertex @xmath122 ( @xmath123 ) , since @xmath122 is adjacent to @xmath10 , @xmath122 and @xmath10 have @xmath120 common neighbors and thus @xmath124 must be their common neighbors . hence the subgraph induced by @xmath125 is a complete graph of order @xmath126 .
but if @xmath127 , then the subgraph induced by @xmath125 would form a component of @xmath2 which contradicts the hypothesis that @xmath2 is connected .
hence @xmath2 is @xmath128-regular , that is , @xmath2 is a complete graph .
if @xmath2 is regular , then we could obtain an immediate consequence of theorem [ cd ] .
let @xmath2 be a r - rugular graph with @xmath3 vertices .
then @xmath129 denote by @xmath130 the maximum degree of @xmath2 .
then theorem [ cd ] yields let @xmath2 be a connected graph with @xmath4 edges .
then @xmath131 we first introduce some notions and results about the majorization order and schur - convexity ( see @xcite for more details ) .
let @xmath132 given two vectors @xmath133 , the _ majorization order _
@xmath134 means : @xmath135 , and for @xmath136 , @xmath137 . in this paper
, we only consider some subsets of @xmath138 where @xmath139 , @xmath140 . given a closed subset @xmath141 , a vector @xmath142 is said to be maximal for @xmath143 with respect to the majorization order if @xmath144 for each @xmath145 .
analogously , a vector @xmath146 is said to be minimal for @xmath143 with respect to the majorization order if @xmath147 for each @xmath145 .
a symmetric function @xmath148 , is said to be _ schur - convex _ on @xmath149 if @xmath150 implies @xmath151 .
if in addition @xmath152 for @xmath150 but @xmath153 is not a permutation of @xmath154 , @xmath155 is said to be strictly schur - convex on @xmath149 .
@xcite let @xmath156 be an interval and let @xmath157 , where @xmath158 .
if @xmath159 is strictly convex on @xmath160 , then @xmath155 is strictly schur - convex on @xmath161 .
@xcite[lub ] let @xmath162 and @xmath163 .
given the subset @xmath164 we have @xmath165 and @xmath166 where @xmath167 and @xmath168 .
we also need to show two fundamental facts on resistance distances as given in the following two lemmas .
[ day ] let @xmath2 be a connected graph with @xmath3 vertices ( @xmath169 ) .
if @xmath2 has a cut edge , then for any two vertices @xmath91 , @xmath170 suppose that @xmath35 is a cut edge of @xmath2 , and let @xmath99 , @xmath109 be two components of @xmath171 such that @xmath172 , @xmath173 .
if @xmath95 and @xmath89 are in the same component , say @xmath99 , then @xmath174 .
if @xmath95 and @xmath89 are in different components , say @xmath175 and @xmath176 , then @xmath177 .
[ 2n ] let @xmath2 be a connected graph of order @xmath3 ( @xmath169 ) .
if @xmath2 is not complete , there are at most @xmath178 pairs of vertices with resistance distances being equal to @xmath75 .
denote by @xmath179 the graph obtained from @xmath69 by deletion of an edge @xmath97 .
we first show that there are exactly @xmath180 pairs of vertices in @xmath179 with resistance distances being equal to @xmath75 .
it is not hard to compute that @xmath181 and for any pair of vertices @xmath182 , @xmath183 .
but for any vertex @xmath184 , by lemma [ ouv ] , we have @xmath185 hence in @xmath179 , there are exactly @xmath186 pairs of vertices with resistance distances being equal to @xmath75 .
note that @xmath2 is a subgraph of @xmath179 since @xmath2 is not complete .
then the desired result follows since @xmath187 for all @xmath91 according to rayleigh short - cut method . before proving the main results of this section
, we still need to define a class of graphs which will be frequently used in what follows . a connected graph @xmath2 is called _ electrically - edge - equivalent _ if all the resistance distances between pairs of adjacent vertices are equal , that is , for any two edges @xmath188 and @xmath189 , @xmath190 .
as can be easily seen , electrically - edge - equivalent graphs contain some interesting classes of graphs , such as trees and edge - transitive graphs .
now we are ready to give our main results in this section .
[ bound ] let @xmath2 be a connected graph with @xmath3 vertices ( @xmath169 ) and @xmath4 edges . then @xmath191 where @xmath192 .
moreover , the first equality holds if and only if @xmath2 is electrically - edge - equivalent , and the second does if and only if @xmath2 is a tree or @xmath80 .
let @xmath193 , @xmath194, ,@xmath195 be edges of @xmath2 . for simplicity ,
denote the resistance distance between the end vertices of @xmath196 by @xmath197 , @xmath198 . without loss of generality ,
suppose that @xmath199 .
again by the foster s first formula mentioned in the second section , we have @xmath200 in order to obtain the desired inequalities , we evaluate the extremal values of the schur - convex function @xmath201 and consider the set @xmath202 it is easily verified that @xmath203 is strictly schur - convex on @xmath143 . by lemma [ lub ] , we know that the minimal element of @xmath143 with respect to majorization order is given by @xmath204 the function @xmath205 attains its minimum at this point with the minimum value given by @xmath206 .
hence the minimum value of @xmath0 is @xmath207 note that the minimum value of @xmath0 is achieved if and only if resistance distances between all pairs of adjacent vertices are equal .
hence the first equality in ( [ ie1 ] ) holds if and only if @xmath2 is electrically - edge - equivalent .
now we prove the upper bound . by lemma
[ lub ] , the maximal element of @xmath143 with respect to the majorization order is given by @xmath208 where @xmath209 and @xmath210 .
the function @xmath205 attains its minimum at this point with the minimum value given by @xmath211
. hence if we let @xmath212 , then @xmath213 note that achieving the second equality in ( [ ie1 ] ) requirs @xmath214 , @xmath215 and @xmath216 . actually ,
as shown below , the requirement can be satisfied only when @xmath2 is a tree or @xmath80 . for convenience , we distinguish the following three cases . _ case 1 .
_ @xmath217 . in this case ,
together with the fact @xmath219 , we have @xmath220 .
hence @xmath221 and @xmath2 is a tree .
_ @xmath222 .
by @xmath214 , we know that @xmath223 are cut edges of @xmath2 . by lemma [ day ] , for any two vertices @xmath43 , @xmath224 .
but this clearly contradicts the fact that @xmath216 . hence in this case , there does not exist any graph satisfying the desired condition . _
_ @xmath225 . in this case , @xmath226 and @xmath227 .
now we show that @xmath228 .
suppose to the contrary that @xmath229 .
then @xmath2 is not complete .
so on one hand , by lemma [ 2n ] , there are at most @xmath178 pairs of vertices with resistance distances being equal to @xmath75 ; on the other hand , since it is readily verified that @xmath230 according to @xmath231 , there are @xmath232 pairs of vertices in @xmath2 with resistance distances being equal to @xmath75 . but this is an obvious contradiction since @xmath233 .
thus @xmath228 and @xmath2 is complete .
as the upper bound in theorem [ bound ] is somewhat complicated , in the following we give a much simpler upper bound by throwing @xmath234 off .
[ bouud1 ] let @xmath2 be a connected graph with @xmath3 vertices ( @xmath169 ) and @xmath4 edges .
then @xmath235 with equality if and only if @xmath2 is a tree or @xmath80 . by theorem [ bound ]
, we have @xmath236 where @xmath192 . clearly , @xmath237 .
now define the function @xmath205 of @xmath238 as @xmath239 in what follows , we will show that for real variable @xmath240 , @xmath241 attains its maximum value at @xmath242 , which thus will lead to @xmath243 as desired . to this end , consider the first derivative of @xmath205 with respect to @xmath238 , @xmath244 ^ 2}+\frac{n-2}{2}.\ ] ] solving @xmath245 , we have @xmath246 the only thing left is to compare @xmath247 with @xmath248 .
@xmath249 hence @xmath241 attains its maximum value at @xmath242 . in view of the fact that the value of @xmath234 corresponding to @xmath2 is 0
if @xmath2 is a tree or @xmath80 , we could conclude that the equality in ( [ e6 ] ) holds if and only if @xmath2 is a tree or @xmath80 . noticing that @xmath250 is none other than the cyclomatic number @xmath1 , theorem [ bound ] and corollay [ bouud1 ] yield a nice relationship between @xmath0 and @xmath1 .
[ cu ] for a connected graph @xmath2 with @xmath3 vertices and @xmath4 edges , we have @xmath5 the first equality holds if and only if @xmath2 is electrically - edge - equivalent , and the second does if and only if @xmath2 is a tree or @xmath80 .
a nordhaus - gaddum - type result is a ( tight ) lower or upper bound on the sum or product of a parameter of a graph and its complement @xcite . in this section
, we will consider nordhaus - gaddum - type results for the global cyclicity index .
there is only one connected graph @xmath251 on 4 vertices with the connected complement @xmath252 . since @xmath253
, we have @xmath254 . [ ng ]
let @xmath2 be a connected ( molecular ) graph on @xmath255 vertices with a connected @xmath256 . then @xmath257 with equality if and only if @xmath258 and both @xmath2 and @xmath256 are electrically - edge - equivalent .
suppose that @xmath259 . then @xmath260 .
according to the lower bound in theorem [ bound ] , we have @xmath261 ^ 2}{n-1}-[\frac{n(n-1)}{2}-m]\\ & = \frac{m^2+[\frac{n(n-1)}{2}-m]^2}{n-1}-\frac{n(n-1)}{2}\\ & \geq \frac{2[\frac{n(n-1)}{4}]^2}{n-1}-\frac{n(n-1)}{2}=\frac{n(n-1)(n-4)}{8}.\end{aligned}\ ] ] equality holds if and only if @xmath262 , @xmath263 ^ 2}{n-1}-[\frac{n(n-1)}{2}-m]$ ] , and @xmath264 , that is , @xmath258 and both @xmath2 and @xmath256 are electrically - edge - equivalent . for the upper bound , since neither @xmath2 nor @xmath256 is complete and it is impossible that both @xmath2 and @xmath256 are trees , by corollary [ bouud1 ] , we have @xmath265 it is interesting to note for any two connected graphs @xmath99 and @xmath109 of the same order , @xmath266 is no more than twice of @xmath267 and no less than half of @xmath267 , though @xmath268 and @xmath269 may different very much . _
_ graphs satisfying @xmath258 and both @xmath2 and @xmath256 are electrically - edge - equivalent do exist .
an typical example is the famous paley graph . for a prime power @xmath270 ( mod 4 ) , a paley graph
@xmath271 is the graph with vertices the elements of the finite field @xmath272 and an edge between @xmath238 and @xmath273 if and only if @xmath274 is a non - zero square in @xmath272 .
paley graphs satisfy the desired conditions since they are self - complementary and edge - transitive @xcite .
_ remark 2 .
_ although the upper bound in theorem [ ng ] is not tight , it is asymptotically tight . to see this ,
let us consider graphs @xmath2 and @xmath256 as shown in figure 1 .
it is readily seen that the subgraph induced by vertices @xmath275 is the complete graph @xmath276 .
hence @xmath277 is equal to the global cyclicity number of the complete graph @xmath276 , that is , @xmath278 thus @xmath279 hence we could conclude that the upper bound is asymptotically tight .
we complete this section by giving lower and upper bounds for the product of @xmath0 and @xmath277 .
let @xmath2 be a connected ( molecular ) graph on @xmath255 vertices with a connected @xmath256 . then @xmath280 with equality if and only if @xmath2 or @xmath256 is a tree . since the global cyclicity index of a tree is 0 , the lower bound is trivial .
for the upper bound , suppose that @xmath259 .
then by corollary [ bouud1 ] , we have @xmath281\\ & = \frac{n^2}{4}\big[-\big(m-\frac{n(n-1)}{4}\big)^2+\big(\frac{n(n-1)}{4}\big)^2-\frac{n(n-1)^2}{2}+(n-1)^2\big]\\ & \leq \frac{n^2}{4}\big[\big(\frac{n(n-1)}{4}\big)^2-\frac{n(n-1)^2}{2}+(n-1)^2\big]\\ & = \big(\frac{n(n-1)(n-4)}{8}\big)^2.\end{aligned}\ ] ]
the author acknowledges the support of national natural science foundation of china under grant no.11201404 , china postdoctoral science foundation under grant no.2012m521318 , special funds for postdoctoral innovative projects of shandong province under grant no.201203056 .
d. bonchev , a.t .
balaban , x. liu , d.j .
klein , molecular cyclicity and centricity of polycyclic graphs .
i. cyclicity based on resistance distances or reciprocal distances , int .
j. quantum chem .
50 ( 1994 ) 120 . | being motivated in terms of mathematical concepts from the theory of electrical networks , klein & ivanciuc introduced and studied a new graph - theoretic cyclicity index the global cyclicity index ( graph cyclicity , excess conductance , and resistance deficit , j. math .
chem . 30
( 2001 ) 271287 ) . in this paper , by utilizing techniques from graph theory , electrical network theory and real analysis , we obtain some further results on this new cyclicity measure , including the strictly monotone increasing property , some lower and upper bounds , and some nordhuas - gaddum - type results .
in particular , we establish a relationship between the global cyclicity index @xmath0 and the cyclomatic number @xmath1 of a connected graph @xmath2 with @xmath3 vertices and @xmath4 edges : @xmath5 * key words : * resistance distance , global cyclicity index , cyclomatic number , nordhuas - gaddum - type result |
the detection of strong maser emission at the frequencies of pure rotational transitions of some molecules is a common phenomenon in circumstellar envelopes ( cse s ) of evolved stars @xcite .
the maser is usually produced in a small region of the envelope and sometimes provides valuable information on the physical conditions of the emitting region . due to the different chemistry
, masers are produced by different molecules in o- and c - rich stars . in o - rich stars ,
sio exhibits strong maser emission in different rotational transitions within several vibrational states , from @xmath1=1 to 4 @xcite .
these masers are formed in a region of the cse very close to the stellar surface and seem to be driven by nir radiation @xcite . in c - rich stars , although sio is present with similar abundances than in o - rich stars @xcite , no sio maser has been detected .
the explanation could be that sio is formed at @xmath23 - 5 r@xmath3 , where the angular dilution of the star is high and the density and temperature lower than in the regions where sio masers are produced in o - rich stars @xcite . in c - rich stars only hcn shows strong maser emission in several pure rotational lines within vibrational states from @xmath7=1 to 4 @xcite .
these masers must be formed in the innermost regions of the cse .
sis has been previously found to show weak maser emission in the @xmath0=1 - 0 @xmath1=0 transition in irc+10216 @xcite . in this letter
, we report on the first detection of maser emission from the @xmath0=11 - 10 , 14 - 13 and 15 - 14 transitions in the @xmath1=0 vibrational state of sis ( hereafter m@xmath8 , m@xmath5 , and m@xmath9 ) observed toward the c - rich star irc+10216 .
we have also obtained observations of @xmath1=1 rotational lines which exhibit thermal emission .
we propose that overlaps of @xmath1=1 - 0 ro - vibrational transitions of sis with mid - ir lines of c@xmath5h@xmath5 and hcn could provide the pumping mechanism for these masers as well as higher-@xmath0 @xmath1=0 sis masers in the submillimeter range .
this discovery is interesting because this species could play in c - rich stars a role similar to that of sio in o - rich stars : the energy level pattern of both molecules is similar and it is also formed close to the star , as chemical equilibrium and interferometric observations imply @xcite .
the observations of the @xmath1=0 @xmath0=6 - 5 and @xmath0=8 - 7 to @xmath0=15 - 14 transitions of sis ( see left panels , fig . [
fig : figure ] ) were carried out on 2004 june 19@xmath10 with the iram 30 m radio telescope .
four sis receivers operating at 1 , 1.3 , 2 , and 3 mm were used simultaneously .
system temperatures were in the range 120 - 225 k for the 1 , 1.3 and 2 mm receivers and 200 - 600 k for the 1 mm receivers .
atmospheric opacities ranged between 0.08 at 108 ghz to 0.28 at 267 ghz . for the @xmath0=10 - 9 sis line at 181.5 ghz ,
the system temperature was significantly higher , @xmath210@xmath11 k , due to proximity to the atmospheric water line at 183.3 ghz .
the intensity scale was calibrated using two absorbers at different temperatures according to the atmosphere transmission model atm @xcite . pointing and focus were regularly checked on the nearby quasar oj 287 .
the observations were made in wobbling mode , with 180 offset and the secondary nutating at a rate of 0.5 hz . the backends were a filter bank with 256 mhz bandwidth and 1 mhz resolution and an autocorrelator with 80 khz resolution ( @xmath12@[email protected]@xmath130.2 km s@xmath4 ) .
the lines of sis in the @xmath1=1 state ( see right panel , fig .
[ fig : figure ] ) were also observed with the iram 30 m telescope between 1989 and 2001 .
additionally we show some spectra corresponding to ro - vibrational lines of sis ( see center panels in fig . [
fig : figure ] ) from a mid - ir high resolution spectrum ( 11 - 14 @xmath14 m ) obtained in 2002 december with the texes spectrometer mounted on the 3 m telescope irtf ( observational procedures described in * ? ? ?
* ; * ? ? ?
the @xmath1=1 rotational lines show single cusped profiles and their relative intensities indicate that the rotational levels are thermally populated .
the linewidths correspond to velocities of 9 - 11 km s@xmath4 , lower than the terminal velocity @xmath214.5 km s@xmath4 @xcite , thus the emission arises from the innermost region of the cse , between the photosphere and the inner dust formation zone , placed at @xmath25 r@xmath3 @xcite .
we can derive the sis abundance in that region from the @xmath1=1 lines assuming an uniform sphere at a distance of 180 pc with a radius of 10 r@xmath3 , illuminated by the central star ( r@xmath3=5@xmath1510@xmath6 cm , t@xmath3=2300 k ) , with t@xmath16=1000 k and n@xmath17=1.6@xmath1510@xmath18 @xmath19 ( mean values for this region derived by @xcite ) and an expansion velocity of 11 km s@xmath4 .
we solve the statistical equilibrium for sis considering 100 rotational levels and 3 vibrational states and apply the lvg radiative transfer formalism using the code developed by j. cernicharo . with the assumed n@xmath17 ,
the column density for h@xmath5 is @xmath27@xmath1510@xmath20 @xmath21 and the derived one for sis is @xmath25.0@xmath1510@xmath22 @xmath21 .
hence , the sis abundance is @xmath27@xmath1510@xmath23 .
this result is compatible with a higher sis abundance in the innermost cse ( from lte chemistry models , 3@xmath1510@xmath24 , @xcite ) and an abundance further away of 6.5@xmath15@xmath25 according to observations of @xmath1=0 rotational lines over the outer cse by @xcite .
most @xmath1=0 rotational lines show a rounded or slightly double peaked profile with the blue part absorbed by cold sis through the envelope .
however , m@xmath8 , m@xmath5 and m@xmath9 show extra emission in the form of narrow peaks ( fwhm=1 - 3 km s@xmath4 ) . the velocities of these features , within the @xmath26 to 10 km s@xmath4 range , indicate that sis maser emission arises from different regions located between the star and the inner dust formation zone ( @xmath27@xmath25 r@xmath3 ) .
the lines m@xmath5 and m@xmath9 have been previously observed by @xcite but no maser emission was noticed .
this could be due either to the limited sensitivity of their observations or to a time variability of the sis maser phenomenon .
the bottom - center panel of fig .
[ fig : figure ] shows in detail the line profiles of the observed masers .
up to ten maser features labelled ( @xmath28, ,@xmath29 ) are identified .
the most complex line profile is that of m@xmath9 .
it is formed by 5 main features : @xmath28 , @xmath30 , @xmath31 , @xmath32 , @xmath29 with @xmath1=@xmath33 , @xmath34 , @xmath35 , @xmath36 and @xmath37 km s@xmath4 , respectively .
the line profiles show that the strongest peaks are at negative velocities , having their red counterparts rather weak .
this behavior was previously found by @xcite for the @xmath1=0 @xmath0=1 - 0 line , which mostly consists of a narrow peak ( fwhm=0.3 km s@xmath4 ) centered at @xmath38 km s@xmath4 .
the strong asymmetry of the lineshapes can be either due to blanking by the star of the redshifted maser or by amplification of the blueshifted emission by the foreground stellar environment .
the strongest observed maser , m@xmath9 , with t@xmath39@xmath260 k ( f@xmath40@xmath2300 jy ; with thermal and non - thermal emission ) , is weak compared to some sio masers detected in o - rich stars ( e.g. * ? ? ? * ) or to hcn masers observed in irc+10216 @xcite .
however , m@xmath8 , m@xmath5 , and m@xmath9 are stronger than the sis @xmath0=1 - 0 maser observed towards irc+10216 by @xcite .
the similarity between maser features in m@xmath5 and m@xmath9 indicates that they may arise from the same regions and produced by the same pumping mechanism ; the maser in m@xmath8 is probably formed in other regions .
hence , we suggest two possible geometries of the innermost cse to explain the observed features : ( @xmath32 ) an onion - like innermost region , where each maser is produced in a shell .
this hypothesis is supported by the symmetry of features @xmath28@xmath29 and @xmath30@xmath32 .
the peaks at extreme velocities , @xmath28 and @xmath29 , would be produced just in front of and behind the star near the inner dust formation shell ( @xmath27@xmath25 r@xmath3 ) with expansion velocities of @xmath25 - 11 km s@xmath4 .
the features @xmath30 and @xmath32 would be produced in a similar way but in an inner shell with a lower expansion velocity .
finally , the central peak , @xmath31 , would be formed in a shell very close to the star , with the whole shell contributing to the maser emission .
m@xmath8 would be produced in a cap - shaped region in front of the star .
( @xmath41 ) all the masers are formed in different positions of a clumpy shell .
the different features in m@xmath5 and m@xmath9 would be produced in different regions of the shell : peaks @xmath28 and @xmath29 in front of and behind the star and the other peaks ( @xmath30 , @xmath31 , and @xmath32 ) in different clumps , as occur with the only feature of m@xmath8 .
the classic pumping mechanism for the sio @xmath1@xmath420 masers observed in o - rich stars resides in the increase of the trapping lifetime ( a/@xmath43)@xmath44 with @xmath0 for @xmath1@xmath45@xmath1 - 1 transitions , when they become optically thick @xcite .
such mechanism produces masers in adjacent rotational lines of the @xmath1 state , and explains the @xmath1=1 and 2 sio masers @xcite .
however , the masers observed in rotational transitions of @xmath46sio , @xmath47sio , and in @xmath1=3 and 4 of sio do not show the latter behavior and have been interpreted as due to ir overlaps between ro - vibrational lines of sio isotopologues @xcite . for sis ,
the absence of maser emission in @xmath1=1 rotational lines and the odd @xmath1=0 pattern also exclude the kwan & scoville pumping mechanism .
this suggests that overlaps of @xmath1=1 - 0 ro - vibrational transitions of sis with those of mode @xmath48 of c@xmath5h@xmath5 and mode @xmath7 of hcn , could provide the pumping mechanism .
c@xmath5h@xmath5 and hcn are abundant in the inner cse of irc+10216 and dominate the 11 - 14 @xmath14 m spectrum @xcite .
overlaps with these two species have been already proposed by @xcite to explain the different profiles of adjacent @xmath0 lines of sis . however , the sis frequencies used in that work were not as accurate ( @xmath49@xmath5010@xmath4 cm@xmath4 ) as those available today .
we have calculated those frequencies from the dunham coefficients determined by @xcite , for which the error of the band center is @xmath52@xmath53 cm@xmath4 ( @xmath20.04 km s@xmath4 ; the relative accuracy of p and r lines is much better ) .
the frequencies of c@xmath5h@xmath5 , h@xmath6cch , hcn , and h@xmath6cn lines have been taken from the hitran database 2004 @xcite , with an accuracy better than @xmath54 cm@xmath4 ( @xmath20.4 km s@xmath4 ) for c@xmath5h@xmath5 and h@xmath6cch , and @xmath53 cm@xmath4 for hcn and h@xmath6cn .
table [ tab : frequencies_sis ] shows the mid - ir line overlaps of sis with c@xmath5h@xmath5 , hcn , and their most abundant isotopologues . for the overlap search we selected coincidences within @xmath55@xmath5210 km s@xmath4 .
however , since the cse is expanding , every region of the envelope is receding from the others . hence
, if the population of the sis levels is affected by an overlap with a strong line of other species , the frequency of this overlapping transition must be higher than the sis one .
this would restrict the condition to positive @xmath56 . nevertheless , due to the linewidth , lines at @xmath56@xmath520 can overlap the sis lines .
hence , we have set the negative cutoff to one half of the typical linewidth in the innermost cse ( @xmath25 km s@xmath4 ; @xcite ) .
therefore , our search is restricted to @xmath57 km s@xmath4 .
all the ro - vibrational sis lines commented hereafter refer to @xmath1=1@xmath450 transitions and will be labelled with the usual spectroscopic nomenclature r , q , p ( see footnote of table [ tab : frequencies_sis ] ) .
c@c|c@c@c + r 9 & 750.3695 & hcn & @xmath58 r@xmath59 & @xmath60 + p10 & 738.2889 & c@xmath5h@xmath5 & @xmath61 r@xmath62 & 1.2 + r13 & 752.6431 & c@xmath5h@xmath5 & @xmath63 r@xmath64 & 5.9 + p14 & 735.7324 & c@xmath5h@xmath5 & @xmath65 q@xmath66 & @xmath67 + p15 & 735.0861 & c@xmath5h@xmath5 & @xmath65 q@xmath68 & @xmath69 + + p 2 & 743.2624 & c@xmath5h@xmath5 & @xmath63 r@xmath70 & 0.6 + p16 & 734.4368 & c@xmath5h@xmath5 & @xmath65 q@xmath71 & 1.1 + p20 & 731.8114 & c@xmath5h@xmath5 & @xmath65 q@xmath72 & 9.5 + p22 & 730.4815 & c@xmath5h@xmath5 & @xmath73 q@xmath74 & @xmath75 + r22 & 757.5819 & c@xmath5h@xmath5 & @xmath76 r@xmath77 & 0.1 + r22 & 757.5819 & h@xmath6cn & @xmath58 r@xmath78 & 5.2 + p23 & 729.8123 & h@xmath6cch & @xmath65 q@xmath79 & 8.7 + p24 & 729.1402 & c@xmath5h@xmath5 & @xmath65 q@xmath80 & 9.4 + p25 & 728.4654 & h@xmath6cch & @xmath65 q@xmath81 & @xmath82 + r25 & 759.1733 & hcn & @xmath58 r@xmath83 & 3.6 + r26 & 759.6977 & c@xmath5h@xmath5 & @xmath63 r@xmath59 & @xmath84 + r33 & 763.2816 & h@xmath6cn & @xmath58 r@xmath79 & 6.1 + p38 & 719.4363 & c@xmath5h@xmath5 & @xmath76 p@xmath70 & 4.5 + r40 & 766.7130 & c@xmath5h@xmath5 & @xmath63 r@xmath83 & 4.3 + r42 & 767.6652 & c@xmath5h@xmath5 & @xmath85 r@xmath86 & 8.7 + r45 & 769.0697 & c@xmath5h@xmath5 & @xmath63 r@xmath87 & @xmath88 + p48 & 712.1720 & c@xmath5h@xmath5 & @xmath89 p@xmath90 & 4.7 [ tab : frequencies_sis ] in order to qualitatively interpret the effects of ir overlaps on maser emission , we have used the same lvg radiative transfer code modified to account for overlaps , changing the intensity at the overlapping frequency and the escape probability for photons from the overlapped lines .
thus , for some sis lines , the excitation temperature becomes negative ( maser activation ) and the brightness temperature is considerably enhanced .
m@xmath5 and m@xmath9 are naturally explained by the overlap of the r@xmath91 line of sis with the strong c@xmath5h@xmath5 line @xmath92r@xmath64 ( table [ tab : frequencies_sis ] ) .
sis molecules are easily excited from @xmath1=0 @xmath0=13 to @xmath1=1 @xmath0=14 through r@xmath91 and decay to the @xmath1=0 @xmath0=15 via p@xmath93 , creating a population inversion between the @xmath1=0 @xmath0=13 and 15 , and producing maser emission in m@xmath5 and m@xmath9 .
m@xmath8 may be produced by the overlap of the p@xmath94 sis line with the strong c@xmath5h@xmath5 line @xmath95r@xmath62 .
this overlap can pump sis molecules from @xmath1=0 @xmath0=10 to @xmath1=1 @xmath0=9 through p@xmath94 , depopulates the @xmath1=0 @xmath0=10 level and produces the inversion between @xmath1=0 @xmath0=10 and 11 .
we have also looked for overlaps of @xmath1=1 - 0 higher-@xmath0 sis lines with c@xmath5h@xmath5 and hcn transitions to try to predict sis masers at submillimeter wavelengths .
some of them are shown in the second block of table [ tab : frequencies_sis ] .
they suggest , for example , that a maser could be found in rotational transitions involving the @xmath1=0 @xmath0=23 level ( likely @xmath0=24 - 23 and maybe 23 - 22 ) , or the @xmath1=0 @xmath0=25 and 26 states ( perhaps in the @xmath1=0 @xmath0=27 - 26 and maybe 26 - 25 ) .
these overlaps could also produce masers in @xmath1=1 rotational transitions .
there are many overlaps of other sis isotopologues with lines of c@xmath5h@xmath5 , hcn and @xmath96si@xmath97s . with the adopted criteria for the overlap search
( see footnote of table [ tab : frequencies_sis ] ) , we have found 91 , 93 , 76 , and 94 coincidences for @xmath46sis , @xmath47sis , si@xmath98s , and si@xmath99s , respectively .
consequently , although these species are less abundant than @xmath96si@xmath97s , the population of some levels could be inverted producing maser emission .
this study represents the discovery of three new sis masers and should be complemented with future observations of higher-@xmath0 @xmath1=0 and 1 rotational transitions .
furthermore , a detailed multi - molecule non - local radiative transfer model would help to understand the dust formation region and the role of sis in c - rich evolved stars .
we want to thank the iram staff , m. j. richter ( nsf grant ast-0307497 ) , j. h. lacy ( nsf grant ast-0205518 ) , and collaborators for the mid - infrared observations , and spanish mec for funding support through grant esp2004 - 665 , aya2003 - 2785 , and `` comunidad de madrid '' government under pricit project s-0505/esp-0237 ( astrocam ) .
this study is supported in part by the european community s human potential programme under contract mcrtn - ct-2004 - 51230 , molecular universe .
agndez , m. & cernicharo , j. , 2006 , , astro - ph/0605645 bieging , j. , h. & nguyen - q - rieu , 1989 , , 343 , l25 bujarrabal , v. & nguyen - q - rieu , 1981 , , 102 , 65 cernicharo , j. 1985 , atm : a program to compute atmospheric opacity between 0 and 1000 ghz , iram internal report cernicharo , j. , bujarrabal , v. , & lucas , r. , 1991 , , 249 , l27 cernicharo , j. , bujarrabal , v. , & santarn , j. , l. , 1993 , , 407 , l33 cernicharo , j. , gulin , m. , & kahane , c. , 2000 , , 142 , 181 elitzur , m. , 1992 , , 30 , 75 fonfra expsito , j. p. , cernicharo , j. , richter , m. j. , & lacy , j. , 2006 , , submitted gonzlez - alfonso , e. & cernicharo , j. , , 322 , 938 gray , m. , 1999 , philos
london a. , 357 , 3277 henkel , c. , matthews , h. e. , & morris , m. , 1983 , , 267 , 184 keady , j. j. , hall , d. n. b. , & ridgway , s. , t. , 1988 , , 326 , 832 keady , j. j. & ridgway , s. , t. , 1993 , , 406 , 199 kwan , j. & scoville , n. , 1974 , , 194 , l97 lacy , j. h. , richter , m. j. , greathouse , t. k. et al . , 2002 , , 114 , 153 lafont , s. , lucas , r. , & omont , a. , 1982 , , 106 , 201 lockett , p. & elitzur , m. , 1992 , , 399 , 704 lucas , r. & cernicharo , j. , 1989 , , 218 , l20 pardo , j. r. , cernicharo , j. , gonzlez - alfonso , e. , & bujarrabal , v. , 1998 , 329 , 219 pardo , j. r. , cernicharo , j. , & serabyn , g. , 2001 , ieee trans .
antennas propagation , 46 , 1683 pardo , j. r. , alcolea , j. , bujarrabal , v. , colomer , f. , del romero , a. , & de vicente , p. , 2004 , 424 , 145 rothman , l. s. , jacquemart , d. , et al .
, 2005 , j. quant . spec .
transf . , 96 , 139 sahai , r. , wootten , a. , & clegg , r. e. s. , 1984 , , 284 , 144 sanz , m. e. , mccarthy , m. c. , & thaddeus , p. , 2003
, , 119 , 11715 schilke , p. & menten , k. m. , 2003 , , 583 , 446 schier , f. l. , olofsson , h. , & lundgren , a. a. , 2006 , , astro - ph/0604213 | we report on the first detection of maser emission in the @xmath0=11 - 10 , @xmath0=14 - 13 and @xmath0=15 - 14 transitions of the @xmath1=0 vibrational state of sis toward the c - rich star irc+10216 .
these masers seem to be produced in the very inhomogeneous region between the star and the inner dust formation zone , placed at @xmath25 - 7 r@xmath3 , with expansion velocities below 10 km s@xmath4 .
we interpret the pumping mechanism as due to overlaps between @xmath1=1 - 0 ro - vibrational lines of sis and mid - ir lines of c@xmath5h@xmath5 , hcn and their @xmath6c isotopologues .
the large number of overlaps found suggests the existence of strong masers for high-@xmath0 @xmath1=0 and @xmath1=1 sis transitions , located in the submillimeter range .
in addition , it could be possible to find several rotational lines of the sis isotopologues displaying maser emission . |
throughout the paper unless stated otherwise all manifolds are smooth and connected , all metrics are complete riemannian , and the phrase `` nonpositive sectional curvature '' is abbreviated to `` @xmath5 '' . a manifold is _ covered by @xmath6 _ if its universal cover is diffeomorphic to @xmath6 .
the cartan - hadamard theorem yields a basic obstruction to @xmath5 : any @xmath7-manifold with a metric of @xmath5 is covered by @xmath6 .
recall that an open @xmath1 manifold covered by @xmath6 exists if and only if @xmath0 is a countable group of finite cohomological dimension .
countable groups of finite cohomological dimension form a class of torsion - free groups that is closed under subgroups , extensions , and amalgamated products .
a much deeper obstruction to @xmath5 arises from gromov s random , torsion - free , hyperbolic groups with strong fixed point properties : each of the groups can be realized as the fundamental group of open manifold covered by @xmath6 for large enough @xmath7 , yet no such manifold is homotopy equivalent to a manifold of @xmath5 , see @xcite and ( * ? ? ?
* theorem 1.1 ) . even though the focus of this paper is on open manifolds , we next review known obstructions to @xmath5 for closed manifolds .
there exist closed manifolds that are covered by @xmath6 and have one of the following properties : their fundamental groups are not cat(@xmath8 ) , see @xcite , they are homeomorphic but not diffeomorphic to manifolds of @xmath5 @xcite , they admit locally cat(@xmath8 ) metrics but are not homotopy equivalent to closed manifold of @xmath5 @xcite . a manifold is _ covered by @xmath9 _ if it is diffeomorphic to the product of @xmath10 and a manifold covered by @xmath11 . in the paper we describe a sizable class of groups such that for each group @xmath0 in the class 1 .
every @xmath1 manifold of @xmath5 is covered by @xmath9 ; 2 .
there are @xmath1 manifolds that are covered by @xmath6 but not homeomorphic to manifolds covered by @xmath9 .
this yields a homeomorphism obstruction to @xmath5 for if @xmath0 satisfies ( i)-(ii ) , then no manifold in ( ii ) is homeomorphic to a manifold of @xmath5 .
manifolds covered by @xmath9 arise in the context of nonpositive curvature , namely , consider a manifold of @xmath5 written as @xmath12 where @xmath0 is the deck transformations group acting isometrically on the universal cover @xmath13 . if @xmath0 _ fixes a busemann function _
( i.e. @xmath0 fixes a point at infinity and stabilizes each horosphere centered at the point ) , then @xmath12 is covered by @xmath9 .
more precisely , any busemann function is a @xmath14-submersion @xcite , so it defines a @xmath15 diffeomorphism of @xmath13 and @xmath9 , and if the busemann function is @xmath0-invariant , the @xmath15-diffeomorphism descends to quotients , which are therefore @xmath16-diffeomorphic .
the property `` @xmath0 fixes a busemann function '' can be forced by purely algebraic assumptions on @xmath0 summarized in sections [ sec : center]-[sec : grounded ] , where we also describe various classes of groups with this property some of which are listed below .
[ ex - intro : busemann list ] any of the following conditions implies that each isometric , properly discontinuous @xmath0-action on a hadamard manifold @xmath13 fixes a busemann function : * @xmath17 where @xmath18 is an aspherical manifold that has a finitely - sheeted cover which is the total space of an oriented circle bundle with nonzero rational euler class .
this applies when @xmath18 is an closed infranilmanifold of nilpotency degree @xmath19 , or a closed @xmath20-manifold modelled on @xmath21 . *
@xmath22 where @xmath18 is as in ( 1 ) and @xmath23 is any group . *
@xmath24 where @xmath25 and @xmath23 is a finite index subgroup of @xmath26 , and the semidirect product is defined via the standard @xmath23-action on @xmath27 .
* @xmath0 is any group with infinite center and finite abelianization , e.g. the preimage of any lattice under the universal cover @xmath28 , @xmath29 ; actually this example is also covered by ( 1 ) .
* @xmath0 is the amalgamated product along @xmath27 of the group @xmath30 as in ( 4 ) and any group @xmath31 that contains @xmath27 as a normal subgroup .
* @xmath0 has infinitely generated center , e.g. the product of @xmath32 and any group . in section [ sec : visib ] we show that if the hadamard manifold @xmath13 is _ visibility _ ( i.e. any two distinct points at infinity are the endpoints of a geodesic in @xmath13 ) , then the class of groups that fix busemann functions on @xmath13 gets even larger , e.g. it contains any nontrivial product @xmath33 such that there is a @xmath34 manifold .
an essential tool in understanding manifolds covered by @xmath9 is the recent result of guilbault @xcite : if an open manifold @xmath3 of dimension @xmath35 is homotopy equivalent to a finite complex , then @xmath36 is diffeomorphic to the interior of a compact manifold . as an addendum to @xcite
we show : [ thm - intro : reg nbhd ]
let @xmath3 be an open @xmath37-manifold with @xmath38 that is homotopy equivalent to a finite complex of dimension @xmath39 .
then @xmath36 is diffeomorphic to the interior of a regular neighborhood of a @xmath40-dimensional finite subcomplex .
in addition to fixing a busemann function there are other properties that can force any open @xmath1 manifold of @xmath5 to be the interior of a regular neighborhood of a finite subcomplex .
[ intro - ex : flat ] if @xmath0 be the fundamental group of a closed flat manifold , then for any @xmath1 manifold with @xmath5 , the @xmath0-action on the universal cover @xmath13 either fixes a busemann function or stabilizes a flat @xmath41 where it acts cocompactly ; in the latter case @xmath12 is diffeomorphic to the normal bundle of the closed flat submanifold @xmath42 , see lemma [ lem : center has parabolic ] and ( * ? ? ?
* corollary ii.7.2 ) .
[ intro - ex : higher rank ] let @xmath0 be an irreducible torsion - free uniform lattice in a connected higher rank semisimple lie group with finite center and no compact factor .
then harmonic maps superrigidity ( * ? ? ?
* theorem 1.2 ) implies that for any @xmath1 manifold with @xmath5 , the @xmath0-action on the universal cover @xmath13 either acts cocompactly on a totally geodesic subspace @xmath41 , or fixes a point at infinity in which case it also fixes the corresponding busemann function as @xmath43 is finite , see remark [ rmk : no hom into r ] . as we show later , theorem [ thm - intro : reg nbhd ] and example [ intro - ex : flat ] yield an attractive characterization of @xmath6 : [ cor - intro : rn ] an open contractible @xmath44-manifold @xmath3 is homeomorphic to @xmath6 if and only if @xmath4 admits a metric of @xmath5 . if @xmath45 , then `` homeomorphic '' in corollary [ cor - intro : rn ] can not be replaced with `` diffeomorphic '' because if @xmath3 is an exotic @xmath46 , then @xmath4 is diffeomorphic to @xmath47 ( *
* theorem 2.2 )
. since there are only countably many diffeomorphism classes of compact manifolds , there are only countably many possibilities for open regular neighborhoods .
various classification results for manifolds diffeomorphic to the interiors of regular neighborhoods can be found in @xcite and ( * ? ? ?
* corollary 11.3.4 ) . in the stable range , i.e.
when the subcomplex has dimension @xmath48 , the interior of a regular neighborhood of dimension @xmath7 is determined up to diffeomorphism by its tangential homotopy type @xcite .
let @xmath0 be a group as in . if @xmath49 is an open @xmath7-manifold that is homotopy equivalent to a finite @xmath1 complex of dimension @xmath48 , where @xmath38 , then the tangential homotopy type of @xmath49 contains at most one manifold of @xmath5 .
if an @xmath7-dimensional regular neighborhood @xmath50 is homotopy equivalent to a closed manifold @xmath51 , then @xmath52 is the total space of a vector bundle over @xmath51 , provided @xmath53 and @xmath54 @xcite ; the former inequality is known as the metastable range . together with theorem [ thm - intro : reg nbhd ] this implies : let @xmath51 be a closed aspherical @xmath55-manifold as in or , and let @xmath49 be an open @xmath7-manifold of @xmath5 that is homotopy equivalent to @xmath51 . if @xmath56 and @xmath54 , then @xmath49 is diffeomorphic to the total space of a vector bundle over @xmath51 .
we now turn to methods of producing manifolds as in ( ii ) .
a trivial method is to consider a @xmath1 manifold that has the minimal dimension among all @xmath1 manifolds that are covered by @xmath6 , which yields : any aspherical manifold is homotopy equivalent to a manifold covered by @xmath6 but not covered by @xmath9 .
if @xmath0 is as in , then any @xmath1 manifold is homotopy equivalent to a manifold that admits no metric of @xmath5 and is covered by a euclidean space .
this trivial method always works , yet it is non - constructive for it is not easy to decide whether a specific open manifold has the minimal dimension in the above sense , see @xcite . on the other hand ,
if @xmath0 has a finite @xmath1 complex , we produce explicit manifolds satisfying ( ii ) : [ thm - intro : count many ] let @xmath49 be an open aspherical @xmath7-manifold that is homotopy equivalent to a finite complex of dimension @xmath57 , where @xmath38 .
then the tangential homotopy type of @xmath49 contains countably many manifolds such that each of them is covered by @xmath6 and is not homeomorphic to the interior of a regular neighborhood of a finite subcomplex of codimension @xmath2 , and in particular not covered by @xmath9 .
it should be possible to strengthen the conclusion `` countably many '' in theorem [ thm - intro : count many ] to `` a continuum of '' , but we only prove this under a technical assumption which holds in many cases of interest , including when @xmath58 is a poincar duality group , or when @xmath58 does not contain a subgroup isomorphic to @xmath59 , see section [ sec : without r - factors ] .
the structure of the paper is as follows : sections [ sec : center ] , [ sec : grounded ] , [ sec : visib ] discuss when a group fixes a busemann function , and in particular , justify examples [ ex - intro : busemann list ] , [ intro - ex : flat ] , [ intro - ex : higher rank ] . in section [ sec : with r - factors ] we study manifolds with @xmath10-factors , and prove theorem [ thm - intro : reg nbhd ] and corollary [ cor - intro : rn ] . in section
[ sec : without r - factors ] we construct manifolds without @xmath10-factors , and prove theorem [ thm - intro : count many ] as well as its harder uncountable version .
the latter requires a certain family of homology @xmath20-spheres described in appendix [ sec : homology sph ] . throughout the paper
we make implicit use of various finiteness results for groups , which are recalled in appendix [ sec : app on finitenss ] .
* acknowledgments . *
the author is grateful for nsf support ( dms-1105045 ) and to mathoverflow for a stimulating work environment .
thanks are due to karel dekimpe for examples [ ex : dekimpe 3d ] and [ ex : dekimpe poly ] , misha kapovich for example [ ex : kapovich ex ] , andy putman and wilberd van der kallen for help with proposition [ prop : putman ex ] .
background references for nonpositive curvature are @xcite . in this section
@xmath0 is a discrete isometry group of a hadamard manifold @xmath13 .
[ lem : periodic group ] if @xmath3 is a closed convex subset of @xmath13 , then any abelian discrete subgroup @xmath60 that contains no parabolic isometries is finitely generated of rank @xmath61 . by ( * ? ? ?
* lemma 7.1(1 ) ) the set @xmath62 is a closed @xmath63-invariant convex subset of @xmath64 which splits as @xmath65 , where @xmath63 acts trivially on the @xmath66-factor , and by translations on @xmath67 . since @xmath63 is discrete , it acts freely on @xmath67 so @xmath68 for @xmath69 .
[ lem : center has parabolic ] @xmath0 preserves a busemann function if it has a finite index subgroup @xmath70 whose center @xmath71 contains a parabolic isometry .
this result is implicit in ( * ? ? ?
* lemma 7.3 , 7.8 ) , but for completeness we record a proof below .
let @xmath72 be a parabolic isometry , and @xmath73 are right coset representatives of @xmath70 in @xmath0 , then the function @xmath74 is convex , and @xmath0-invariant because @xmath75 is independent of a coset representative in @xmath76 as @xmath72 . since @xmath77 does not assume its infimum , neither does the above convex function . then a limiting process outlined in ( *
* lemma 3.9 ) , and fully explained in ( * ? ? ?
* lemma ii.8.26 ) , gives rise via arzela - ascoli theorem to a @xmath0-invariant busemann function .
note that this busemann function is not explicit and there seems to be no way to find a @xmath0-invariant busemann function associated with a given point of @xmath78 .
flat torus theorem ( * ? ? ?
* chapter ii.7 ) implies that a discrete isometry group of @xmath13 whose center consists of axial isometries is quite special , which for our purposes is summarized as follows .
[ thm : into - center ] let @xmath0 be a group with subgroups @xmath23 , @xmath70 such that their centers @xmath79 , @xmath71 are infinite , @xmath80 , the index of @xmath70 in @xmath0 is finite , and one of the following conditions hold : @xmath79 is not finitely generated ; any homomorphism @xmath81 is trivial .
@xmath23 is finitely generated , and @xmath79 contains a free abelian subgroup that is not a direct factor of any finite index subgroup of @xmath23.then any effective , properly discontinuous , isometric @xmath0-action on a hadamard manifold fixes a busemann function .
the reader may want to first think through the case when @xmath82 , and then go on to observe that if theorem [ thm : into - center ] applies to @xmath23 , @xmath70 , @xmath0 , then it also applies to @xmath23 , @xmath83 , @xmath84 for any group @xmath31 .
the result follows by lemma [ lem : center has parabolic ] once we show that @xmath71 contains a parabolic isometry . since @xmath80 it suffices to find a parabolic element in @xmath79 .
suppose no such element exists .
then @xmath79 is finitely generated by lemma [ lem : periodic group ] , which implies ( 1 ) . to prove ( 2 ) note that if @xmath23 has no nontrivial homomophism into @xmath10 , then by ( * ? ? ?
* remark ii.6.13 ) @xmath79 contains no axial isometries , so it is a finitely generated , abelian , and periodic , and hence finite which contradicts the assumptions .
finally , ( 3 ) is deduced as follows .
since @xmath79 is finitely generated and infinite , it has a free abelian subgroup @xmath63 of positive rank . under the assumption that @xmath23 is finitely generated , ( * ? ? ?
* theorem ii.7.1 ) implies that @xmath23 has a finite index subgroup that contains @xmath63 as a direct factor .
theorem [ thm : into - center](1 ) applies , e.g. to any infinitely generated , torsion - free , countable abelian group of finite rank , such as @xmath32 , where finiteness of rank ensures finiteness of cohomological dimension ( * ? ? ?
* theorem 7.10 ) .
is there a group of type _ f _ with infinitely generated center ?
there exists a finitely presented group with solvable word problem whose center contains every countable abelian group ( * ? ? ?
* corollary 3 ) . on the other hand ,
if @xmath23 is a group with infinitely generated center , then @xmath23 can not be embedded into hyperbolic or cat(@xmath8 ) groups because their abelian subgroups are finitely generated .
if in addition @xmath23 has type _ f _ , or even of type _ fp _ , then @xmath23 can not be linear over a field of characteristic zero ( * ? ? ?
* corollary 5 ) , and can not be elementary amenable , as we explain below for completeness .
every elementary amenable group of type _ fp _ has finitely generated center . since @xmath0 has type _ fp _
, its cohomological dimension is finite .
any elementary amenable of finite cohomological dimension is virtually solvable @xcite and of course torsion - free . as @xmath0 has type _ fp _ ,
so does the solvable finite index subgroup @xmath70 of @xmath0 ( * ? ? ?
* theorem viii.6.6 ) .
solvable groups of type _ fp _ , are constructable @xcite and hence quotients of @xmath70 are finitely presented ( * ? ? ?
* theorem 4 ) .
so @xmath71 is finitely generated , for if the quotient of a finitely generated group by the center is finitely presented , then the center is finitely generated ( * ? ? ?
* corollary 1 , page 54 ) .
hence the subgroup @xmath85 of @xmath71 is finitely generated , and therefore the same holds for @xmath86 .
here is an example of a group of type _ f _ with infinite center and finite abelianization ( to which theorem [ thm : into - center](2 ) applies ) .
let @xmath0 be the preimage of a lattice @xmath87 under the universal cover , then @xmath86 is infinite , and
@xmath0 has kazhdan property ( t ) when @xmath29 ( * ? ? ?
* example 1.7.13 ) , and hence finite abelianization . if @xmath88 is torsion - free , then it has type _ f _
@xcite , and hence so does @xmath0 , see appendix [ sec : app on finitenss ] .
note that theorem [ thm : into - center](3 ) also applies to @xmath0 because its finite index subgroups have property ( t ) and hence finite abelianization which implies that they can not have @xmath89 as a direct factor .
theorem [ thm : into - center](3 ) also applies when @xmath0 is virtually nilpotent , finitely generated , and not virtually abelian , namely , we let @xmath90 be a nilpotent torsion - free subgroup of finite index , and use the following .
[ prop : nil not virt split ] if @xmath23 is a torsion - free , nonabelian , nilpotent group @xmath23 , then @xmath79 is not a virtual direct factor of @xmath23 .
arguing by contradiction suppose @xmath91 is a finite index subgroup of @xmath23 that contains @xmath79 as a direct factor , i.e. there is @xmath92 with @xmath93 and @xmath94 .
then @xmath95 where @xmath96 .
a key observation is that @xmath97 : if we take @xmath98 and @xmath99 , and pick @xmath55 with @xmath100 , then @xmath101 but in torsion - free nilpotent groups roots are unique ( * ? ? ?
* theorem 16.2.8 ) , so @xmath102 , and hence @xmath103 .
it follows that @xmath104 is trivial , so since @xmath31 is nilpotent we conclude that @xmath31 is trivial , i.e. @xmath105 . now
if @xmath106 , then @xmath107 so @xmath108 and again uniqueness of roots implies @xmath109 so that @xmath23 is abelian . it is a well - known fact ( which we will not use , and hence do not prove ) that the group @xmath23 in proposition [ prop : nil not virt split ] can be realized as the fundamental group of the circle bundle over a closed nilmanifold , and noncommutativity of @xmath23 allows one to choose the bundle so that its rational euler class is nonzero .
thus proposition [ prop : nil not virt split ] is also implied by the proposition below .
[ prop : euler class ] if @xmath51 is an aspherical manifold , and @xmath110 is a principal circle bundle with nonzero rational euler class , then the fiber circle is a nontrivial element @xmath111 in the center of @xmath112 that is not a virtual direct factor . the circle action on @xmath18 lifts to any cover @xmath113 whose fundamental group contains @xmath111 ( * ? ? ?
* theorem i.9.3 ) .
therefore , the covering @xmath114 descends to a covering of base spaces @xmath115 .
it follows that @xmath116 is a pullback of @xmath110 via @xmath115 .
if @xmath116 is a finite cover , then it has a nonzero rational euler class because finite covers induce injective maps on rational cohomology ( due to existence of a transfer ) . finally ,
if @xmath111 were a direct factor in @xmath117 , then @xmath116 would have a section , so the euler class would be zero .
theorem [ thm : into - center](3 ) also applies to the fundamental group of any closed seifert fibered space modelled on @xmath21 , because it has a finite index subgroup to which proposition [ prop : euler class ] applies .
further examples of groups that satisfy theorem [ thm : into - center](3 ) can be obtained by amalgamating one such example with any group along a common central subgroup , which uses the following fact .
let @xmath118 where @xmath63 lies in the center of @xmath119 , @xmath120 .
if @xmath121 is a virtual direct factor of @xmath0 , then @xmath51 is a virtual direct factor of @xmath119 and @xmath120 .
note that @xmath122 .
if @xmath31 is a finite index subgroup of @xmath0 of the form @xmath123 , where @xmath124 is normal in @xmath31 and @xmath125 , then @xmath126 is a finite index subgroup of @xmath127 of the form @xmath128 where @xmath129 is normal in @xmath126 and @xmath130 , so @xmath51 is a virtual direct factor of @xmath127 .
in this section @xmath0 is a discrete isometry group of a hadamard manifold @xmath13 . if @xmath0 fixes a point at infinity , then @xmath0 permutes busemann functions ( or equivalently , permutes horospheres ) associated with the point .
an axial isometry @xmath131 of the upper half permutes horospheres centered at @xmath8 , as well as horospheres centered at @xmath132 .
if @xmath0 fixes a point @xmath133 $ ] at infinity represented by a busemann function @xmath134 , then the map @xmath135 is a group homomorphism @xmath136 that depends only on @xmath133 $ ] , and not on @xmath137 or @xmath134 .
this is immediate from the cocycle property of the busemann function , e.g. see the discussion after ( * ? ? ?
* proposition 3.11 ) where the above homomorphism was called the _ busemann character_. [ rmk : no hom into r ] thus if @xmath138 has no nontrivial homomorphism into @xmath10 , then @xmath0 preserves a busemann function if and only if @xmath0 fixes a point at infinity . to further enlarge the class of groups that preserve busemann functions we discuss , following schroeder in ( * ? ? ?
* appendix 3 ) and eberlein ( * ? ? ?
* section 4.4 ) , groups which we call _ grounded _ ( and which were called _ admissible _ in @xcite ) .
namely , equip the ideal boundary @xmath78 of a hadamard manifold @xmath13 with the tits metric , which is @xmath139 and complete ( * ? ? ?
* theorem ii.9.20 ) , and call a subgroup @xmath138 _ grounded _ if its fixed point set at infinity @xmath140 is a nonempty subset of ( intrinsic ) radius @xmath141 , i.e. @xmath140 lies in the closed @xmath142-ball centered at a point of @xmath140 .
there are discrete cocompact subgroups of @xmath143 with finite abelianizations , such as @xmath144 , or see @xcite for torsion - free examples .
no such group can fix a point at infinity of @xmath6 because no lattice fixes a busemann function .
by contrast , any group of translations , such as the standard @xmath27 , fixes the ideal boundary of @xmath6 pointwise , yet it is not grounded as its tits boundary is isometric to the standard unit sphere , which has radius @xmath145 .
[ thm : intro - grounded ] a discrete , infinite isometry group @xmath0 of a hadamard manifold is grounded if one of the following is true : @xmath0 is as in .
@xmath0 has a grounded normal subgroup .
@xmath0 is virtually solvable and not virtually-@xmath146 for any @xmath55 .
@xmath0 is the union of a nested sequence of grounded subgroups .
a normal abelian subgroup of @xmath0 contains an infinite @xmath0-conjugacy class .
schroeder proved in ( * ? ? ? * appendix 3 ) the key fact that an abelian group that contains a parabolic is grounded , see also ( * ? ? ? * section 4.4 ) .
these references also contain a proof of part ( 2 ) which we outline below .
if @xmath147 is a grounded normal subgroup of @xmath0 , then @xmath0 stabilizes @xmath148 , and hence fixes its center of gravity , which proves that @xmath140 is nonempty .
it is clear that @xmath149 and that @xmath140 is closed in the cone topology on @xmath78 . since the identity map of @xmath78 from the tits metric topology to the cone the topology is continuous , @xmath140 is closed in the tits metric topology .
finally , convexity of @xmath140 is implied by the fact that the @xmath142-balls are uniquely geodesic , so if an element of @xmath0 fixes the endpoints of a geodesic segment , it fixes the segment pointwise . to prove ( 4 ) let @xmath150 where @xmath151 is a countable family of grounded subgroups of @xmath152 ( since @xmath0 is countable , it is not the union of uncountable family of distinct groups ) .
note that @xmath153 . by assumptions
@xmath154 is a nested family of closed convex subsets of radius @xmath141 .
it is shown by caprace - lytchak ( * ? ? ?
* lemma 5.2 ) that any such family must have a nonempty intersection provided the ambient metric space is finite dimensional , which follows from ( * ? ? ?
* proposition 2.1 ) because @xmath13 is finite dimensional .
thus @xmath140 is a nonempty , closed , convex subset of radius @xmath141 , so @xmath0 is grounded .
now we are ready to prove ( 1 )
. the conclusion would be immediate if @xmath70 were normal in @xmath0 because by the proof of theorem [ thm : into - center ] @xmath71 contains a parabolic element , so @xmath70 , and hence @xmath0 is grounded by the above - mentioned result of schroeder . in general , since @xmath0 is countable , we can realize @xmath0 as the union of a nested sequence of finitely generated subgroups @xmath155 where @xmath156 for all @xmath157 .
we may also assume that @xmath158 contains a parabolic element @xmath72 .
the subgroup @xmath159 is normal in @xmath160 , and moreover , the index of @xmath161 in @xmath160 is finite because @xmath160 is a finitely generated , and the index of @xmath162 is finite .
hence @xmath161 contains a power of @xmath163 , which is a parabolic element in @xmath164 .
it follows that @xmath161 is grounded , and hence so is @xmath160 .
now ( 4 ) implies that @xmath0 is grounded as the union of @xmath160 s . to prove ( 3 ) we first recall a result of burger - schroeder @xcite that an amenable subgroup of @xmath152 either stabilizes a flat , or fixes a point of @xmath78 . in the former case
the subgroup is virtually-@xmath146 , which is excluded by assumption .
( if a discrete subgroup @xmath138 stabilizes a @xmath165-dimensional flat , then restricting the @xmath0-action to the flat gives a homomorphism @xmath166 with finite kernel and discrete image . by bieberbach s theorem the image of @xmath0 stabilizes an @xmath55-dimensional flat in @xmath67 , which implies that @xmath0 is virtually-@xmath146 , see e.g ( * ? ? ?
* theorem 2.1 ) ) .
thus we may assume that @xmath140 is nonempty .
by lemma [ lem : periodic group ] abelian subgroups of @xmath0 are finitely generated , so @xmath0 is virtually polycyclic as proved by malcev @xcite
. a polycyclic subgroup of @xmath152 that contains no parabolics must be virtually abelian ( * ? ? ?
* theorem ii.7.16 ) .
thus @xmath0 contains a parabolic element . if @xmath167 is the cyclic subgroup of @xmath0 generated by a parabolic element , then @xmath168 , where @xmath169 has radius @xmath141 because @xmath167 is grounded . to prove ( 5 ) consider an normal abelian subgroup @xmath63 of @xmath0 that contains an infinite @xmath0-conjugacy class .
then @xmath63 contains a parabolic element ( * ? ? ?
* lemma ii.7.17(2 ) ) , which implies that @xmath63 , and hence @xmath0 are grounded .
[ cor : elem amen ] let @xmath0 be a finitely generated , torsion - free , discrete subgroup of @xmath152
. if @xmath0 has a nontrivial , normal , elementary amenable subgroup , then either @xmath0 is grounded , or @xmath0 has a nontrivial , finitely generated , abelian , normal subgroup that is a virtual direct factor of @xmath0 .
since @xmath0 is torsion - free and discrete , it has finite cohomological dimension , so by ( * ? ? ?
* corollary 2 ) @xmath0 has a nontrivial , abelian , normal subgroup @xmath63 .
if @xmath63 contains a parabolic element , then @xmath0 is grounded .
otherwise , by lemma [ lem : periodic group ] @xmath63 is isomorphic to some @xmath146 , @xmath170 , and then ( * ? ? ?
* theorem ii.7.1(5 ) ) implies that @xmath63 is a virtual direct factor of @xmath0 .
if @xmath0 is an amenable , discrete , torsion - free isometry group of @xmath13 that is not virtually-@xmath146 for any @xmath55 , then @xmath0 fixes a point at infinity @xcite , yet i do not know whether it is always grounded . a counterexample , if it exists , would have to be an amenable group of finite cohomological dimension that is not elementary amenable , for if @xmath0 were elementary amenable , it would be virtually solvable @xcite , and hence grounded by theorem [ thm : intro - grounded](3 ) . in view of remark [ rmk : no hom into r ]
one wants to produce grounded groups with no nontrivial homomorphisms into @xmath10 .
[ ex : dekimpe 3d ] a complete list of @xmath20-dimensional closed infranilmanifolds , as found in @xcite , contains an infinite family of manifolds whose fundamental group has finite abelianization .
namely , in the family 2 on @xcite , also listed on @xcite , let @xmath171 be a positive even integer , and set @xmath172 , @xmath173 . taking products yields examples in each dimension divisible by @xmath20 .
this example does not really illustrate the power of theorem [ thm : intro - grounded](3 ) because it is already covered by theorem [ thm : into - center ] , yet it can be used as a building block in other examples below .
for that purpose we note that any group in the above family surjects onto the infinite dihedral group .
the following virtually solvable group has finite abelianization , finite cohomological dimension , and is not virtually polycyclic , so it satisfies assumptions of theorem [ thm : intro - grounded](3 ) .
[ ex : kapovich ex ] ( misha kapovich ) let @xmath23 be a virtually polycyclic group of finite cohomological dimension such that @xmath174 is a finite group that has @xmath175 among its cyclic factors for even @xmath165 ( e.g. we could take @xmath23 to be the fundamental group of hantsche - wendt s flat manifold , or any group in example [ ex : dekimpe 3d ] ) . set @xmath176 , where @xmath177 acts on @xmath178 by @xmath179 via the surjection onto @xmath180 . set @xmath181 , where @xmath23 acts on @xmath51 so that the action on @xmath178 is trivial and the action on @xmath177 is by @xmath182 ( which is indeed an action as @xmath89 is abelian ) . by construction @xmath183 is finite , and @xmath0 has finite cohomological dimension because its semidirect factors have this property , e.g. @xmath178 is locally cyclic . finally , @xmath0 is not virtually polycyclic as it contains @xmath178 , and abelian subgroups of virtually polycyclic groups are finitely generated . in the above example
@xmath0 is not of type _ f _ ( because it has @xmath178 as a virtual direct factor ) .
the following is an example of a torsion - free polycyclic group with finite abelianization which is not virtually nilpotent , and like every virtually polycyclic group , it is the fundamental group of a closed aspherical manifold @xcite .
[ ex : dekimpe poly ] ( karel dekimpe ) let @xmath23 be any virtually polycyclic group with finite abelianization such that @xmath23 surjects onto the infinite dihedral group @xmath184 .
let @xmath185 \quad\quad\quad \b=\left [ \begin{array}{cc } 0 & i_n \\ i_n & 0 \end{array}\right]\ ] ] be matrices on @xmath186 , where @xmath187 .
since @xmath188 and @xmath189 , the map that sends @xmath190 , @xmath134 to @xmath191 , @xmath192 , respectively , extends to a homomorphism @xmath193 .
it remains to choose @xmath194 so that the corresponding semidirect product @xmath195 has finite abelianization , and is not virtually nilpotent .
the abelianization of the semidirect product @xmath196 is @xmath197 , where @xmath198 is the coinvariants for the @xmath51-action on @xmath199 , i.e. the quotient of @xmath63 by the subgroup generated by the set @xmath200 , where @xmath201 is @xmath202 followed by the projection @xmath203 .
take @xmath204 with @xmath205 so that @xmath206 has nonzero determinant . by considering where @xmath191 , @xmath192 map the standard generators of @xmath207 one gets finiteness of the coivariants . to show that @xmath0 is not virtually nilpotent , we assume in addition that @xmath208 , and consider the subgroup @xmath209 that is the preimage of the subgroup @xmath210 , and note the @xmath211-summand of @xmath207 spanned by the first two basis vectors is @xmath91-invariant .
thus @xmath0 has a subgroup @xmath212 , which surjects onto @xmath213 . with our choice of @xmath194
the latter group is a lattice in sol , a @xmath20-dimensional solvable lie group . since lattices in sol
are not virtually nilpotent , neither is @xmath0 , as being virtually nilpotent is inherited by subgroups and quotients .
the following gives a large family of grounded groups of the form @xmath214 , which have finite abelianization and type _ f _ if @xmath31 does .
this applies e.g. if @xmath31 is a torsion - free , finite index subgroup of @xmath215 where @xmath216 and the semidirect product is defined via its standard action on @xmath146 .
[ prop : grounded semidirect]theorem [ thm : intro - grounded](5)_applies to every semidirect product @xmath217 defined by a homomorphism @xmath218 whose image contains an infinite order element . _ if @xmath219 is mapped to an infinite order element @xmath220 , then there is @xmath221 such that @xmath222 for @xmath223 ( e.g. @xmath224 can be chosen as a vector of the standard basis @xmath225 , for otherwise for each @xmath157 we have @xmath226 for some @xmath227 , so if @xmath228 , then by linearity @xmath229 is the identity , contradicting the infinite order assumption ) .
thus @xmath230 in @xmath0 for all @xmath223 so the conjugacy class of @xmath190 intersects @xmath146 along an infinite subset , and theorem [ thm : intro - grounded](5 ) applies . [ prop : putman ex ] in the notations of proposition [ prop : grounded semidirect ] if @xmath231 is a finite index subgroup of @xmath215 , and if @xmath232 is finite , then @xmath43 is finite . in particular
, @xmath43 is finite if @xmath202 is injective and @xmath25 . as in example
[ ex : dekimpe poly ] , the abelianization of @xmath196 is @xmath197 .
finiteness of @xmath233 follows via the elementary computation below shown to me by wilberd van der kallen .
for @xmath234 , let @xmath235 be the matrix with @xmath236 s on the diagonal , @xmath165 as @xmath237-entry , and zero elsewhere .
for every @xmath165 , and any basis vector @xmath238 one can be easily find @xmath221 and @xmath239 such that @xmath240 .
thus it remains to find @xmath165 such that @xmath231 contains every matrix @xmath239 .
let @xmath66 be a finite index normal subgroup of @xmath215 that lies in @xmath231 .
since the @xmath165th power of @xmath241 equals @xmath239 , it suffices to pick @xmath165 divisible by the index of @xmath66 in @xmath215 .
an elegant way to prove finiteness of @xmath233 was explained to me by andy putman : one needs to show that @xmath242 , and since @xmath31 is a lattice in @xmath243 , the borel density theorem shows that @xmath67 is an irreducible @xmath31-module , so the submodule generated by @xmath244 can not be proper . finally , if @xmath202 is injective and @xmath25 , then @xmath31 has kazhdan property ( t ) , and hence @xmath232 is finite .
the property of having finite abelianization is inherited by amalgamated products and extensions : the former is obvious and the latter follows from right exactness of the abelianization functor .
suppose @xmath63 is a normal abelian subgroup of groups @xmath23 and @xmath41 , so that @xmath63 is also normal in @xmath245 .
suppose @xmath23 and @xmath63 satisfy an algebraic condition that forces the existence of a parabolic in @xmath63 for every isometric , properly discontinuous @xmath23-action on a hadamard manifold , which happens e.g. if @xmath23 is as in theorem [ thm : into - center ] , or if @xmath246 as in proposition [ prop : putman ex ]
. then @xmath0 is grounded , and the above discussion allows one to construct many such groups that have finite abelianization and type _ f_.
grounded groups are much easier to find among isometry groups of visibility manifolds . the main geometric ingredient is supplied by the following .
[ lem : lem - visibi ] if @xmath13 is visibility , @xmath0 is not virtually cyclic , and @xmath140 is nonempty , then @xmath0 fixes a busemann function , @xmath0 contains no axial isometry , @xmath140 is a single point , and in particular @xmath0 is grounded .
if @xmath140 contains two distinct points , then these points can be joined by a geodesic as @xmath13 is visibility .
the union of all geodesics that join the points splits as @xmath247 , where @xmath66 is a compact convex subset .
there exists an element of @xmath0 that acts as a nontrivial translation on the @xmath10 factor ( for if @xmath0-action on the @xmath10-factor were trivial , compactness of @xmath66 and proper discontinuity of the action on @xmath247 would imply that finiteness of the group @xmath0 ) .
thus @xmath0-action on @xmath247 is cocompact , so @xmath0 is two ended and hence virtually-@xmath89 , contradicting the assumptions .
thus @xmath140 is a single point , to be denoted @xmath248 .
first we deal with the case when @xmath0 is periodic .
any elliptic element @xmath249 pointwise fixes the ray that connects @xmath248 with any point of @xmath13 fixed by @xmath250 , and hence @xmath250 preserves the busemann function for that ray , and hence every busemann function for @xmath248 as they differ by a constant .
a key property of a parabolic isometry of a visibility manifold is that is has one fixed point at infinity , and similarly , any axial isometry fixes precisely two points at infinity ( * ? ? ?
* lemma 6.8 ) .
it follows from ( * ? ? ?
* proposition 4 , page 8) that the fixed point sets of any two infinite order isometries are either disjoint or equal .
since any element of @xmath0 fixes @xmath248 , all infinite order isometries in @xmath0 have the same fixed point set .
suppose @xmath0 contains an axial isometry and no parabolic isometries .
let @xmath251 such that each axial isometry in @xmath0 fixes @xmath191 and @xmath192 . if @xmath252 is elliptic and @xmath253 is axial , then @xmath254 is axial with fixed points @xmath255 and @xmath256 , so @xmath257 .
this contradicts that above conclusion that @xmath258 .
finally , suppose @xmath0 that contains a parabolic .
then all infinite order elements of @xmath0 are parabolic with unique fixed point @xmath248 .
therefore , they fix a busemann function at @xmath248 . by the same argument as above , any elliptic element in @xmath0
also fixes the busemann function . in @xcite karlsson - noskov studied isometry groups of visibility spaces with nonempty @xmath140 .
to state their result we call a group @xmath0 _ no-@xmath259-connected _ if @xmath0 is generated by a set @xmath260 of infinite order elements such that the following graph is connected : the vertices are elements of @xmath260 , and two vertices @xmath261 are joined by an edge if and only if the subgroup generated by @xmath262 does not contain a nonabelian free subgroup . by ( * ?
* theorem 2 ) and proposition [ lem : lem - visibi ] if @xmath0 is a no-@xmath259-connected , not virtually - cyclic , discrete isometry group of a visibility manifold , then @xmath0 is grounded , and it fixes a busemann function .
the class of no-@xmath259-connected groups trivially contains each torsion - free group without free nonabelian subgroups
. also the product of any two nontrivial , torsion - free groups is no-@xmath259-connected ( here @xmath260 is the set of all elements in the product for which one of the coordinates equals the identity ) .
more examples of no-@xmath259-connected groups can be found in ( * ? ? ?
* section 8) .
we conclude that if @xmath13 visibility , then part ( 3 ) of theorem [ thm : intro - grounded ] can be improved to `` @xmath0 is amenable and not virtually cyclic '' or by `` @xmath0 does not contains a nonabelian free subgroup and is generated by the set of its infinite order elements '' , and similarly , part ( 5 ) can be improved to `` a normal abelian subgroup of @xmath0 is not cyclic '' .
in this section we discuss when an open manifold is homeomorphic to the product of @xmath10 and another manifold . for background on ends of manifolds
we refer to @xcite .
guilbault showed in @xcite that if @xmath3 is an open manifold that is homotopy equivalent to a finite complex , then @xmath263 has stable fundamental group at infinity and zero siebenmann s end obstruction , so that if @xmath264 , then @xmath263 is diffeomorphic to the interior of a compact manifold .
if the complex has codimension @xmath2 in @xmath263 , one can say more : [ thm : collar reg nbhd ] let @xmath3 be an open @xmath165-manifold that is homotopy equivalent to a finite complex of dimension @xmath265 .
if @xmath266 , then @xmath267 is diffeomorphic to the interior of a regular neighborhood of @xmath55-dimensional subcomplex .
the case @xmath268 follows from the classification of open surfaces @xcite , which implies that @xmath3 is diffeomorphic to @xmath269 .
suppose @xmath270 so that @xmath271 has dimension @xmath35 . stalling s
theorem @xcite allows to embed any codimension @xmath2 finite complex up to homotopy type , so let @xmath31 be a finite @xmath55-dimensional subcomplex onto which @xmath49 deformation retracts .
let @xmath50 be a regular neighborhood of @xmath31 .
the inclusions @xmath272 induce @xmath273-isomorphisms as @xmath31 has codimension @xmath274 .
excision in the universal cover as in the proof of ( * ? ? ?
* theorem 2.1 ) implies that @xmath275 deformation retracts onto @xmath276 .
the manifolds @xmath49 and @xmath3 are one - ended as their homology groups ( with twisted coefficients ) vanish in degrees @xmath277 ( * ? ? ?
* proposition 1.2 ) .
denote their ends by @xmath278 , @xmath279 . by (
* propositions 3.1 , 3.2 ) the end @xmath278 has arbitrary small finitely dominated @xmath8-neighborhoods and its fundamental group is stable .
let @xmath280 denote the inverse limit of the fundamental groups of the @xmath8-neighborhoods ; this group is finitely presented being a retract of the ( finitely presented ) fundamental group of a @xmath8-neighborhood .
as @xmath50 is compact , the above discussion of @xmath278 applies to the end @xmath281 of @xmath282 , and in particular , @xmath283 and @xmath280 are isomorphic via the inclusion .
if we show that the canonical map @xmath284 is an isomorphism , then siebenmann s open collar theorem ( * ? ?
* theorem 1.6 ) would imply that @xmath282 is diffeomorphic to @xmath285 , which would prove the theorem . since the inclusion @xmath286 induces a @xmath273-isomorphism
, @xmath287 is essentially the canonical map @xmath288 .
next we note that @xmath287 is onto and has perfect kernel . to see that multiply @xmath282 by a circle to make the dimension @xmath289 . by siebenmann s product formula for the end obstruction (
* ? ? ? * theorem 7.5 ) , @xmath290 is the interior of a compact manifold , which is a one - sided @xmath291-cobordism that deformation retracts onto the boundary component of the form @xmath292 .
the other boundary component @xmath51 satisfies @xmath293 .
restricting the retraction to @xmath51 gives a map @xmath294 , which is a @xmath273-surjection with perfect kernel by a standard argument involving poincar duality in the universal cover as e.g. in ( * ? ? ?
* theorem 2.5 ) . under the obvious identifications this @xmath273-surjection corresponds to the product of @xmath287 and
the identity of @xmath295 , so @xmath287 is surjective with perfect kernel .
on the other hand , it follows from ( * ? ? ?
* propositions 3.5 ) that @xmath280 is isomorphic to a ( possibly trivial ) amalgamated product of the form @xmath296 , and moreover , under the isomorphism @xmath297 is identified with the standard projection @xmath298 that extends the identity maps on each factor .
the kernel of this is a free group .
( indeed , the kernel is a normal subgroup that intersects the factors trivially , and hence intersects trivially each of their conjugates .
thus the kernel acts freely on the bass - serre tree for the amalgamated product , and hence it is a free group ) . since a free perfect group must be trivial
, @xmath287 is an isomorphism , which finishes the proof .
[ rmk : m=3 case collar ] if @xmath299 and @xmath300 is trivial or @xmath89 , which are the fundamental groups of @xmath301-complexes for which the topological surgery is known to work , then the same proof , based on the open collar theorem of @xcite , shows that @xmath263 is homeomorphic to the interior of a regular neighborhood of @xmath31 , i.e. @xmath302 , or @xmath303 , respectively .
[ rmk : prod of contr mnflds ] the product of @xmath10 with any aspherical @xmath37-manifold is covered by @xmath304 .
in fact , the references below imply that the product of any two open contractible manifolds is diffeomorphic to @xmath6 ( provided their dimensions are positive and add up to @xmath7 ) .
this is due to stallings @xcite if @xmath38 . the case of the product of an open contractible @xmath20-manifold with @xmath10 is due to luft ( * ? ? ?
* theorem 5 ) and perelman s solution of poincar conjecture .
finally , any contractible open @xmath305-manifold is diffeomorphic to @xmath269 by the classification of surfaces @xcite .
these results only give a pl homeomorphism onto @xmath6 , but a smooth structure on @xmath6 is uniquely determined by the pl structure it induces ( * ? ? ?
* corollary 6.6 ) .
[ cor : product aspherical rn ] let @xmath66 be an open aspherical @xmath306-manifold and @xmath51 be a closed manifold such that @xmath307 is homeomorphic to the product of @xmath10 and an open manifold . then the universal cover of @xmath66 is homeomorphic to @xmath308 .
we can assume @xmath309 , and that @xmath66 is contractible , because the property of being homeomorphic to the product of @xmath10 and an open manifold in inherited by any cover .
let @xmath3 be an open manifold such that @xmath307 is homeomorphic to @xmath263 . by ( * ? ? ?
* propositions 3.1 - 3.2 ) @xmath263 , and hence @xmath307 , has only one end at which @xmath273 is stable , and @xmath310 is finitely presented .
hence the same holds for the end of @xmath66 .
it follows that @xmath310 is isomorphic to @xmath311 and the projection on the first factor can be identified with the canonical map @xmath312 , which is an isomorphism by theorem [ thm : collar reg nbhd ] and remark [ rmk : m=3 case collar ] .
thus @xmath313 is trivial .
any open contractible @xmath306-manifold that is simply - connected at infinity is homeomorphic to @xmath308 .
( this is due to @xcite if @xmath314 and to @xcite if @xmath315 , while the case @xmath316 follows from @xcite and the non - existence of fake @xmath20-cells by perelman s solution of the poincar conjecture ) .
let @xmath317 be a fiber bundle whose base @xmath51 is a closed manifold with @xmath318 , and fiber @xmath41 is a compact contractible @xmath306-manifold . if @xmath319 is homeomorphic to @xmath263 for a manifold @xmath3 , then @xmath320 is homeomorphic to @xmath308 .
we can assume @xmath321 ; in particular if @xmath322 , then @xmath323 .
theorem [ thm : collar reg nbhd ] and remark [ rmk : m=3 case collar ] imply that @xmath319 is homeomorphic to the interior of a regular neighborhood of a subcomplex of codimension @xmath324 , hence the inclusion @xmath325 induces a @xmath273-isomorphism .
the same is then true for the the bundle projection @xmath326 , whose fiber is @xmath327 .
since @xmath318 , the homotopy sequence of the fibration @xmath328 implies @xmath329 , so @xmath330 is simply - connected at infinity , and hence homeomorphic to @xmath308 , as in the proof of corollary [ cor : product aspherical rn ] .
if @xmath3 is homeomorphic to @xmath6 , then it is diffeomorphic to @xmath6 when @xmath331 , and @xmath4 is diffeomorphic to @xmath47 ( * ? ? ?
* theorem 2.2 ) .
of course , @xmath332 admits a metric of @xmath5 .
conversely , suppose @xmath4 admits a metric of @xmath5 , and write it as @xmath12 where @xmath0 is infinite cyclic .
the generator of @xmath0 is either axial or parabolic . in the former case
@xmath12 is a vector bundle over @xmath333 as the normal exponential map to a totally geodesic submanifold of @xmath13 is a diffeomorphism , and the bundle is trivial because @xmath4 is orientable . in the latter case @xmath0
fixes a busemann function , so @xmath12 is the product of @xmath10 and an open manifold , and the claim follows from corollary [ cor : product aspherical rn ] .
the purpose of this section is to give a method of constructing @xmath6-covered manifold that are not homeomorphic to @xmath334 covered manifolds .
[ thm : count many not rn-1xr ] let @xmath49 be an open manifold that is covered by @xmath6 , @xmath38 , and is homotopy equivalent to a finite complex of dimension @xmath57 .
if @xmath58 is nontrivial , then the tangential homotopy type of @xmath49 contains countably many open manifolds that are covered by @xmath6 but are not homeomorphic to manifolds covered by @xmath9 .
stalling s theorem @xcite allows one to embed any codimension @xmath2 finite complex up to homotopy type , so let @xmath335 be a regular neighborhood of that subcomplex .
since the subcomplex has codimension @xmath274 , we know that @xmath336 is connected and the inclusions @xmath337 induce @xmath273-isomorphisms .
the universal cover of @xmath338 is diffeomorphic to @xmath6 because it is a regular neighborhood of a codimension @xmath2 subcomplex , which easily implies that it is simply - connected at infinity , as mentioned e.g. in the statement of ( * ? ? ?
* theorem 2.1 ) .
thus there is an embedded circle in @xmath336 that is homotopically nontrivial in @xmath147 . by replacing it with a embedded circle homotopic to the square of the original one
, we may assume its normal bundle is trivial ( and it is still homotopically nontrivial in @xmath339 since this group is torsion - free ) .
denote the circle by @xmath340 , and its closed tubular neighborhood by @xmath341 .
let @xmath66 be any compact contractible @xmath37-manifold bounded by a non - simply - connected homology sphere whose fundamental group is freely indecomposable ( see e.g. appendix [ sec : homology sph ] , or @xcite ) . fix a closed coordinate disk @xmath342 , attach @xmath343 to @xmath147 by identifying @xmath344 with @xmath341 , and denote the resulting compact manifold with boundary by @xmath345 ( the choices involved in this identification will be irrelevant for our purposes ) .
the fundamental group of the boundary of @xmath345 is the amalgamated product of @xmath346 and @xmath347 along the infinite cyclic subgroup generated by the homotopy class of @xmath340 . taking the quotient by the normal closure of @xmath346 we see that the amalgamated product surjects onto @xmath348 , so its rank is at least as large as the rank of @xmath348 . by taking boundary connected sums of @xmath66 with itself
one gets a sequence @xmath349 of compact contractible manifolds such that @xmath350 bounds the connected sum @xmath351 of @xmath157 copies of the homology sphere @xmath352 .
grushko s theorem implies @xmath353 , so @xmath354 are pairwise non - homeomorphic because their fundamental groups of infinity are pairwise non - isomorphic .
deformation retraction @xmath355 extends to a deformation retraction of @xmath356 , so @xmath357 is tangentially homotopy equivalent to @xmath49 .
next we show that @xmath357 is covered by @xmath6 .
denote the universal cover of @xmath147 and @xmath49 by @xmath358 and @xmath359 , respectively . since @xmath340 is homotopically nontrivial ,
it unravels in @xmath358 to a proper embedding @xmath360 .
let @xmath361 be the union of all lifts of @xmath362 to @xmath363 .
each component of @xmath361 is an open disk , and these disks have disjoint closures .
then @xmath364 is a noncompact manifold with boundary @xmath361 .
note that @xmath364 can be compactified to @xmath365 where @xmath361 becomes the union of a countable collection of round open disks with pairwise disjoint closures , namely , attach a collar to @xmath361 and invoke the strong version of cantrell - stalling s hyperplane linearization theorem ( * ? ? ?
* corollary 9.3 ) . on the other hand ,
the universal cover of @xmath366 is a noncompact manifold with boundary whose interior is diffeomorphic to @xmath6 @xcite and whose boundary is an open disk , so it also compactifies to @xmath365 so that the disk @xmath367 becomes a round disk .
it follows that attaching each lift of @xmath366 to @xmath368 simply amounts to attaching an open collar to each component of @xmath361 , and hence the result has interior diffeomorphic to @xmath359 , which is @xmath6 .
it remains to show that @xmath357 is not homeomorphic to a manifold covered by @xmath9
. if it is , then theorem [ thm : collar reg nbhd ] applies and the map @xmath369 is an isomorphism . hence the same is true for the inclusion @xmath370 , which on the other hand is not injective because any homotopically nontrivial loop in @xmath351 is null - homotopic in @xmath350 .
this contradiction completes the proof .
we can replace the conclusion `` countably many '' in theorem [ thm : count many not rn-1xr ] by " a continuum of under the following technical assumption .
let @xmath371 be the class of groups of the form @xmath372 where @xmath373 is an aspherical homology @xmath20-sphere that ( smoothly ) bounds a contractible manifold .
let us fix a set @xmath374 of pairwise non - isomorphic groups in @xmath371 such that any group in @xmath371 is isomorphic to a group in @xmath374 ; the set @xmath374 is countably infinite ( see appendix [ sec : homology sph ] ) , hence we write @xmath375 .
given an infinite subset @xmath376 of @xmath374 , we say that a group @xmath0 is _ no-@xmath376 _ if * one * of the following conditions ( 1 ) , ( 2 ) holds : * no group in @xmath376 can be embedded into @xmath0 .
* @xmath0 splits as a ( possibly infinite and possibly consisting of a single vertex ) graph of groups such that * * all edge groups have cohomological dimension @xmath377 , * * no vertex group is isomorphic to a group in @xmath376 . * * no vertex group splits over a trivial or an infinite cyclic subgroup . here _ splits _ stands for _ splits as an amalgamated product or an hnn - extension_. some examples of no-@xmath376 groups are given below : 1 . if @xmath0 has cohomological dimension @xmath378 , then @xmath0 satisfies ( 1 ) because groups in @xmath376 has cohomological dimension @xmath379 .
2 . if @xmath0 does not splits over a trivial or an infinite cyclic subgroup , and @xmath380 , then @xmath0 satisfies ( 2 ) for the graph of groups that consists of a single vertex .
3 . suppose @xmath376 is associated with the infinite family of @xmath351 s that are modelled on @xmath21 geometry and described in appendix [ sec : homology sph ]
. then @xmath59 and @xmath381 embed into every group in @xmath376 .
thus if @xmath59 or @xmath381 does not embed into @xmath0 , then @xmath0 is no-@xmath376 by ( 1 ) .
if @xmath0 splits as a finite graph of groups that satisfies ( a ) and ( c ) , then it can be arranged to satisfy ( b ) by removing the ( finitely many ) vertex groups from @xmath376 .
let @xmath0 be an @xmath7-dimensional poincar duality group ( e.g. the fundamental group of a closed aspherical @xmath7-manifold )
. then @xmath0 is no-@xmath376 if and only if @xmath380 .
in particular , @xmath0 is no-@xmath376 for some @xmath376 .
the case @xmath382 is trivial : @xmath0 is no-@xmath376 by ( 1 ) and @xmath380 . in the case @xmath383 a key ingredient is that any splitting of @xmath0 as an amalgamated product @xmath384 over a subgroup @xmath63 of cohomological dimension @xmath385 is trivial , i.e. @xmath0 equals a vertex subgroup ( see e.g. @xcite for a proof in the case of closed aspherical manifolds which extends to poincar duality groups ) . thus if @xmath380 , then @xmath0 satisfies ( 2 ) for the graph of group consisting of a single vertex .
conversely , arguing by contradiction suppose that @xmath0 is no-@xmath376 and @xmath386 .
then ( 1 ) is not true , and since @xmath387 admits no nontrivial splitting over a subgroup of dimension @xmath377 , the only possible graph in ( 2 ) in a single vertex which contradicts ( b ) .
if @xmath0 is the free product of all groups in @xmath374 , then @xmath0 has finite cohomological dimension , yet there is no @xmath376 for which @xmath0 is no-@xmath376 .
clearly @xmath0 is the fundamental group of a locally finite @xmath379-dimensional complex , so it has finite cohomological dimension .
write @xmath388 , assume arguing by contradiction that @xmath0 is no-@xmath376 , and fix a group @xmath389 . clearly ( 1 ) is not true , so @xmath0 has a graph of group decomposition coming from ( 2 ) .
let us compare that graph of groups decomposition with the factorization @xmath390 . since no @xmath160 splits over subgroups of dimension @xmath377 ,
bass - serre theory considerations show that @xmath391 lies in a conjugate @xmath31 of a vertex group in the graph of group , and similarly ( c ) implies that @xmath31 lies in a conjugate of some @xmath392 .
now @xmath393 by ( b ) .
hence @xmath391 fixes two vertices in the bass - serre tree for @xmath394 , which contradicts nontriviality of @xmath391 .
is it true that every group of type _ f _ is no-@xmath376 for some @xmath376 ?
[ thm : not reg nbhs ] let @xmath49 be an open aspherical @xmath7-manifold that is homotopy equivalent to a finite complex of dimension @xmath57 , where @xmath38 and @xmath58 is no-@xmath376 for some @xmath376 .
then the tangential homotopy type of @xmath49 contains a continuum of manifolds covered by @xmath6 , and not homeomorphic to manifolds covered by @xmath9 . as in the proof of theorem [ thm : count many not rn-1xr ] we can change @xmath49 within its tangential homotopy type to assume that @xmath49 is the interior of a regular neighborhood @xmath147 of a codimension @xmath2 finite subcomplex . set @xmath395 , so that @xmath396 is an isomorphism induced by the inclusion .
fix @xmath397 such that @xmath58 is no-@xmath376 .
results of kervaire , recalled in appendix [ sec : homology sph ] ensure the existence of an infinite sequence of compact contractible @xmath37-manifolds @xmath349 such that @xmath398 . set @xmath399 .
pick two disjoint closed coordinate disks @xmath400 , @xmath401 in @xmath351 .
given a subset @xmath402 , let @xmath403 be a boundary connected sum of @xmath350 s with indices in @xmath191 defined as follows : if @xmath404 with the ordering induced from @xmath405 , then we attach @xmath406 to @xmath407 by identifying @xmath408 and @xmath409 for all @xmath55 .
pick a nontrivial element in @xmath410 . by replacing the element with its square , which we denote @xmath111
, we can assume that it can be represented by an embedded circle in @xmath51 with trivial normal bundle .
attach @xmath147 to @xmath411 by identifying @xmath412 with a closed tubular neighborhood of the above embedded circle in @xmath51 via an arbitrary homeomorphism , denote the result by @xmath413 , and set @xmath414 .
since @xmath413 deformation retracts to @xmath147 , the manifolds @xmath415 and @xmath49 are tangentially homotopy equivalent .
the corresponding part of the proof of theorem [ thm : count many not rn-1xr ] implies that @xmath415 is covered by @xmath6 but is not homeomorphic to a manifold covered by @xmath9 .
it remains to show that @xmath416 fall into a continuum of homeomorphism types .
we next construct a nested sequence @xmath417 of codimension zero submanifolds that are neighborhoods of infinity in @xmath418 such that @xmath419 , the boundary @xmath420 is compact , and @xmath421 .
( any sequence of neighborhoods that satisfy the last property is called _ cofinal _ ) .
imagine a worm that eats the inside of @xmath422 except for a collar neighborhood of @xmath423 , which gives @xmath424 .
for the second helping the worm eats some more of @xmath425 leaving a smaller collar neighborhood of @xmath423 , and then eats a tunnel into @xmath426 through @xmath427 , by leaving a collar neighborhood of @xmath428 in @xmath429 , and then eats the inside of @xmath426 except for a collar neighborhood of @xmath430 , which results in @xmath431 .
the process is repeated so that the worm gets into @xmath406 on the step @xmath55 , which yields @xmath432 .
then @xmath433 , and in general @xmath434 .
now @xmath435 is a sequence of neighborhoods of infinity in @xmath436 .
let @xmath437 be a nested cofinal sequence of neighborhoods of infinity of @xmath338 such that @xmath438 is a collar neighborhood of @xmath51 .
let @xmath439 be a nested cofinal sequence of closed collar neighborhoods of @xmath440 in @xmath441 , and let @xmath442 be the complement of @xmath443 in @xmath441 .
we let the worm enlarge @xmath444 by eating through the collar neighborhoods of @xmath445 in @xmath147 and @xmath446 until @xmath447 connects to @xmath448 through a tunnel homeomorphic to @xmath449\times z_k\times s^1 $ ] ; we denote the result by @xmath450 .
thus @xmath451 is a nested , cofinal sequence of neighborhood of infinity @xmath452 in @xmath453 such that @xmath454 and the boundary @xmath455 is compact . by van kampen s theorem @xmath456 is the amalgamated product of @xmath457 and @xmath410 along the infinite cyclic subgroup generated by @xmath111 . given basepoints @xmath458 and paths @xmath459 in @xmath460 connecting @xmath461 to @xmath462
, we get an inverse sequence of homomorphisms @xmath463 induced by the inclusion @xmath464 followed by the basepoint change conjugation via @xmath459 s .
each @xmath465 has a section defined as the composition of homomorphisms induced by the following operations : deformatation retract @xmath466 onto @xmath467 inside @xmath450 , intersect the result with @xmath468 , include the intersection into @xmath468 , and change the base point via @xmath459 . except for including into @xmath468 ,
these operations induce fundamental group isomorphisms . two inverse sequence of homomorphisms @xmath469 , @xmath470
are _ pro - equivalent _ if after passing to subsequences there is a commutative diagram @xmath471 \ar[ld ] & & a_{l_3}\ar[ll ] \ar[ld ] & & \dots\ar[ll ] \ar[ld ] \\ & b_{m_1}\ar[lu ] & & b_{m_2}\ar[lu]\ar[ll ] & & b_{m_3}\ar[lu]\ar[ll ] & & \dots\ar[ll ] } \ ] ] in which the horizontal arrows are the composites of the given homomorphisms ( see e.g. @xcite for details ) .
a sequence of homomorphisms is _ semistable _ if it is pro - equivalent to a sequence of surjective homomorphisms .
since @xmath465 has a section , it is surjective , so the sequence @xmath472 is semistable , which implies ( see e.g. @xcite ) that its pro - equivalence class is independent of any of the choices of @xmath450 , @xmath461 , @xmath459 , and hence it depends only on the homeomorphism type of @xmath473 .
if we set @xmath474 , and identify @xmath475 with a subgroup of @xmath476 via the above section , then @xmath465 becomes a retraction @xmath477 .
for the rest of the proof we think of @xmath475 as an amalgamated product of subgroups isomorphic to @xmath478 along @xmath479 and refer to these subgroups as _
factors_. note that @xmath475 is the subgroup of @xmath476 generated by the above factors , and the kernel of the retraction @xmath477 equals the normal closure of the subgroup @xmath480 of the @xmath481-factor , which is mapped to @xmath475 by projection onto @xmath89 , identified with @xmath479 . to get a continuum of homeomorphism types of @xmath473 s it suffices to show that that the pro - equivalence of @xmath482 and @xmath483 implies equality of the subsets @xmath192 , @xmath484 , and by symmetry we only need to show that any @xmath485 also lies in @xmath484 .
look at a portion of the commutative diagram from the definition of pro - equivalence : @xmath486_{q_1 } \ar[ld ] & & g_{l_3}^\b\ar[ll]_{q_2 } \ar[ld]^{\psi } & \\ & g_{m_1}^\g\ar[lu ] & & g_{m_2}^\g\ar[lu]^{\phi}\ar[ll]_p & } \ ] ] with @xmath487 and @xmath488 . with the above identifications @xmath489 and @xmath490 both restrict to the identity on @xmath491 .
commutativity implies that @xmath492 and @xmath493 restricted to @xmath491 are injective , and hence so is @xmath202 restricted to @xmath494 .
let @xmath495 denote the subgroup corresponding to the @xmath496-factor .
note that @xmath194 does not split over a trivial or an infinite cyclic subgroup because @xmath497 is a closed aspherical @xmath379-manifold , and the fundamental group of a closed aspherical manifold of dimension @xmath324 can not have such splittings . since @xmath498 is an amalgamated product over @xmath479 ,
the group @xmath499 lies in a conjugate , denoted @xmath50 , of one of the factors of @xmath498 .
the map @xmath500 restricted to a conjugate of a factor is either injective or is a projection onto @xmath479 .
the latter is impossible for @xmath501 since the image contains @xmath502 .
it follows that @xmath202 restricted to @xmath50 is injective .
now @xmath50 is isomorphic either to @xmath410 or to a group in @xmath376 .
let us consider the former case , and use that @xmath503 is no-@xmath376 .
as @xmath504 , the assumption ( 1 ) gives a contradiction .
the assumption ( 2 ) implies that @xmath499 lies in a vertex group @xmath505 in a graph of groups decomposition of @xmath339 as in ( 2 ) such that @xmath505 does not split over a trivial or an infinite cyclic subgroup .
then @xmath506 lies in a conjugate , denoted @xmath260 , of one of the factors of @xmath507 , which implies @xmath508 .
it follows that @xmath509 for otherwise @xmath260 fixes a vertex ( in the bass - serre tree of the amalagamated product decomposition of @xmath510 ) that is distinct from the vertex fixed by @xmath194 , so @xmath194 would fix the path joining the two vertices , but edge stabilizers are cyclic and @xmath194 is not .
thus @xmath511 , so injectivity of @xmath202 on @xmath505 implies @xmath512 , which contradicts ( b ) . finally ,
if @xmath50 is isomorphic to a group in @xmath376 , then @xmath50 also does not split over a trivial or an infinite cyclic subgroup , so the argument of the previous paragraph gives @xmath513 , which means that a group in @xmath376 is among the factors of @xmath514 other than the @xmath410-factor , i.e. @xmath515 lies in @xmath484 .
the same methods yield the following two results that give other examples of manifolds that are covered by @xmath6 and admit no complete metric of @xmath5 .
[ thm : comp countable ] let @xmath147 be a compact @xmath7-manifold with boundary such that @xmath339 is nontrivial , and @xmath516 is covered by @xmath6 , @xmath38 .
then the tangential homotopy type of @xmath147 contains countably many compact manifolds @xmath161 such that @xmath517 is covered by @xmath6 and a component of @xmath518 is not @xmath273-injectively embedded . by assumptions @xmath339 is nontrivial , and hence infinite .
fix a component @xmath51 of @xmath336 .
the inclusion @xmath519 can not be trivial on @xmath273 .
( otherwise , passing to the universal cover we conclude that @xmath51 is homeomorphic to the boundary of a noncompact contractible manifold @xmath3 .
noncompactness of @xmath3 implies it admits no constant function with compact support , so @xmath520 vanishes but by the poincar duality @xmath521 and the latter group is isomorphic to @xmath522 by the long exact sequence of the pair .
vanishing of @xmath522 contradicts compactness of @xmath51 ) .
so there is an embedded curve @xmath523 with trivial normal bundle that is homotopically nontrivial in @xmath147 , and the proof proceeds as in theorems [ thm : count many not rn-1xr ] .
[ thm : uncount compact ] let @xmath147 be a compact manifold with boundary such that @xmath339 is nontrivial , @xmath516 is covered by @xmath6 with @xmath38 , and there is a component @xmath51 of @xmath336 such that @xmath410 is no-@xmath376 for some @xmath376 . then the tangential homotopy type of @xmath49 contains a continuum of manifolds covered by @xmath6 none of which is homeomorphic to the interior of a compact manifold . find a curve @xmath340 as in the proof of theorem [ thm : comp countable ] .
the proof proceeds as in theorems [ thm : not reg nbhs ] except that @xmath524 is the sequence of the neighborhoods of the end that corresponds to @xmath51 .
if @xmath191 is infinite , then the sequence of groups @xmath475 does not stabilize , so @xmath447 is not the interior of a compact manifold .
let @xmath64 be closed aspherical @xmath37-manifold such that any properly discontinuous , isometric @xmath525-action on a hadamard @xmath7-manifold fixes a busemann function .
let @xmath49 be complete @xmath7-manifold of @xmath5 that is homotopy equivalent to @xmath64 .
then @xmath49 is the total space of a real line bundle over @xmath64 , and in particular , @xmath49 is the interior of a compact manifold @xmath147 such that no homotopically nontrivial loop in @xmath336 is null - homotopic in @xmath147 . now theorems
[ thm : comp countable ] and [ thm : uncount compact ] applied to this @xmath147 give infinitely many manifolds in the tangential homotopy type of @xmath49 that are covered by @xmath6 and admit no complete metric of @xmath5 .
this argument applies when * @xmath64 is an infranilmanifold ( see example [ intro - ex : flat ] and proposition [ prop : nil not virt split ] ) ; * a finite cover of @xmath64 is the total space of a circle bundle with nonzero rational euler class ( see proposition [ prop : euler class ] ) ; * @xmath526 is a higher rank , irreducible lattice as in example [ intro - ex : higher rank ] ; * @xmath526 a polycyclic group as in example [ ex : dekimpe poly ] .
a group @xmath0 _ has type _ f _ _ if there is a finite cw - complex that is @xmath1 .
our main results apply to groups of type _ f _ , or more generally , groups of finite cohomological dimension .
this appendix contains references justifying that various groups constructed in sections [ sec : center]-[sec : grounded ] have these properties ( provided they are torsion - free ) .
we refer to @xcite for properties of cohomological and geometric dimensions of a group , of which we mostly use the former , denoted @xmath527 .
finiteness conditions on groups , such as _ f _ and _ fp _ , are reviewed in ( * ? ? ?
* appendix f ) .
the following facts are known : 1 .
finiteness of cohomological dimension is inherited by extensions and amalgamated products ( * ? ? ?
* proposition viii.2.4 ) .
2 . if @xmath119 , @xmath120 , @xmath63 have type @xmath41 , then so does @xmath384 , cf .
* theorem ii.7.3 ) .
3 . the property of having type _ f _ or _ fp _ is inherited by extensions .
( combine ( * ? ? ? * exercise viii.6.8 ) with the fact that finite presentability of a group inherited by extensions @xcite ) 4 .
the fundamental group of an aspherical @xmath7-manifold has cohomological dimension @xmath528 with equality precisely when the manifold is closed ( * ? ? ?
* viii.8.1 ) .
more generally , an infinite index subgroup of the @xmath7-dimensional poincar duality group has cohomological dimension @xmath529 @xcite .
5 . @xmath0 is a countable group of finite cohomological dimension if and only if there is a @xmath1 manifold whose universal cover is a euclidean space .
( find a finite - dimensional @xmath1 ( * ? ? ? * viii.7.1 ) , make it a locally finite and of the same dimension ( * ? ? ?
* theorem 13 ) , use simplicial approximation to make the space polyhedral , embed it into a euclidean space ( * ? ? ?
* theorem 3.2.9 ) , take the open regular neighborhood , and note that the universal cover of any codimension @xmath2 open regular neighborhood is simply - connected at infinity , as mentioned e.g. in the statement of ( * ? ? ?
* theorem 2.1 ) .
alternatively , one can take the product of the open regular neighborhood with @xmath10 , see remark [ rmk : prod of contr mnflds ] . )
results of section [ sec : without r - factors ] require a certain infinite family of compact contractible manifolds whose construction is reviewed below . the problem which homology @xmath20-spheres bound contractible manifolds is not well understood .
( freedman famously showed that all homology @xmath20-spheres bound topological contractible @xmath379-manifolds , but in this paper we only consider smooth manifolds ) . an infinite family of homology @xmath20-spheres that bound contractible manifolds was found by casson - harer @xcite among brieskorn spheres @xmath530 with @xmath531 , which are closed aspherical @xmath20-manifolds modelled on @xmath21 geometry @xcite . for example , casson - harer s list contains the family @xmath532 where @xmath500 is any odd integer @xmath2 .
the groups @xmath533 are pairwise non - isomorphic because their quotients by the centers is the fundamental groups of @xmath305-orbifolds with euler characteristics @xmath534 , which are all distinct .
one can also construct an infinite family of hyperbolic homology @xmath20-spheres that bound contractible manifolds , see the mathoverflow answers @xcite for details , but the above casson - harer s family is enough for our purposes .
kervaire shows in ( * ? ? ?
* theorems 1 and 3 ) that for each @xmath535 the fundamental group of any homology @xmath20-sphere is also the fundamental group of a homology @xmath165-sphere that bounds a contractible manifold .
( even though ( * ? ? ?
* theorems 1 ) is not stated for @xmath536 , it still applies in our case because as kervaire remarks on @xcite the fundamental group of any homology @xmath20-sphere has a presentations with an equal number of relations and generators ) .
i. agol and b. martelli , _
irreducible homology 3-spheres that bound smooth contractible manifolds _ , http://mathoverflow.net/questions/104451/irreducible-homology-3-spheres-that-bound-smooth-contractible-manifolds .
m. r. bridson and a. haefliger , _ metric spaces of non - positive curvature _ , grundlehren der mathematischen wissenschaften [ fundamental principles of mathematical sciences ] , vol .
319 , springer - verlag , berlin , 1999 .
m. i. kargapolov and ju .
i. merzljakov , _ fundamentals of the theory of groups _ , graduate texts in mathematics , vol . 62 ,
springer - verlag , new york , 1979 , translated from the second russian edition by robert g. burns .
j. milnor , _ on the @xmath20-dimensional brieskorn manifolds @xmath539 _ , knots , groups , and 3-manifolds ( papers dedicated to the memory of r. h. fox ) , princeton univ . press ,
princeton , n. j. , 1975 , pp .
175225 . ann . of math .
studies , no . 84 .
c. t. c. wall , _ surgery on compact manifolds _
, second ed .
, mathematical surveys and monographs , vol .
69 , american mathematical society , providence , ri , 1999 , edited and with a foreword by a. a. ranicki . | we study algebraic conditions on a group @xmath0 under which every properly discontinuous , isometric @xmath0-action on a hadamard manifold has a @xmath0-invariant busemann function , and for such @xmath0 we prove the following : ( 1 ) every open complete nonpositively curved riemannian @xmath1 manifold that is homotopy equivalent to a finite complex of codimension @xmath2 is an open regular neighborhood of a subcomplex of the same codimension ; ( 2 ) each tangential homotopy type contains infinitely many open @xmath1 manifolds that admit no complete nonpositively curved metric even though their universal cover is the euclidean space .
a sample application is that an open contractible manifold @xmath3 is homeomorphic to a euclidean space if and only if @xmath4 admits a complete riemannian metric of nonpositive curvature . |
a complex - valued harmonic function @xmath3 defined on the unit disk @xmath0 is called a _ harmonic mapping _ if it maps @xmath0 univalently onto a domain @xmath4 .
note that it is not required that the real and the imaginary part of @xmath3 satisfy the cauchy - riemann equations . in 1984 ,
clunie and sheil - small @xcite showed that many classical results for conformal mappings have natural analogues for harmonic mappings , and hence they can be regarded as a generalization of conformal mappings .
each harmonic mapping in @xmath0 has a canonical presentation @xmath5 , where @xmath6 and @xmath7 are analytic in @xmath0 and @xmath8 .
a harmonic mapping @xmath5 is called _ sense - preserving _ if the jacobian @xmath9 is positive in @xmath0 .
then @xmath3 has an _ analytic dilatation _
@xmath10 such that @xmath11 for @xmath12 . for basic properties of harmonic mappings
we refer to @xcite .
a domain @xmath4 is said to be _ convex in the horizontal direction _ ( chd ) if its intersection with each horizontal line is connected ( or empty ) .
a univalent harmonic mapping is called a chd mapping if its range is a chd domain .
construction of a harmonic mapping @xmath3 with prescribed dilatation @xmath1 can be done effectively by the _ shear construction _ the devised by clunie and sheil - small @xcite .
let @xmath13 be a harmonic and locally univalent in the unit disk @xmath0 .
then @xmath3 is univalent in @xmath0 and its range is a chd domain if and only if @xmath14 is a conformal mapping of @xmath0 onto a chd domain .
suppose that @xmath15 is a chd conformal mapping .
for a given dilatation @xmath1 , the harmonic shear @xmath5 of @xmath15 is obtained by solving the differential equations @xmath16 from the above equations , we obtain @xmath17 for the anti - analytic part @xmath7 , we have @xmath18 observe that @xmath19 - \overline{\varphi(z)}.\ ] ] we shall use ( [ eqn : fun - h ] ) to find the analytic part @xmath6 of the harmonic mapping @xmath5
. then the anti - analytic part @xmath7 of the harmonic mapping @xmath3 can be obtained from the identity @xmath20 , or computed via ( [ eqn : fun - f ] ) .
it is known that the class of harmonic mappings has a close connection with the theory of minimal surfaces . in the space @xmath21 , the _ minimal surface _ is a surface which minimizes the area with a fixed curve as its boundary .
this minimization problem is called _ plateau s
problem_. discussion concerning the differential geometric approach to the subject can be found from the book by pressley @xcite .
our results concerning minimal surfaces are based on the weierstrass - enneper representation .
let @xmath22 be a non - parametric minimal surface over a simply connected domain @xmath23 in @xmath24 given by @xmath25 where @xmath26 identifies the complex plane @xmath27 , which underlies the domain of @xmath28 .
the following result is known as the weierstrass - enneper representation .
this representation provides a link between harmonic univalent mappings and minimal surfaces .
the surface @xmath22 is minimal surfaces if and only if @xmath22 has the following representation @xmath29 where @xmath30 are analytic such that @xmath31 , and @xmath32 is a sense - preserving univalent harmonic mapping from @xmath0 onto @xmath23 . in this case , the surface @xmath22 is called a minimal surface over @xmath23 with the projection @xmath33 .
further information about the relation between harmonic mappings and minimal surfaces can be found from the book of duren @xcite .
systematical construction of harmonic shears of mappings of the unit disk and unbounded strip domains , and their boundary behaviour are presented in the article by greiner @xcite . in most cases
the dilatation is chosen to be @xmath34 . in this paper
, we study two classes of conformal mappings , each of which map @xmath0 univalently onto a domain which is convex in the horizontal direction .
the first one involves the mapping @xmath35 which maps @xmath0 onto @xmath24 minus four symmetric half - lines . in @xcite , ganczar and
widomski have studied some special cases of this mapping and its harmonic shears .
analytic examples of harmonic shears of @xmath15 with dilatations @xmath36 along with illustrations , are given in @xcite .
the second case is related to the conformal mapping ( see @xcite ) @xmath37 which maps the unit disk @xmath0 onto a regular @xmath2-gon . in @xcite , driver and
duren discussed the harmonic shear of @xmath15 by the choice of the dilatations @xmath38 , @xmath39 and @xmath40 . with the last choice of the dilatation
, it is assumed that @xmath2 is even .
the minimal surfaces of these harmonic shears were studied as well .
the outline of the paper is as follows : in the first part , we study the four - slit conformal mapping , and its harmonic shears , with the dilatation @xmath34
. then we consider conformal mappings of regular @xmath2-gons , with dilatations @xmath41 and @xmath42 .
these dilatations lead naturally to the appel s hypergeometric functions , which are formal extensions of the gaussian hypergeometric functions into two variables .
finally , we consider minimal surfaces obtained by shearing conformal mapping of the regular @xmath2-gon described by .
the results are also illustrated by using mathematica .
in this section we shall give examples of harmonic shear of unbounded conformal mappings with a suitable dilatations .
for @xmath43 and @xmath44 $ ] , let us consider the function @xmath15 defined by @xmath45 recall from @xcite that @xmath15 is univalent and it maps the unit disk @xmath0 onto a domain convex in the direction of the real axis .
in special cases , namely @xmath46 and @xmath47 , the image of the unit disk @xmath0 under @xmath15 is @xmath48\right\},\ ] ] and @xmath49 respectively . in the case
@xmath50 , the mapping @xmath15 maps the unit disk @xmath0 onto @xmath24 minus the following half - lines : @xmath51\right\},\ ] ] and @xmath52 by writing @xmath53 , @xmath54 , the equation ( [ eqn : fun - slit ] ) takes the form @xmath55 also in @xcite , the authors considered harmonic shears of @xmath15 for some choices of @xmath56 and @xmath57 with dilatations @xmath58 let us now consider the harmonic shear of @xmath15 defined in ( [ eqn : fun - slit-2 ] ) with the dilatation @xmath34 , @xmath59 .
first , we need the derivative of @xmath15 and the direct calculation gives @xmath60.\ ] ] thus by ( [ eqn : fun - h ] ) and ( [ eqn : fun - slit - der ] ) , we have @xmath61 . \end{split}\ ] ] we shall write this in the form @xmath62 where @xmath63 by partial fractions and @xmath64 , @xmath65 , we have @xmath66 therefore the first integral @xmath67 can be rewritten as @xmath68 we remark that for the case @xmath69 , the latter part of the integral inside the summation sign should be omitted . for the first integral
, we get @xmath70 the latter integral @xmath71 depends on whether @xmath2 is odd or even . assuming that @xmath2 is odd
, we easily see that @xmath72 thus @xmath73 note that , by assumption , @xmath74 . to simplify our notation ,
let @xmath75 be an index set .
suppose that @xmath76 , then we define @xmath77 , and @xmath78 . with this notation , in case @xmath2 is even , we have @xmath79 where @xmath80 next we compute the integrals @xmath81 and @xmath82 .
assuming that @xmath83 , i.e , @xmath84 , @xmath65 , we compute @xmath85 \\ & = -\frac{1}{n } \sum_{k=0}^{n-1 } \left\ { \frac{1}{\eta - z_k } \left ( \frac{1}{\eta - z } - \frac{1}{\eta } \right ) - \frac{1}{(\eta - z_k)^2 } \left [ \log \left ( \frac{\eta - z}{\eta } \right ) - \log \left ( \frac{z_k - z}{z_k }
\right ) \right ] \right\}.\end{aligned}\ ] ] in the case of @xmath86 , for @xmath87 , we have @xmath88 the sum can be computed as above and the last integral is @xmath89.\ ] ] note that , we have an identity @xmath90 and if @xmath91 , we have @xmath92 .
finally , the function @xmath7 can be solved readily from the following identity @xmath93 in table [ table : slit - map - i1 ] we have the integral @xmath67 of @xmath6 , which depends on whether @xmath2 is odd or even . on the other hand , integrals @xmath81 and @xmath82
depend on whether @xmath94 or @xmath86 , for an @xmath87 and the result is given in table [ table : slit - map - i2 ] . .
.analytic part @xmath95 of the harmonic shear @xmath3 of the slit mapping @xmath15 with a dilatation @xmath96 .
note that in the case of @xmath69 , the summation should be omitted .
see table [ table : slit - map - i2 ] for the integrals @xmath81 and @xmath82 . [ cols="^,^",options="header " , ] therefore , we have obtained the following result : let @xmath15 be given by .
then the harmonic shear @xmath97 , where @xmath6 is given in table [ table : slit - map - i1 ] and table [ table : slit - map - i2 ] and the anti - analytic part @xmath20 , maps the unit disk @xmath0 univalently onto a domain which is convex in the horizontal direction . in the case of @xmath98 , @xmath99 and @xmath50
, we have the corollary .
[ cor : nehari ] let @xmath100 be defined as @xmath101 then @xmath100 maps the unit disk @xmath0 univalently onto @xmath24 minus the following half - lines @xmath102\right\},\ ] ] and @xmath103 which is convex in the horizontal direction . then the harmonic shear @xmath104 , with @xmath105 where @xmath106 are given in table [ table : slit - map - i1 ] and table [ table : slit - map - i2 ] , and @xmath107 , maps the unit disk @xmath0 onto a chd domain .
the above conformal mapping @xmath100 is given in @xcite .
in figure [ fig : conf - nehari ] , we have the conformal mapping @xmath100 of the unit disk @xmath0 onto a four - slit domain with @xmath108 .
harmonic shears of this mapping as given in corollary [ cor : nehari ] are shown in figure [ fig : nehari - pi3 ] with @xmath109 , for @xmath110 .
of the unit disk @xmath0 onto a four - slit domain with @xmath109 , and the omitted half - lines are dashed .
illustrations of harmonic shears are given in figure [ fig : nehari - pi3 ] .
, scaledwidth=50.0% ]
as pointed out in the introduction , the authors in @xcite considered the conformal mapping @xmath15 given by .
then the harmonic shear @xmath3 ( with dilatations @xmath111 , @xmath112 , and @xmath39 ) can be given in terms of the gaussian hypergeometric function , which is defined by @xmath113 where @xmath114 is the pochhammer symbol .
for @xmath115 , this can also be written as the euler integral @xmath116 now we consider the dilatation @xmath41 .
therefore , by ( [ eqn : fun - h ] ) , ( [ eqn : fun - g ] ) and ( [ eqn : n - gonmap ] ) , we have @xmath117 notice that we can write @xmath6 and @xmath7 as @xmath118 using the change of variable @xmath119 , we obtain @xmath120 in order to rewrite @xmath6 in a compact form for our purpose , we recall the first appel s hypergeometric function @xcite , which is defined by @xmath121 where @xmath122 is the pochhammer symbol given above . as for hypergeometric functions , appel s hypergeometric functions can be defined by euler s integral as follows @xcite : @xmath123 where @xmath124 . set @xmath125 .
we have @xmath126 a direct computation shows that @xmath7 defined by ( [ eqn : fun - hg - w2 ] ) can be written in the form @xmath127 this proves the following result .
[ thm : polygon-2n ] let @xmath15 be given by . then the harmonic shear @xmath97 , where @xmath6 and @xmath7 are given in and , respectively , maps the unit disk @xmath0 univalently onto a domain which is convex in the horizontal direction . in figure
[ fig : polygon - double ] , we have illustrations of the conformal mapping @xmath15 onto regular @xmath2-gon and the harmonic shear @xmath128 with dilatation @xmath41 , for @xmath129 .
it is worth remarking that this situation was not considered by duren and driver in @xcite . onto a regular @xmath2-gon and
its harmonic shears with the dilatation @xmath41 , @xmath129 for a comparison with a dilatation @xmath40 , see @xcite .
, title="fig:",scaledwidth=43.0% ] onto a regular @xmath2-gon and its harmonic shears with the dilatation @xmath41 , @xmath129 for a comparison with a dilatation @xmath40 , see @xcite .
, title="fig:",scaledwidth=43.0% ] onto a regular @xmath2-gon and its harmonic shears with the dilatation @xmath41 , @xmath129 for a comparison with a dilatation @xmath40 , see @xcite .
, title="fig:",scaledwidth=43.0% ] onto a regular @xmath2-gon and its harmonic shears with the dilatation @xmath41 , @xmath129 for a comparison with a dilatation @xmath40 , see @xcite .
, title="fig:",scaledwidth=43.0% ] onto a regular @xmath2-gon and its harmonic shears with the dilatation @xmath41 , @xmath129 for a comparison with a dilatation @xmath40 , see @xcite .
, title="fig:",scaledwidth=43.0% ] onto a regular @xmath2-gon and its harmonic shears with the dilatation @xmath41 , @xmath129 for a comparison with a dilatation @xmath40 , see @xcite . ,
title="fig:",scaledwidth=43.0% ] next we let @xmath2 be odd and we consider the dilatation @xmath42 . thus , by ( [ eqn : fun - h ] ) , ( [ eqn : fun - g ] ) and ( [ eqn : n - gonmap ] ) , we have @xmath130 since @xmath2 is assumed to be odd , we have @xmath131 and as a consequence of this observation , @xmath6 defined above takes the form @xmath132 a similar expression holds for @xmath7 as well . finally , by computation
, we obtain that @xmath133 we have shown the following result : [ thm : polygon - z2 ] let @xmath15 be given by . then the harmonic shear @xmath97 , where @xmath6 and @xmath7 are given in , maps the unit disk @xmath0 univalently onto a domain which is convex in the horizontal direction . in figure [
fig : polygon - square ] , illustrations of the conformal mapping @xmath15 onto a regular @xmath2-gon and the harmonic shear @xmath3 with dilatation @xmath42 , for @xmath134 , are given .
a harmonic function @xmath97 can be lifted to a minimal surface if and only if the dilatation @xmath1 is the square of an analytic function .
suppose that @xmath135 for some analytic function @xmath136 in the unit disk @xmath0 .
then the corresponding minimal surfaces has the form @xmath137 where @xmath138 let @xmath15 be a conformal mapping , which maps the unit disk @xmath0 onto a regular @xmath2-gon and let the dilatation be @xmath41 .
then the minimal surfaces is determined by the integral @xmath139 by the substitution @xmath140 , we have @xmath141 appel s hypergeometric presentation gives @xmath142 in figure [ fig : minimal-2n ] , we have the minimal surfaces for the above mapping for @xmath143 . in the second polygonal example , let @xmath2 be odd and let the dilatation be @xmath144 . in this case the minimal surface lifting is given by integral @xmath145 again the above integral can be written by appel s hypergeometric function @xmath146 in figure [ fig : minimal - z2 ] , minimal surfaces of , for @xmath134 , are illustrated . | in this paper , we study harmonic mappings by using the _ shear construction _ , introduced by clunie and sheil - small in 1984 .
we consider two classes of conformal mappings , each of which maps the unit disk @xmath0 univalently onto a domain which is convex in the horizontal direction , and shear these mappings with suitable dilatations @xmath1 .
mappings of the first class map the unit disk @xmath0 onto four - slit domains and mappings of the second class take @xmath0 onto regular @xmath2-gons .
in addition , we discuss the minimal surfaces associated with such harmonic mappings .
furthermore , illustrations of mappings and associated minimal surfaces are given by using mathematica .
= 1 the author was supported by a grant ( ma2011n25 ) from the magnus ehrnrooth foundation . ] |
we consider the free motion of an incompressible viscous fluid in a two - dimensional exterior domain @xmath1 , where @xmath2 is a compact obstacle with a smooth boundary .
we do not assume that @xmath3 is connected , hence we include the case where the fluid moves around a finite collection of obstacles . as for the boundary conditions ,
we suppose that the velocity of the fluid vanishes on @xmath4 and decays to zero at infinity .
the evolution of our system is thus governed by the navier - stokes equations @xmath5 where @xmath6 and @xmath7 denote , respectively , the velocity and the pressure of the fluid at a space - time point @xmath8 . as can be seen from the first equation in , we assume that the fluid density is constant and that the kinematic viscosity is equal to @xmath9 .
since includes no forcing , the motion of the fluid originates entirely from the initial data @xmath10 , which we assume to be divergence - free and tangent to the boundary on @xmath4 .
the behavior of the solutions of depends in a crucial way on the decay rate of the velocity field @xmath11 as @xmath12
. if the initial velocity @xmath13 belongs to the energy space @xmath14 where @xmath15 denotes the interior unit normal on @xmath16 , we have the following classical result : [ thm1 ] for all initial data @xmath17 , eq . has a unique global solution @xmath18 which satisfies for all @xmath19 the energy equality : @xmath20 global well - posedness for the navier - stokes equations was first established by leray @xcite in the particular case where @xmath21 .
when @xmath22 is bounded , the first results also go back to leray @xcite , but global existence of large solutions was shown only later by ladyzhenskaia @xcite , see also @xcite . to prove theorem [ thm1 ] , one can use a regularization or a discretization procedure to construct global _ weak _ solutions of which satisfy the energy inequality , and then prove that these solutions are unique and have the desired regularity .
alternatively , one can construct local _ mild _ solutions by transforming into an integral equation and solving it by a fixed point argument , and then use the energy equality to show that all solutions can be extended to the whole time interval @xmath23 .
although most of the literature is devoted to the situation where @xmath24 is either a bounded domain or the whole plane @xmath25 , the case of an exterior domain can be treated without essential modifications , see e.g. @xcite .
it follows from that the kinetic energy @xmath26 is nonincreasing in time , and a result of masuda @xcite shows that @xmath27 converges to zero as @xmath28 .
moreover , under additional assumptions on the initial data , it is possible to specify a decay rate in time .
for instance , if @xmath29 for some @xmath30 , the solution of lies in the same space for all @xmath31 and @xmath32 see @xcite .
it is interesting to notice that fails in the limiting case @xmath33 . indeed ,
if @xmath34 , then in general the velocity field @xmath11 decays like @xmath35 as @xmath12 , so that @xmath36 for @xmath31
. as an aside , we mention that this loss of spatial decay is related to the net force @xmath37 exerted by the fluid on the obstacle @xmath3 . to see this
, we first observe that any velocity field @xmath38 satisfies @xmath39 . indeed , if @xmath40 is smooth and compactly supported , then using gauss theorem and the fact that @xmath41 on @xmath4 we find @xmath42 the general easily case follows by a density argument @xcite . now , if @xmath43,l^1(\omega)^2)$ ] is a solution of the navier - stokes equation , then @xmath44 because @xmath45 where @xmath46 is the stress tensor ( we recall that all physical parameters have been normalized to @xmath9 ) .
the formal calculation above can be made rigorous @xcite and shows that , no matter how localized the initial data may be , the velocity field @xmath47 does not stay integrable for positive times , unless the net force @xmath37 vanishes identically .
of course this is not the case in general , but in highly symmetric situations it is possible to construct solutions of for which @xmath48 , and which decay faster as @xmath0 than what is indicated in , see @xcite .
much less is known about the solutions of if the initial data @xmath13 are not square integrable .
although the physical relevance of infinite - energy solutions can be questioned , we believe that such solutions naturally occur when studying the dynamics of in a two - dimensional exterior domain @xmath24 . one way to realize that is to consider the relation between the velocity field @xmath40 and the associated vorticity @xmath49 . given @xmath50 , we denote @xmath51 in other words @xmath52 is the completion with respect to the norm @xmath53 of the space of all smooth , divergence - free vector fields with compact support in @xmath24 , see @xcite .
we then have the following result : [ bs0 ] if @xmath54 for some @xmath50 , and if @xmath55 , then @xmath56 for almost every @xmath57 . here
@xmath58 and @xmath59 if @xmath60 .
the proof of lemma [ bs0 ] is very simple : if @xmath61 denotes the extension of @xmath40 by zero outside @xmath24 , then @xmath62 , @xmath63 , and @xmath64 .
moreover @xmath65 is the extension of @xmath66 by zero outside @xmath24 .
thus @xmath67 can be expressed in terms of @xmath68 using the classical biot - savart law in @xmath25 , and restricting that relation to @xmath24 we obtain .
we emphasize that the representation is only valid if @xmath66 is the curl of a divergence - free velocity field @xmath40 which _ vanishes on _ @xmath4 .
in contrast , if @xmath69 is an arbitrary smooth function with compact support , the velocity field _ defined by _ does not even satisfy @xmath41 on @xmath70 !
we now assume that the vorticity distribution @xmath66 is sufficiently localized so that @xmath71 , and we define the _ total circulation _
@xmath72 where the second equality follows from green s theorem , since @xmath73 and @xmath40 vanishes on @xmath74 . using the vorticity formulation of the navier - stokes equation , it is not difficult to verify that the total circulation is a conserved quantity .
but it follows from that @xmath75 hence @xmath76 as soon as @xmath77 .
this shows that finite - energy solutions of necessarily have zero total circulation .
on the other hand , in many important examples of two - dimensional flows such as vortex patches , vortex sheets , or point vortices , the vorticity distribution typically has a constant sign , hence the total circulation is necessarily nonzero . in our opinion , it is thus important to enlarge the class of admissible solutions of , so as to allow for velocity fields which decay like @xmath78 as @xmath12 .
a possible framework for the study of infinite - energy solutions of the navier - stokes equation is the _ weak energy space _
@xmath79 where @xmath80 is the weak @xmath81 space on @xmath24 , see @xcite .
we recall that @xmath82 in the sense that the norm @xmath83 is equivalent to the quantity in the right - hand side of . clearly @xmath84 , but the weak energy space is large enough to include velocity fields which decay slowly at infinity , as in . concerning the solvability of in @xmath85 ,
the following general result was obtained by kozono and yamazaki : [ thm3 ] @xcite there exists @xmath86 such that , for all initial data @xmath87 satisfying @xmath88 eq . has a unique global solution such that , for all @xmath89 , @xmath90 and such that @xmath91 as @xmath92 in the weak-@xmath93 topology of @xmath94 .
theorem [ thm3 ] shows that the cauchy problem for the navier - stokes equations is globally well - posed in the weak energy space @xmath85 , provided that the local singularity of the initial data @xmath13 is sufficiently small , in the sense of . to illustrate the meaning of this smallness condition
, we consider the simple situation where the initial flow is just a point vortex of circulation @xmath95 located at @xmath96 .
the initial vorticity is thus given by @xmath97 , and using the classical biot - savart law in the exterior domain @xmath24 ( see e.g. @xcite ) it is easy to verify that the corresponding velocity field @xmath13 lies in @xmath85 , is smooth in @xmath98 , and satisfies @xmath99 so that is fulfilled if and only if @xmath100 .
this example shows that , if the initial vorticity @xmath101 is a finite measure , condition implies a restriction on the size of the atomic part of @xmath101 .
such a restriction also arises in the analysis of the two - dimensional vorticity equation in the whole space @xmath25 , see @xcite , but in that particular case the uniqueness of the solution can be established when the initial vorticity is an arbitrary finite measure @xcite .
although theorem [ thm3 ] provides the existence of a large class of infinite - energy solutions , very little is known about the asymptotic behavior of these solutions as @xmath0 .
in fact , we do not even know whether they stay bounded in the weak energy space @xmath85 , because we are lacking a priori estimates .
indeed , if @xmath102 the energy equality does not make sense , and because of the no - slip boundary condition on @xmath4 it is quite difficult to obtain estimates on the vorticity distribution if @xmath103 . in the rest of this paper , however , we consider a particular class of infinite - energy solutions of the navier - stokes equations , for which the asymptotic behavior in time can be accurately described .
in the particular case where @xmath21 , the navier - stokes equations have a family of self - similar solutions of the form @xmath104 , @xmath105 , where @xmath95 is a free parameter ( the total circulation ) and @xmath106 these solutions are usually called the _ lamb - oseen vortices_. if @xmath104 , the corresponding vorticity distribution is @xmath107 , where @xmath108 note that @xmath109 and @xmath110 for all @xmath19 .
oseen vortices play an important role in the dynamics of the navier - stokes equations in @xmath25 . in particular , we have the following result : [ thm4 ] @xcite .
for all initial data @xmath111 such that the vorticity distribution @xmath101 is integrable , the solution of the navier - stokes equation in @xmath25 satisfies @xmath112{}\ , 0~ , \qquad \hbox{where } \quad \alpha \,=\ , \int_{{{\mathbb r}}^2 } \omega_0 { \,{\rm d}}x~.\ ] ] in other words , oseen vortices describe the leading order asymptotics of all solutions of the navier - stokes equations in @xmath25 with integrable vorticity distribution and nonzero total circulation . in the case of an exterior domain @xmath113 ,
approximate oseen vortices can be constructed in the following way .
let @xmath114 $ ] be a smooth , radially symmetric cut - off function such that @xmath115 is nondecreasing along rays , @xmath116 on a neighborhood of @xmath3 , and @xmath117 when @xmath118 is sufficiently large .
the _ truncated oseen vortex _ with unit circulation is defined as follows : @xmath119 for any @xmath19 , it is clear that @xmath120 is a smooth divergence - free vector field which vanishes in a neighborhood of @xmath3 .
the corresponding vorticity distribution @xmath121 has the explicit expression @xmath122 where @xmath123 is defined in .
in particular @xmath124 and @xmath125 for all @xmath19 .
moreover , a direct calculation shows that @xmath126 hence there exists a radially symmetric function @xmath127 such that @xmath128 .
now , given @xmath95 , we consider solutions of of the particular form @xmath129 where @xmath130 is the truncated oseen vortex defined in , and @xmath131 is a _ finite - energy _ perturbation . in this situation , we expect that @xmath132 converges to zero in energy norm as @xmath0 , so that the long - time behavior of @xmath47 is described , to leading order , by the oseen vortex @xmath133 .
our main result , which was obtained in collaboration with y. maekawa , shows that this is indeed the case , provided the total circulation @xmath134 is sufficiently small .
[ main ] @xcite fix @xmath30 , and let @xmath135 .
there exists a constant @xmath136 such that , for any smooth exterior domain @xmath22 and for all initial data of the form @xmath137 with @xmath138 and @xmath139 , the solution of the navier - stokes equations satisfies @xmath140 as @xmath0 . to understand the scope and the limitations of this statement , a few comments are in order .
first of all , theorem [ main ] is a _ global stability result _ for oseen vortices with small circulation at infinity , because we do not impose any restriction on the size of the perturbation @xmath139 . in the particular case
where @xmath141 , there is no vortex at all and we just recover the asymptotics with @xmath142 instead of @xmath143 in the right - hand side . also , in the simple situation where @xmath21 , our result is comparable to that of carpio @xcite , although the proof is very different .
the main limitation of theorem [ main ] is of course the restriction on the size of the circulation @xmath134 , which we believe is purely technical . in this respect
, the fact that @xmath144 can be taken independent of the domain @xmath24 is quite significant , because we know that there is no restriction on the circulation in the particular case where @xmath21 , see theorem [ thm4 ] .
obviously @xmath144 is a decreasing function of @xmath145 , and the proof shows that @xmath146 as @xmath147 . thus the limiting case @xmath148 is not included , which means that we are not able to control arbitrary finite - energy perturbations of the oseen vortex ( see however @xcite for a partial result in that direction ) . on the other hand the limit of @xmath144 as @xmath149 can be estimated and is found to be approximately @xmath150 , see @xcite .
we also mention that the decomposition @xmath151 of the initial data is automatically satisfied if we assume that the initial vorticity is sufficiently localized .
more precisely , we have the following auxiliary result , which follows quite easily from lemma [ bs0 ] . [ vort1 ]
@xcite given @xmath152 , assume that @xmath153 for some @xmath154 and that the associated vorticity @xmath155 satisfies @xmath156 for some @xmath157 .
if we denote @xmath158 , then @xmath137 for some @xmath159 . in particular , if @xmath160 , the conclusion of theorem [ main ] holds . in view of proposition [ vort1 ] , it would be nice to extend the conclusion theorem [ main ] so as to include a convergence result for the vorticity distribution in the critical space @xmath161 .
this is not immediately obvious , because the classical @xmath162@xmath163 estimates for the stokes semigroup in an exterior domain do not include the limiting case @xmath164 , see @xcite .
however , combining theorem [ main ] with a relatively standard estimate , which shows that the @xmath165 norm of the vorticity can not leak to infinity , we obtain the following result which is the main original contribution of the present paper : [ vort2 ] under the assumptions of theorem [ main ] , if we suppose in addition that holds for some @xmath166 , then the vorticity @xmath167 satisfies @xmath168 where @xmath123 is defined in . in the rest of this paper , we give a simplified proof of theorem [ main ] , which does not yield the optimal conclusion .
in particular , we shall find a suboptimal convergence rate in , and our limitation on the size of the circulation will ( a priori ) depend on the domain @xmath24 .
we refer the reader to @xcite for a complete proof , including all details . in the last section ,
we briefly show how proposition [ vort2 ] follows from theorem [ main ] , using some additional information on the vorticity near infinity .
given @xmath95 we consider solutions of of the form .
the perturbation @xmath131 vanishes on the boundary @xmath4 and satisfies the equation @xmath169 where @xmath170 is the remainder term given by below .
it is not difficult to verify that the cauchy problem for equation is globally well - posed in the energy space @xmath171 .
the goal of this section is to control the long - time evolution of the perturbation @xmath172 using energy estimates .
first of all , we multiply both sides of by @xmath173 and integrate by parts over @xmath24 . taking into account the no - slip boundary condition
, we find @xmath174 where @xmath175 denotes the usual scalar product in @xmath171 , so that @xmath176 .
to estimate the right - hand side of , we use the results of section [ sec6 ] below .
first , in view of , we have @xmath177 moreover , applying with @xmath178 , we see that @xmath179 we thus obtain the energy inequality @xmath180 using gronwall s lemma , we deduce that @xmath181 for @xmath182 .
this simple estimate shows that the energy of the perturbation @xmath131 grows at most polynomially in time as @xmath0 . such a conclusion is rather pessimistic , however , because by a relatively simple modification of the previous argument it is possible to establish a logarithmic bound , which is clearly superior for large times
. [ logbound ] there exists a constant @xmath183 such that , for any circulation @xmath95 and any @xmath184 , the solution of with initial data @xmath185 satisfies @xmath186 for all @xmath19 , where @xmath187 .
if @xmath188 then follows from with @xmath189 , hence we can assume that @xmath190 .
given any @xmath191 , we denote @xmath192 for all @xmath57 and all @xmath19
. then @xmath193 satisfies where @xmath130 and @xmath194 are replaced by @xmath195 and @xmath196 , respectively . proceeding exactly as above
, we thus obtain the energy estimate @xmath197 now , we fix @xmath190 and choose @xmath198 . from , , we have @xmath199 similarly , using , we find @xmath200 thus it follows from ( with @xmath198 ) that @xmath201 for all @xmath31 , where @xmath202 . finally , we have by @xmath203 hence easily follows from .
[ logrem ] the logarithmic energy estimate is the main new ingredient in the proof of theorem [ main ] . to a certain extent , we use it as a substitute for the energy equality , which does not make sense for the solutions we consider .
as is clear from the proof , the logarithmic energy estimate relies on the fact that oseen s vortex has `` nearly finite energy '' , in the sense that the integral defining @xmath204 diverges only logarithmically at infinity .
let @xmath205 be the leray - hopf projection in @xmath24 , and @xmath206 be the stokes operator , see e.g. @xcite .
we recall that @xmath207 is self - adjoint and nonnegative in @xmath171 , so that the fractional power @xmath208 can be defined for all @xmath209 .
the following result shows that the range of @xmath210 contains the ( dense ) subspace @xmath211 .
[ amu ] @xcite let @xmath30 and @xmath212 . for all @xmath213 ,
there exists a unique @xmath214 such that @xmath215 .
moreover , there exists a constant @xmath216 ( independent of @xmath173 ) such that @xmath217 .
[ a - mu ] if @xmath218 are as in the above statement , we denote @xmath219 . roughly speaking
, the proof of lemma [ amu ] argues as follows . by classical sobolev embedding , we know that the domain of @xmath210 is contained in @xmath220 , where @xmath221 , and by duality we deduce that the range of @xmath222 contains @xmath223 .
we go back to the study of the perturbation equation , which can be written in the equivalent form @xmath224 so far we only considered solutions in the energy space @xmath225 , but now we assume in addition that @xmath226 , for some fixed @xmath30 , and we denote @xmath227 . then it is not difficult to verify that the solution @xmath228 of lies in @xmath211 for all @xmath229 .
in particular , invoking lemma [ amu ] , we can define @xmath230 for all @xmath19 .
this quantity solves the modified equation @xmath231 where @xmath232 is the bilinear term formally defined by @xmath233 we refer to ( * ? ? ?
* section 2 ) for a rigorous definition and a list of properties of the bilinear map @xmath234 .
our goal here is to establish the following estimate : [ wenergy ] there exists @xmath235 and , for all @xmath95 , there exist positive constants @xmath236 and @xmath237 such that , if @xmath173 is any solution of with initial data @xmath238 , the function @xmath239 satisfies @xmath240 for all @xmath19 . moreover @xmath236 and @xmath237 are @xmath241 as @xmath242 .
taking the scalar product of both sides of by @xmath243 , we obtain @xmath244 it is well known that @xmath245 for all @xmath246 . to bound the other terms ,
we observe that @xmath247 where in the last inequality we used the interpolation inequality for fractional powers of @xmath207 .
the same argument shows that @xmath248 . in a similar way , @xmath249 finally , since @xmath250 by
, we can use interpolation and young s inequality to obtain @xmath251 for some @xmath252 satisfying @xmath253 .
thus implies @xmath254 for some positive constants @xmath255 .
now , using with @xmath178 and the logarithmic energy estimate , we obtain @xmath256 where @xmath257 , @xmath258 , and @xmath259 . applying gronwall s lemma to
, we thus find @xmath260 where @xmath261 , and follows since @xmath262 by lemma [ amu ] .
[ vcor ] under the assumptions of proposition [ wenergy ] , there exists a positive constant @xmath263 depending on @xmath264 and @xmath265 such that , for any @xmath266 , there exists a time @xmath267 $ ] for which @xmath268 fix @xmath266 . in view of
, there exists a time @xmath267 $ ] such that @xmath269 where @xmath270 .
moreover , @xmath271 by .
thus , we obtain using the interpolation inequality @xmath272 .
we are now able to conclude our sketch of the proof of theorem [ main ] . given @xmath95 and @xmath238 , let @xmath131 be the solution of the perturbation equation .
for any @xmath266 , we choose @xmath267 $ ] as in corollary [ vcor ] , and we apply estimate .
we thus obtain @xmath273 where @xmath274 is @xmath241 as @xmath242 .
now , if @xmath264 is small enough to that @xmath275 , the right - hand side of converges to zero ( at a suboptimal rate ) as @xmath0 .
in particular , the perturbation @xmath132 becomes very small in energy norm for large times . in that regime ,
the perturbation equation can be solved by a global fixed point argument , which allows to show that @xmath276 as @xmath0 , see ( * ? ? ?
* section 3 )
. finally follows from , because @xmath277 and @xmath278 for all @xmath19 .
this final section is devoted to the proof of proposition [ vort2 ] .
we first show that , under the assumptions of theorem [ main ] , one can control the @xmath165 norm of the vorticity sufficiently far away from the obstacle @xmath3 .
[ noleak ] under the assumptions of proposition [ vort2 ] , the vorticity @xmath167 satisfies @xmath280 since by the initial vorticity is assumed to be square integrable , the solution @xmath11 of given by theorem [ main ] satisfies @xmath281 for all @xmath19 , where @xmath282 depends only on the initial data .
the associated vorticity @xmath167 is a solution of the advection - diffusion equation @xmath283 but the no - slip boundary condition becomes very complicated when expressed in terms of @xmath66 .
it is thus difficult to use to obtain estimates in the whole domain @xmath24 , and in particular near the boundary @xmath4 . here
, however , our goal is to bound @xmath66 near infinity , so we can avoid that problem using localized energy estimates and invoking to control the flux terms in the regions where the localization function is not constant .
given @xmath284 and @xmath285 , we define the cut - off function @xmath286 where @xmath287 , and @xmath288 $ ] is a smooth , nondecreasing function satisfying @xmath289 for @xmath290 and @xmath291 for @xmath292 .
we always assume that @xmath284 is large enough so that the support of @xmath293 is contained in @xmath24 , and that @xmath285 is large enough ( depending on @xmath294 and @xmath295 ) so that @xmath293 is not identically zero . given @xmath296
, we also denote @xmath297 and we observe that @xmath298 and @xmath299 for all @xmath300 .
now , using , we obtain by a direct calculation @xmath301 because @xmath302 and @xmath299 . if we integrate this inequality over @xmath303 $ ] , we find @xmath304 where @xmath305 .
using lebesgue s monotone convergence theorem , we can take the limit @xmath306 in both sides , and we arrive at the simpler estimate @xmath307 our next task is to take the limit @xmath308 in .
again , we use the monotone convergence theorem , except in the last integral where it does not apply . to treat that term ,
we observe that @xmath309 vanishes identically except in the region @xmath310 , where for any @xmath311 we denote @xmath312 .
taking @xmath313 sufficiently large and using , we easily obtain @xmath314 for some @xmath315 independent of @xmath303 $ ] .
the contribution of the annulus @xmath316 is therefore negligible for large @xmath317 , hence taking the limit @xmath308 in we arrive at @xmath318 where @xmath319 .
in particular @xmath320 for @xmath321 and @xmath322 for @xmath323 , hence the last inequality implies @xmath324 to conclude the proof of proposition [ noleak ] , it remains to estimate both terms in the right - hand side of .
first , using and hlder s inequality , we easily find @xmath325 for some @xmath326 independent of @xmath295 . in the last inequality
, we used the hypothesis @xmath327 and the fact that @xmath328 . on the other hand , since @xmath329 is given by with @xmath330 , there exists @xmath331 such that @xmath332 where @xmath333 .
it follows that @xmath334 but using and we easily find , @xmath335 where in the last inequality we used the fact that @xmath336 . in a similar way
, @xmath337 inserting these estimates in the right - hand side of , we obtain @xmath338 for some @xmath339 independent of @xmath295 .
thus , if we combine , , and , we conclude that @xmath340 for all sufficiently large @xmath89 , which is the desired result .
it is now easy to conclude the proof of proposition [ vort2 ] . for @xmath31
sufficiently large , we denote @xmath341 and we decompose @xmath342 the last integral in the right - hand side is controlled using proposition [ noleak ] and the explicit expression of @xmath123 . to estimate the first integral , we use and hlder s inequality : @xmath343 summarizing , we find @xmath344 for all sufficiently large @xmath31 .
this concludes the proof .
in this appendix , we collect a few estimates on the truncated oseen vortices which are used throughout the paper .
we first list a few bounds which follow from and by rather straightforward calculations , see ( * ? ? ?
* lemma 2.1 ) .
[ uchiest ] + * 1 . * for any @xmath345 $ ] , there exists a constant @xmath346 such that @xmath347 * 2 . * for any @xmath348 $ ] , there exists a constant @xmath349 such that @xmath350 * 3 .
* there exists a constant @xmath351 such that , for all @xmath352 , @xmath353 * 4 .
* there exists a constant @xmath354 such that , for all @xmath352 , @xmath355 [ rchiest ] there exists a constant @xmath358 such that , for any @xmath359 $ ] , @xmath360 moreover , for any vector field @xmath361 , we have @xmath362 where @xmath363 is a compact annulus containing the support of @xmath364 .
it is clear from that @xmath365 for all @xmath366 and all @xmath19 , and follows immediately .
moreover , we have @xmath367 for some radially symmetric scalar function @xmath368 , hence @xmath369 has zero mean over the annulus @xmath370 . if @xmath361 and if we denote by @xmath67 the average of @xmath40 over @xmath370 , the poincar - wirtinger inequality implies @xmath371 and using with @xmath372 we obtain . | we study the long - time behavior of infinite - energy solutions to the incompressible navier - stokes equations in a two - dimensional exterior domain , with no - slip boundary conditions .
the initial data we consider are finite - energy perturbations of a smooth vortex with small circulation at infinity , but are otherwise arbitrarily large . using a logarithmic energy estimate and some interpolation arguments ,
we prove that the solution approaches a self - similar oseen vortex as @xmath0 .
this result was obtained in collaboration with y. maekawa ( kobe university ) .
nous tudions le comportement asymptotique en temps des solutions de lquation de navier - stokes incompressible dans un domaine extrieur du plan , avec condition de non - glissement la frontire .
les donnes initiales que nous considrons sont des perturbations dnergie
finie dun tourbillon rgulier do nt la circulation linfini est petite , mais nous nimposons aucune autre restriction leur taille . en utilisant une estimation dnergie logarithmique
et des arguments dinterpolation , nous montrons que la solution converge lorsque @xmath0 vers un tourbillon doseen autosimilaire .
ce rsultat a t obtenu en collaboration avec y. maekawa ( universit de kobe ) . |
the topological susceptibility in the yang mills ( ym ) gauge theory can be formally defined in euclidean space - time as @xmath4 where the topological charge density @xmath5 is @xmath6\ ; .\ ] ] besides its interest within the pure gauge theory , @xmath7 plays a crucial rle in the qcd - based explanation of the large mass of the @xmath3 meson proposed by witten and veneziano ( wv ) @xcite .
the wv mechanism predicts that , at leading order in @xmath8 , the contribution due to the anomaly to the mass of the @xmath9 particle is given by @xcite @xmath10 where @xmath11 is the pion decay constant .
the discovery of a fermion operator @xcite that satisfies the ginsparg
wilson ( gw ) relation @xcite triggered a breakthrough in the understanding of the topological properties of the ym vacuum @xcite , and made it possible to give a precise and unambiguous implementation of the wv formula @xcite .
indeed the naive lattice definition of the topological susceptibility has a finite continuum limit , which is universal @xcite , if the topological charge density , defined as suggested by gw fermions , is employed @xcite .
this yields the suggestive formula @xmath12 with @xmath13 being the topological charge , @xmath14 the volume , and @xmath15 ( @xmath16 ) the number of zero modes of @xmath17 with positive ( negative ) chirality in a given background . using new simulation algorithms @xcite , it is now possible to investigate the wv scenario from first principles for the first time .
more precisely , the aim of the work presented here , and fully described in @xcite , is to achieve an accurate and reliable determination of @xmath7 in the continuum limit , which in turn allows a verification of the wv mechanism for the @xmath3 mass .
several exploratory computations have already studied the susceptibility employing the gw definition of the topological charge @xcite .
comparison with other computations.,title="fig:",width=264 ] -0.75 cm -1.0 cm
the ensembles of gauge configurations are generated with the standard wilson action and periodic boundary conditions , using a combination of heat - bath and over - relaxation updates .
more details on the generation of the gauge configurations can be found in refs .
table [ tab : lattices ] shows the list of simulated lattices , where the bare coupling constant @xmath18 , the linear size @xmath19 in each direction and the number of independent configurations are reported for each lattice .
the topological charge density is defined as @xmath20 , \ ] ] with @xmath17 being the massless neuberger
dirac operator @xcite , and @xmath21 an adjustable parameter in the range @xmath22 . for a given gauge configuration ,
the topological charge is computed by counting the number of zero modes of @xmath17 ( the interested reader should refer to @xcite for more details ) .
as @xmath21 is varied , @xmath17 defines a one - parameter family of fermion discretizations , which correspond to the same continuum theory but with different discretization errors at finite lattice spacing .
our computation includes data sets computed for @xmath23 and @xmath24 . a comparison with results previously obtained with various implementation of the neuberger s operator is shown in fig .
[ fig : comparison ] . .
[ tab : lattices ] simulation parameters and results . for lattices @xmath25@xmath26 and @xmath27@xmath28 , @xmath23 and @xmath24 respectively . [ cols="<,<,^,>,^,^ " , ] -1.0 cm the systematic uncertainties stem from finite - volume effects and from the extrapolation needed to reach the continuum limit .
the pure gauge theory has a mass gap , and therefore the topological susceptibility approaches the infinite - volume limit exponentially fast with @xmath29 .
since the mass of the lightest glueball is around 1.5 gev , finite - volume effects are expected to be far below our statistical errors as soon as @xmath30 fm . in order to further verify that no sizeable finite - volume effects are present in our data
, we simulated four lattices at @xmath31 but with different linear sizes .
the results obtained for @xmath7 , reported in table [ tab : lattices ] , show no dependence on @xmath29 within our statistical errors .
continuum extrapolation of the adimensional product @xmath32 .
the @xmath24 and @xmath23 data sets are represented by black circles and white squares respectively .
the dashed lines represent the results of the combined fit described in the text .
the filled diamond at @xmath33 is the extrapolated value in the continuum limit.,title="fig:",width=264 ] -0.75 cm -0.75 cm at finite lattice spacing , @xmath7 is affected by discretization effects starting at @xmath34 , which are not universal , and , in our case , depend on the value of @xmath21 chosen to define the neuberger operator .
the values of the adimensional quantity @xmath32 that we obtain , where @xmath35 is taken from @xcite , are reported in table [ tab : lattices ] .
data , displayed in fig .
[ fig : continuum ] as a function of @xmath36 , show sizeable @xmath34 effects for both the @xmath23 and @xmath24 samples . for @xmath37 ,
the difference between the two discretizations is statistically significant .
a detailed data analysis indicate a linear dependence in @xmath38 of our results within our statistical errors @xcite .
a robust estimate of @xmath7 in the continuum limit can be obtained by performing a combined linear fit of the data .
this fit gives a very good value of @xmath39 when all sets are included , and is very stable if some points at larger values of @xmath36 are removed . in particular
a combined fit of all points with @xmath40 gives @xmath41 with @xmath42 .
since @xmath43 is not directly accessible to experiments , we express our result in physical units by using the lattice determination of @xmath44 in the quenched theory @xcite and we obtain @xmath45 which has to be compared with @xcite @xmath46 notice that eq .
( [ eq : wv ] ) being only valid at the leading order in a @xmath47 expansion , the ambiguity in the conversion to physical units in the pure gauge theory is of the same order as the neglected terms .
our result supports the witten
veneziano mechanism for explaining the large @xmath3 mass .
a measurement of the connected expectation values @xmath48 with the methods presented here would provide interesting informations about the dependence of the free energy density on the @xmath49 angle , or equivalently on the probability distribution of the topological charge @xmath50 , putting on solid ground the results presented in ref .
unfortunately much higher statistics are required in order to highlight the deviations from a gaussian distribution ; higher momenta of the topological charge distribution measured on our data are all compatible with zero within large statistical errors . | we present the results of a computation of the topological susceptibility in the su(3 ) yang mills theory performed by employing the expression of the topological charge density operator suggested by neuberger s fermions . in the continuum limit
we find @xmath0 , which corresponds to @xmath1 if @xmath2 is used to set the scale .
our result supports the witten
veneziano explanation for the large mass of the @xmath3 .
-0.25 cm |
the perihelion precession of planetary orbits and the bending of light by a massive star were examined a long time ago in the context of einstein s general theory of relativity ( @xcite@xcite ) in a schwarzschild metric for the background spacetime . since then , there have been various proposals of alternative theories of gravitation beyond einstein s gravity . especially in the context of string theory the effects of the extra scalar field , namely , the dilaton , lead to a new model of gravity in the low - energy effective theory in four dimensions ( see gibbons and hull ( @xcite , @xcite ) , hawking ( @xcite ) , witten ( @xcite ) , mayers et.al .
( @xcite , @xcite ) and chandrasekhar and hartle ( @xcite ) .
in addition there are some other higher order gravity theories which admit spherical symmetry for the spacetime . among them
einstein - maxwell - gauss - bonnet ( henceforth referred to as emgb ) theory ( see @xcite , @xcite ) and spherically symmetric solution in quadratic gravity ( see @xcite , @xcite ) are of utmost importance .
+ in this paper , we present a general formalism based on the perturbative method to determine the timelike or null geodesics and estimate the perihelion precession and light bending for alternative gravity theories . among them
one very important candidate is dilaton gravity in the presence of a @xmath0 gauge field which yields a charged black hole solution in the low - energy limit of string theory compactified to four dimensions .
such a solution has been analyzed in great detail by garfinkle , horowitz and strominger @xcite ( see also coleman @xcite ) .
we have shown that , while in the case of a pure reissner - nordstrm ( r - n ) solution the bending of light , in the leading order , is similar to the schwarzschild scenario , it differs significantly for a charged dilaton coupled solution .
for such a solution the bending angle depends nontrivially on the vacuum expectation value of the dilaton which is a measure of string coupling .
we have also presented the calculation of the perihelion shift and bending of light for the emgb gravity theory where we have a higher order gauss - bonnet term and action for an electromagnetic field .
the natural coupling in this type of theory is the inverse of string constant @xmath1 , which we have estimated by using known observational results . for the emgb theory
we have also obtained that the shift decreases with charge of the black hole as in the r - n scenario ; however the expression for the bending angle of light differs significantly and increases with the coupling @xmath1 .
recently yunes and stein ( see @xcite ) have obtained spherically symmetric solution in quadratic gravity in the context of dynamical theories .
we have used that solution to calculate the perihelion shift and bending of light and obtained nontrivial results .
our results further reveal that while for r - n spacetime the perihelion shift decreases with increase in electric charge , for a dilaton coupled electromagnetic background it increases with the electric charge .
+ as an illustration , we first very briefly present the perihelion precession and bending of light in a reissner - nordstrm geometry and then extend our calculations for alternative theories of gravity .
our analysis includes charged dilaton black hole , the emgb gravity theory , and a spherically symmetric solution in quadratic gravity . in every case we have tried to estimate a bound on different coupling parameters which represent the characteristic scale of the corresponding model beyond einstein gravity .
the paper ends with a discussion on our results .
the metric for a r - n black hole is given by ( see @xcite , @xcite ) @xmath2 where @xmath3 is the mass of the black hole and @xmath4 is the charge . since the metric does not contain @xmath5 and @xmath6 explicitly , the corresponding two conserved quantities are the energy and angular momentum such that @xmath7 here @xmath8 is the mass of the particle , @xmath9 is the energy per particle mass , @xmath10 is the angular momentum per particle mass , and @xmath11 and @xmath12 are the components of four - momentum . in our chosen unit system , mass has same dimension as energy and @xmath9 is dimensionless
. we shall use @xmath13 for massive particles and @xmath14 for massless photons .
also , we can choose the motion of the particle in the equatorial plane ( @xmath15 ) ( see @xcite , @xcite ) such that @xmath16 .
the geodesic equation can be obtained by using the energy - momentum relation @xmath17 and defining @xmath18 where @xmath19 is some affine parameter connected with the proper time of the particle . then using the metric coefficients and the energy - momentum relation we get an expression for @xmath20 . to eliminate @xmath19
, we introduced @xmath21 , and finally by simple algebraic manipulation with a substitution @xmath22 we arrive at , @xmath23 here@xmath24 denotes derivative with respect to @xmath6 . on further differentiation and keeping terms @xmath25
, we arrive at @xmath26 where we have taken @xmath27 , @xmath28 and @xmath29 .
we assume that after each revolution the period of @xmath6 changes by a small amount , i.e. , @xmath30 $ ] , where @xmath1 is the measure of perihelion precession which depends on @xmath31 , @xmath32 and @xmath33 . by using the perturbation method ,
the solution for @xmath34 is taken to be @xmath35 .
then , by keeping terms linear in @xmath33 ( since @xmath36 as well as the quantity @xmath37 are small ) , we obtain @xmath38 note that the solution of an equation of the form @xmath39 will be nonresonant when the @xmath40 term vanishes ( see @xcite ) , which in turn implies @xmath41 .
hence the perihelion precession would be @xmath42 [ fig1 ] figure 1 : the figure shows the variation of the perihelion shift with charge .
thus for the r - n case the shift is given by @xmath43 figure 1 depicts how the perihelion shift decreases with an increase in charge of the source . from eq .
@xmath44 we see that for a given @xmath45 the perihelion shift @xmath46 decreases with @xmath4 due to the negative sign in front . for @xmath47 ,
the curve in the figure touches the @xmath48 axis which corresponds to the value of @xmath46 for the schwarzschild scenario .
for example , if we took @xmath49 , then the curve should touch the @xmath48 axis at @xmath50 , i.e. , at @xmath51 , which can be observed from the figure . to estimate the bending of light we can write the null geodesic as @xmath52\ ] ] simple algebra , as shown in the previous section ,
leads the above expression to @xmath53 where @xmath54 .
equation ( @xmath55 ) indicates that the term that contains @xmath4 is of the order of @xmath56 and therefore has a very negligible influence on the measure of bending of light when the minimum distance of approach is large .
note that the term containing @xmath57 also appears in the massive case .
however there we have a factor of @xmath58 , which makes a nonzero contribution from that term .
hence we can conclude that for r - n geometry the bending of light is almost identical to that of a schwarzschild black hole at a sufficiently large distance from the source such that terms beyond the leading order , i.e. , @xmath59 can be ignored .
static uncharged black holes in general relativity are described by schwarzschild solution .
if the mass of the black hole is large compared to the planck mass , then this also , to a good approximation , describes the uncharged black hole in string theory except regions near a singularity .
however , the einstein - maxwell solution for the string - inspired theory differ widely from the known classical solution due to dilaton coupling .
the dilaton coupling with electromagnetic field tensor @xmath60 implies that every solution with nonzero @xmath61 will come with a non zero dilaton .
thus the charged black hole solution in general relativity ( which is the reissner - nordstrm solution ) appears in a new form in string theory due to the presence of the dilaton .
the effective four - dimensional low - energy lagrangian obtained from string theory is @xmath62\ ] ] where @xmath63 is the maxwell field lagrangian associated with a @xmath0 subgroup of @xmath64 or @xmath65 .
we have set the remaining gauge and antisymmetric tensor field @xmath66 to zero and focus into the presence of @xmath67 , the dilaton field ( see garfinkle , horowitz and strominger @xcite , coleman @xcite , vega and sanchez @xcite , bekenstein @xcite , and witten @xcite ) . extremizing with respect to the @xmath0 potential @xmath68 , @xmath69 , and @xmath70 , we have the following field equations : @xmath71 we wish to find a static spherically symmetric solution corresponding to the above field equations ( [ n1 ] ) that are asymptotically flat and have a regular horizon . for this purpose
, the general metric ansatz can be taken as @xmath72 , where @xmath73 and @xmath74 are functions of @xmath75 only . for a purely magnetic maxwell field @xmath76
, we have @xmath77 and can deduce that there exist only three independent components for the ricci tensor , namely , @xmath78 , @xmath79 , and @xmath80 .
we can also show from eq .
( [ n1]c ) that @xmath81 , which leads to @xmath82 , with the prime denoting derivative with respect to the radial coordinate .
again from eq .
( [ n1 ] ) we find the equation , @xmath83 .
using these equations and ( [ n1 ] ) , we can find out the spherically symmetric solution ( see garfinkle , horowitz and strominger @xcite ) @xmath84 where , @xmath85 .
once again due to isometry , we have taken our motion in the equatorial plane such that @xmath86 . here
@xmath87 is the asymptotic value of the dilaton , and @xmath4 represents the black hole charge .
note that this is almost identical to the schwarzschild metric , with a difference that areas of spheres of constant @xmath75 and @xmath5 now depend on @xmath4 .
in particular the surface @xmath88 is singular , and @xmath89 is the regular event horizon .
the evolution of the scalar field @xmath69 may also be derived from the field equations mentioned before eq .
( [ 31 ] ) and the solution given by @xmath90 hence , as @xmath91 , @xmath92 .
we can define the dilaton charge as @xmath93 where the integral is over a two sphere at spatial infinity and @xmath94 is the normal to the two sphere at spatial infinity . for a charged black hole we could compute this by using eq .
( [ n1]b ) and the expression for a purely magnetic maxwell field @xmath76 , which leads to @xmath95 this indicates that the matter particle feels the presence of the dilaton with a coupling modified by @xmath96 . here
@xmath97 depends on the asymptotic value of dilaton field , which is determined once @xmath3 and @xmath4 are given and is always negative .
note that the actual dependence on the dilaton field is described by @xmath98 .
since we have worked in the unit @xmath99 , the term is modified to @xmath100 . as @xmath101
, this term is expected to become significant .
+ following the same procedure as in the previous section , we find for a test particle @xmath102\ ] ] where @xmath9 and @xmath10 are conserved energy and angular momentum respectively .
also , @xmath103 is @xmath104 for a massless photon and is @xmath105 for a massive particle . defining @xmath106 and using @xmath107 we arrive at @xmath108\ ] ] for massive particles ( @xmath109 ) , eq .
( @xmath110 ) becomes @xmath111 -u^{2}[1+\frac{a^{2}+4ma}{l^{2}}-\frac{a^{2}e^{2}}{l^{2 } } ] + u^{3}[2m+a+\frac{2ma^{2}}{l^{2}}]\ ] ] differentiation yields @xmath112 note that in the limit @xmath113 the above equation reduces to the orbit equation for schwarzschild solution . just as in the previous case , here also the quantity @xmath114 is small since the quantity @xmath31 contains mass of the source in inverse power . + by comparing with eq .
( @xmath115 ) , the perihelion precession is now estimated as @xmath116[1+\frac{q^{2}}{2m^{2}}e^{2\phi_{0}}+ \frac{q^{4}}{m^{2}l^{2}}e^{4\phi_{0 } } ] - \frac{\pi q^{2}}{ml^{2}}e^{2\phi_{0}}[4m+\frac{q^{2}}{m}(1-e^{2})e^{2\phi_{0}}]\ ] ] neglecting terms of the order of @xmath117 , we obtain @xmath118\ ] ] again for @xmath119 it reduces to the schwarzschild case . from eq .
( @xmath120 ) it is evident that the perihelion shift increases with the asymptotic value of dilaton , which is shown explicitly in fig .
2 . figure 2 : the figure shows the variation of perihelion shift with a dilaton field for different choices of @xmath4 and @xmath3 .
it may be observed that the perihelion shift increases very fast with an increase on dilaton strength and also increases with an increase in charge of the star or the black hole .
note that both the r - n solution and the dilaton solution describe a black hole with mass @xmath3 and charge @xmath4 only when @xmath121 is small ; otherwise they describe naked singularity . however , we have an important difference between these solutions , as in this case there is no analog of the inner horizon which exists for the r - n scenario .
but , for string theory , the statement that the horizon is singular when @xmath122 is actually irrelevant , since the strings do not couple to the metric @xmath70 but rather to @xmath123 .
this metric appears in the string @xmath124 model . in terms of the string metric ,
the effective lagrangian becomes ( see @xcite ) @xmath125\ ] ] hence the charged black hole metric is obtained as @xmath126 this metric is identical to the metric given in eq .
@xmath127 where we have just rescaled the metric by some conformal factor which is finite everywhere outside and on the horizon .
if we try to calculate the perihelion precession by using this metric following the same method , we arrive at the same result presented by eq . @xmath128 . from the available experimental data on the perihelion precession of mercury
, the error bar is estimated as @xmath129 arcsec / century .
so we can estimate a bound for the dilaton charge , which is @xmath130 .
this clearly explains why the stringy signature can still not be determined within the precision of the present day astrophysical experiments .
figure 3 : the figure shows the variation of bending of light with a dilaton strength for different choices of mass and charge . for the null geodesic the orbit equation yields @xmath131 introducing @xmath132 ,
i.e. , angular momentum per unit energy or , equivalently , angular momentum per particle mass , we obtain @xmath133 by neglecting terms of order @xmath59 , the above differential equation can be integrated such that @xmath134 where @xmath135 and @xmath136 .
the lower limit @xmath137 represents the position of the photon when it does not feel the presence of the field , i.e. , at @xmath138 .
the upper limit yields the minimum distance of approach , which in this case is the solution of the equation @xmath139 . using these we determine the bending of light as @xmath140 where @xmath141 .
note that in the limit @xmath142 this vanishes , which is consistent , since in eq .
( @xmath143 ) we have eliminated the @xmath59 term which appears in the schwarzschild case . as we have neglected the dominant term in the schwarzschild case ( @xmath144 ) , we should not expect the schwarzschild result to be retrieved .
thus , the above bending is dependent crucially on the dilaton asymptotic value and comes from a more dominant term ( @xmath25 ) than that in the schwarzschild scenario .
so this bending can be interpreted as a signature of a string - inspired dilaton spacetime . here
( [ n3 ] ] where @xmath97 is the dilaton charge . using this
we finally obtain @xmath146 with @xmath147 defined after eq .
( [ 43a ] ) .
however , the above solution is valid for @xmath148 . for the other choice
the bending of light is given by @xmath149 where @xmath150 .
this also vanishes in the limit @xmath151 and hence is consistent with the above analysis .
the first solution is applicable when the @xmath152 of the particle is much greater than the dilaton strength , and the second condition is just the reverse of the first condition .
following the classic test of general relativity by cassini spacecraft @xcite , we could put bounds on the dilaton charge @xmath97 .
the quantity @xmath153 could be calculated as @xmath154 .
then , following the constraint put on this @xmath155 by cassini @xcite , we get , @xmath156 .
this bound on @xmath97 is distinctly several orders larger than the bound derived from the perihelion precession test .
however one should also need to consider damour - polyakov mechanism , which generates a screening effect and may keep the estimate within experimental bounds ( @xcite , @xcite ) .
string - loop corrections of low - energy matter couplings of the dilaton provide a mechanism to fix the vacuum expectation value of the massless dilaton , compatible with existing experimental data . by using some assumptions regarding the universal nature of the dilaton coupling function ,
it is possible to show that the cosmological evolution of a graviton - dilaton - matter system drives the dilaton to a value when it decouples from matter , called the least coupling principle @xcite .
the origin of this evolution is explained as follows : masses of different particles depend on the dilaton @xmath6 , while the source for the dilaton mass is the gradient of @xmath6 .
it is therefore not surprising to have a fixed point where the gradient of @xmath6 is zero leading to a massless dilaton .
this mechanism is similar to the generic attractor mechanism in scalar tensor theories discussed in ref .
other types of couplings of a long - range scalar field @xmath6 atre given as a universal multiplicative coupling of @xmath6 to all the other fields , where @xmath157 with @xmath158 admitting a universal extremum .
the universal coupling function @xmath158 can be expanded in powers of @xmath159 in the small coupling limit , i.e. , @xmath160 or , equivalently , @xmath161 such that @xcite @xmath162 in the low - energy limit of the underlying string theory , which is the one we are discussing here , or equivalently the coupling constant , @xmath160 leads to the first term in the above equation to be the dominant one validating our results . in this limit
we may ignore the backreaction of the other terms in the expansion of @xmath158 on the metric and can use the form of the metric discussed in ( @xmath163 ) . in the low - energy limit
the original damour - polyakov action written in the `` string frame '' can be reduced to that in the `` einstein '' frame with a conformal transformation such that the action in the einstein frame is identical to that of the effective low - energy lagrangian used in this section @xcite .
this justifies the use of the damour - polyakov mechanism to lower the bounds on the dilaton charge .
we summarize in qualitative terms the basic reasons why a massless dilaton is rendered nearly invisible during its cosmological evolution : ( i ) during the radiation era , when the universe passes through a temperature @xmath164 , the @xmath31-type particles and antiparticles become nonrelativistic before annihilating and hence disappearing from the cosmic soup ; this provides a source term for the dilaton , which attracts @xmath6 toward a minimum @xmath165 of @xmath166 . according to ref .
@xcite this attraction mechanism is efficient .
( ii ) in the subsequent matter era , @xmath6 will be continuously attracted toward a minimum of the mass function @xmath167 corresponding to the matter dominating the universe .
( under the same universality condition this minimum will be again @xmath168 . ) thus , by , following this damour - polyakov mechanism it is possible to reduce the dilaton charge beyond the experimentally observable value by the screening effect ( @xcite , @xcite and @xcite ) .
hence the effective bound on dilaton charge is lower than estimated bound following the mechanism illustrated above .
we , however , reiterate that the angle for the bending of light increases in magnitude with the dilaton strength ( see fig .
3 ) and also with the electric charge .
theories with an extra spatial dimension have been an active area of interest ever since the original work of kaluza and klein and the advent of string theory , which predicts the inevitable presence of an extra spatial dimension . among many alternatives ,
the brane world scenario is considered as a strong candidate within some theoretical basis in some underlying string theory .
usually , the effect of string theory on classical gravitational physics ( see @xcite , @xcite ) is investigated by means of a low - energy effective action , which in addition to the einstein - hilbert action contains squares and higher powers of curvature term .
however , the field equations become fourth order and bring in ghosts ( see @xcite ) . in this context , lovelock ( see @xcite ) showed that , if the higher curvature terms appear in a particular combination , the field equation become second order and consequently the ghosts disappear . in emgb gravity , the action in five - dimensional spacetime ( @xmath169 ) can be written as @xmath170,\ ] ] where @xmath171 is the gb lagrangian and @xmath172 is the lagrangian for the electromagnetic field . here
@xmath1 is the coupling constant of the gb term having dimension ( length)@xmath173 . as @xmath1 is regarded as inverse string tension , so @xmath174 . for the gravitational and electromagnetic field equations obtained by varying the above action with respect to @xmath70 , and @xmath68 , we obtain ( see @xcite )
@xmath175-\frac{1}{2}g_{\mu \nu}l_{gb } \end{array}\right \}\end{aligned}\ ] ] where @xmath176 is the electromagnetic field tensor .
a spherically symmetric solution to the above action was obtained by dehghani ( see @xcite ) , and the corresponding line element is given by @xmath177 where the metric coefficient is @xmath178\ ] ] it should be noted that , under asymptotic expansion ( this is identical to small @xmath1 expansion ) , the function @xmath179 can be written as @xmath180 here @xmath181 is the curvature , @xmath182 is the geometrical mass , and @xmath183 is the metric of a 3d hypersurface such that @xmath184 the range is given by @xmath185 $ ] . note that the gauss - bonnet term decreases the mass of the spacetime for negative @xmath1 and increases for positive @xmath1 .
it is worthwhile to mention that this happens only in five - dimensional spacetime @xcite .
we assume that there is a constant charge @xmath186 at @xmath187 and the vector potential @xmath188 such that @xmath189 in five - dimensional spacetime . for four - dimensional effective theory ,
this reduces to the usual form @xmath190 . in this metric
the metric function @xmath179 will be real for @xmath191 , where @xmath192 is the largest real solution of the cubic equation @xmath193 by a transformation of the radial coordinate we can show that @xmath194 is an essential singularity of the spacetime .
we shall consider the negative sign in front of the square root in eq .
( [ 49 ] ) since the positive sign does not lead to an asymptotically flat solution @xcite .
one may note that for @xmath195 the function @xmath179 tends to @xmath104 as @xmath75 tends to infinity and is not an acceptable solution . for @xmath196
the function @xmath179 tends to @xmath105 as @xmath75 tends to infinity and thus we have an asymptotically flat solution . under this condition eq .
( [ 49a ] ) reduces to @xmath197 thus the presence of an additional dimension leads to a reissner - nordstrm black hole with mass parameter @xmath198 @xcite . in this scenario , we shall take the following choice , which we have used in the previous two sections by exploiting the spherical symmetry such that @xmath199 , @xmath200 , and @xmath201 .
then the equation of motion is given by @xmath202\ ] ] since we are interested in the trajectory of massive particles , we choose @xmath203 , defining the variable @xmath22 , and the above equation of motion reduces to @xmath204\ ] ] where we have @xmath205\ ] ] differentiating the above expression with respect to the angular coordinate , we readily obtain the following differential equation @xmath206 this equation can be simplified in the large @xmath75 or small @xmath34 limit as @xmath207u^{3}\ ] ] now , following the same procedure , i.e. , assuming a periodic solution as in the r - n scenario and rewriting the differential equation as @xmath208 where we have taken @xmath209 and @xmath210 , we finally obtain the following expression for the precession angle : @xmath211\ ] ] this can be rewritten by algebraic simplification as @xmath212 the solution though asymptotically flat , does not reduce to the schwarzschild solution in the limit @xmath213 , since the action modified by quadratic terms is not the usual einstein - hilbert action even if @xmath214 ( see @xcite ) . note that the perihelion shift varies linearly with the coupling parameter @xmath1 . as coupling increases , the shift also increases .
however , in this scenario also the shift decreases with the charge as in the r - n scenario . in this case
, we can estimate the parameter @xmath1 by using the estimate of 43 arcsec / century .
then this poses a limit on @xmath1 which is @xmath215 .
hence , this effect is also very small in respect to the present experimental resolution .
now we shall discuss the null geodesic solution to the above metric .
this can be obtained easily by substituting @xmath216 in eq .
@xmath217 , which leads to @xmath218\ ] ] here we have kept terms up to @xmath56 , and @xmath219 and @xmath220 are the energy and angular momentum of the photon , respectively . then from this the bending angle can be determined as @xmath221}}\ ] ] where @xmath54 , @xmath137 represents the asymptotic value from where the photon came , and the upper limit is the minimum distance of approach . then following the procedure presented in a work by chakraborty and chakraborty ( see @xcite ) , we finally obtain the bending angle as @xmath222 hence , we find that bending increases in magnitude with the coupling @xmath1 , which is expected , since the increase of coupling implies the increase in the strength of the field .
testing strong - field features of general relativity ( gr ) is very important for astrophysical signatures , as they could indicate departure from general relativity with deep implications in fundamental theory .
thus in order to test such features we need deviations from schwarzschild or kerr and look for non - gr solutions .
these non - gr solutions are known through numerical studies , where one choose an alternative theory , constructs the field equations , and by postulating a metric ansatz tries to address various astrophysical signatures .
the differential equations satisfied by these functions are solved and studied numerically ( see @xcite , @xcite ) .
we can solve those differential equations by approximation methods ; for example , we can expand in terms of the coupling constants of the theory .
this small - coupling approximation ( see @xcite ) treats the alternative theory as an effective and approximate model that allows for small perturbation about gr . in this section
we consider a class of alternative theories of gravity in four dimensions defined by modifying the einstein - hilbert action through all possible quadratic , algebraic curvature scalars , multiplied by constants or nonconstant couplings as ( see @xcite ) , + @xmath223 @xmath224\ ] ] where @xmath225 is the determinant of the metric @xmath226 ; @xmath74 , @xmath227 , @xmath228 , @xmath229 are the ricci scalar and tensor and the riemann tensor and its dual ( see @xcite ) , respectively ; @xmath230 is the lagrangian density for other matter ; @xmath231 is a scalar field ; @xmath232 are coupling constants ; and @xmath233 .
all other quadratic curvature terms are linearly dependent on these terms . theories of this type are motivated from low - energy expansion of string theory ( see @xcite , @xcite ) .
varying eq .
@xmath234 with respect to the metric and setting @xmath235 , we find the modified field equations @xmath236 where @xmath237 is the stress energy of matter and @xmath238 with @xmath239 , @xmath240 , and @xmath241 are the first- and second - order covariant derivative and the dalembertian
. the scalar field equation can be given by @xmath242 + the spherically symmetric solution to the above field equations , imposing dynamical arguments , could be written using the metric ansatz as ( see @xcite ) @xmath243dt^{2}+ f_{0}^{-1}\left[1+\epsilon k_{0}(r)\right]dr^{2}+r^{2}d\omega ^{2}\ ] ] and @xmath244 , where @xmath245 , with @xmath246 the bare mass and @xmath247 the line element on the two sphere .
the free functions @xmath248 are small deformations about the schwarzschild metric .
the scalar field equation can be solved to yield @xmath249 we can use this scalar field solution to solve modified field equations to an order linear in @xmath103 . requiring the metric to be asymptotically flat and regular at @xmath250
, we find the unique solution @xmath251 and @xmath252 , where @xmath253 and @xmath254 @xmath255 here we have defined the dimensionless coupling function @xmath256 , which is of the order of @xmath103 .
such a solution is most general for all dynamical , algebraic , quadratic gravity theories , in spherical symmetry .
we can define the physical mass @xmath257 $ ] such that the modified metric components become @xmath258 and @xmath259 where @xmath260 and @xmath261 . here ,
@xmath262 @xmath263 where @xmath264 .
note from the above expression for metric element that physical observables are related to renormalized mass @xmath3 , not bare mass @xmath246 . in this metric ansatz
we shall use the choice @xmath265 exploiting the spherical symmetry of the solution . then the orbit equation for a massive particle
would be given in the @xmath266 plane to yield @xmath267\ ] ] introducing the new variable @xmath268 we rewrite the orbit equation , using the metric co - efficient as , @xmath269}\ ] ] by substituting for @xmath270 and @xmath271 and using the fact that @xmath34 is a small quantity , the above orbit equation is simplified to @xmath272 + \frac{e^{2}}{l^{2}}\left[1+\zeta m^{2}u^{2}+\frac{8}{3}\zeta m^{3}u^{3}\right]\ ] ] this equation can be rearranged to yield @xmath273u^{2 } + \left[2m- \frac{\zeta m^{3}}{l^{2}}(1 - 8/3 e^{2})\right]u^{3}\ ] ] differentiating this equation again , we finally obtain , @xmath274\ ] ] now taking a trial solution of the form @xmath35 and keeping terms linear in @xmath33 , we arrive at the following expression for perihelion precession : @xmath275 as a check note that as @xmath276 we have @xmath277 and hence the above expression for the perihelion shift reduces to the schwarzschild result . rewriting the above expression in terms of the bare mass ,
we finally obtain @xmath278 from this equation we see that the second term in the right - hand side expression increases with @xmath279 , while the other two terms decrease
. however if @xmath9 is greater than 1 then all terms are positive .
so the perihelion shift depends on the energy of the particle .
also since we have kept terms linear in @xmath279 ( the metric itself is valid for linear terms in @xmath279 ; see @xcite ) , the shift varies linearly , but whether it will increase or decrease depends on the energy value . in this case
, we can estimate the parameter @xmath279 using the fact that perihelion precession measurements for mercury show an error bar of 3 arcsec / century . then using the above expression for the perihelion precession we can limit @xmath279 to a small value of @xmath280 .
so the effect is not as small as in the above two cases , though not significant enough to be within the present day experimental observation .
the null geodesic in this case is given by @xmath281 where we have kept terms up to the order of @xmath282 , which is the leading order contribution .
also @xmath219 and @xmath220 are the energy and angular momentum , respectively , of the photon under consideration .
this equation can be integrated to obtain the bending angle as @xmath283 here @xmath137 represents the asymptotic value of the inverse radius from where the photon approaches . since the solution is asymptotically flat this value is @xmath104 to be precise .
however the upper limit implies the minimum distance i.e. the solution of the equation @xmath139 .
then the above equation can be integrated to yield @xmath284 where @xmath132 . in this case
the bending depends on the leading order term @xmath25 and we have kept terms just up to that order .
so in the limit of @xmath285 the bending angle of light vanishes .
this shows that the parameter @xmath279 provides a longer range effect on bending of light than the schwarzschild mass does .
so we again have an instance where the bending angle is coming from a purely string theory effect in a modified schwarzschild solution in the presence of quadratic gravity .
in this paper , we have considered various alternative theories of gravity which have their roots in some underlying string theory .
considering three such variants , we have explored the influence of various modifications beyond einstein gravity on the perihelion shift and lensing . from the present available data for these phenomena , we have established a bound on the characteristic parameter of each model . for a dilaton
coupled charged solution , our work brings out the possibility of a stringy signature and also shows a remarkable feature that the dilaton vacuum expectation value , which decides the magnitude of the bending angle ( though suppressed by the scale of the theory ) , has its presence at one order lower in @xmath34 than the pure schwarzschild scenario .
this implies the possibility of having a more pronounced stringy effect than the pure einstein gravity .
the dependence of the bending angle as well as the precession angle with a dilaton vacuum value has also been determined for different values of mass and charge . for the precession angle
the charge dependence turns out to be opposite to that in r - n scenario , indicating that the dilaton coupled charged spacetime solution belongs to a new class of solutions other than the ordinary r - n scenario .
next in our analysis we have explored the role of the gauss - bonnet term at higher order .
the modification of the perihelion shift due to this additional term puts a stringent bound on the parameter of the gb coupling .
the cassini effect as well as the damour - polyakov mechanism have been discussed in this context . finally , we have considered quadratic gravity in a spherically symmetric spacetime . here
also a bound on the parameter of the model has been established after a critical analysis .
as all these models are inspired by string theory , our work thus establishes bounds on stringy parameters in respect to present astrophysical observations .
p. cagivendelas , g. t. horowitz , a. strominger and e. witten , nucl . phys . *
b258 * , 46 ( 1985 ) ; m. b. greens , j. h. schwarz and e. witten , ( 1987 ) _ superstring theory _ , ( cambridge : cambridge university press ) . | the perihelion precession of planetary orbits and the bending angle of null geodesics are estimated for different gravity theories in string - inspired models .
it is shown that , for dilaton coupled gravity , the leading order measure in the angle of bending of light comes purely from vacuum expectation value of the dilaton field which may be interpreted as an indicator of a dominant stringy effect over the curvature effect .
we arrive at similar results for spherically symmetric solution in quadratic gravity .
we also present the perihelion shift and bending of light in the einstein - maxwell - gauss - bonnet theory with special reference to the casimir effect and damour - polyakov mechanism .
numerical bounds to different coupling parameters in these models are estimated .
* keywords : * perihelion precession ; bending of light ; alternative gravity theories ; |
recently , d6-brane and orientifold 6-plane are of great interest@xcite in m - theory@xcite .
these @xmath0 dimensional objects in type iia theory are expressed as eleven dimensional smooth manifolds in m - theory , while other brane solutions generally have a singularity .
therefore , in low energy and strong coupling limit , their dynamics can be analyzed by means of eleven dimensional supergravity without informations of its microscopic physics .
the manifolds which express the d6-brane ( d6 ) and the orientifold 6-plane ( o6 ) are @xmath1 and @xmath2 , respectively@xcite , where @xmath3 is taub - nut manifold , @xmath4 is atiyah - hitchin manifold@xcite and @xmath5 is the @xmath0-dimensional flat minkowski space .
the d6-brane lying on the orientifold 6-plane ( o6+d6 ) corresponds to @xmath6 , where @xmath7 is covering space of @xmath4@xcite .
the purpose of this paper is to study the propagations of the gravitational wave or the d - particle on d6 , o6 or o6+d6 backgrounds in type iia theory with strong coupling .
we calculate the potential between a 6-plane and a d - particle and reproduce the known result . in the case of the o6 or the o6+d6 background
, we find the process in which gravitational wave colliding with the fixed plane convert into a d - particle . for eigenstates of orbital angular momentum
, we calculate the transition amplitude of this process numerically in the simplest case . as a result
, we obtain the amplitude of order one .
the manifolds @xmath9 , @xmath3 , @xmath4 and @xmath7 have some common properties ( @xmath10 is eguchi - hanson manifold , which is not used in this paper ) .
namely , they are hyper khler manifolds with @xmath8 ( or @xmath11 ) isometry .
such manifolds are studied in detail in @xcite and metrics of these manifolds are given explicitly . in this section ,
we review the properties of these manifolds which are necessary in the later sections .
topologically , these manifold can be expressed as @xmath12 , where @xmath13 is a group manifold of @xmath8 and @xmath14 is a half line ( fig.[xislm ] ) .
@xmath15 is a discrete subgroup of @xmath8 which we shall mention at the end of this section . at the end point of @xmath14
, @xmath13 shrinks to a point ( @xmath3 ) or two sphere ( @xmath10 , @xmath4 and @xmath7 ) .
( 85,5)@xmath14 ( 95,105)@xmath13 ( 170,10)@xmath16 ( -40,55)@xmath17 [ xislm ] each point on @xmath18 is specified by a pair @xmath19 of real number @xmath20 and an element @xmath21 , which parameterize @xmath14 and @xmath18 , respectively . to determine the @xmath8 invariant metric on @xmath18 , we introduce two mappings @xmath22 and @xmath23 from @xmath18 onto itself , @xmath24 and parameterize the tangent space @xmath25 at @xmath26 by the local coordinate @xmath27 defined by @xmath28 where @xmath29 is a representation matrix and @xmath30 is the generator of @xmath8 normalized as @xmath31=-\epsilon_{abc}t^{(s)}_c$ ] .
the index @xmath32 specifies the representation of @xmath8 . because @xmath23 and @xmath22 are commutable each other , this local coordinate system defined by @xmath23 is invariant under the @xmath22 rotation . therefore ,
if we take the @xmath33-independent metric @xmath34 in this local coordinate , the rotation @xmath22 is an isometry of @xmath13 . adding the radial parameter @xmath16 , which we define as the geodesic distance from the center @xmath17 , the metric on @xmath18 is given by @xmath35 we can diagonalize the metric @xmath36 by redefining @xmath37 : @xmath38 the three functions @xmath39 , @xmath40 and @xmath41 give the moving distances of the mappings @xmath42 , which are generated by @xmath43 , @xmath44 and @xmath45 , respectively . for the manifold @xmath18 to be a hyper khler manifold
, the connection should be anti - self - dual in some coordinate .
this requires the functions @xmath39 , @xmath40 and @xmath41 to satisfy the equations @xmath46 where @xmath47 for @xmath10 and @xmath48 for @xmath3 , @xmath4 and @xmath7 .
the functions @xmath39 , @xmath40 and @xmath41 for the manifolds @xmath3 , @xmath4 and @xmath7 , in which we are interested in this paper , have the common behavior in the asymptotic region ( @xmath49 ) : @xmath50 this behavior implies that the structure of @xmath18 in the asymptotic region is @xmath51 , with @xmath52 being the orbit of @xmath53 which will be identified with the @xmath54 direction of m - theory .
the radius of this @xmath52 is given by @xmath55 . because @xmath53 mixes the two directions @xmath56 and
@xmath57 , it becomes an isometry of @xmath18 when @xmath58 . in the asymptotic region ,
all of @xmath59 , @xmath60 and @xmath61 have this isometry , while it is broken in the central region for @xmath60 and @xmath61 .
if eq.([condabc ] ) is satisfied , @xmath18 is a hyper khler manifold with @xmath8 isometry , except for the point or the two - sphere at @xmath17 .
generically , this point or two - sphere is a conical singularity . to avoid such a singularity
, we should divide the manifold @xmath13 by a discrete subgroup @xmath62 , which is given in the table [ table ] ( @xmath63 ) . [ cols="^,^,^,^,^",options="header " , ] [ table ] we mentioned before that the compactification radius is given by @xmath55 .
however , we should take account of the action of @xmath15 on this @xmath52 : this circle should be divided by the discrete group @xmath64 . as a result
, we obtain the correct relation between @xmath41 and compactificaton radius @xmath65 , @xmath66 where @xmath67 for @xmath3 , @xmath7 and @xmath4 , respectively .
the bosonic fields in eleven dimensional supergravity are the metric @xmath68 and the three form field @xmath69 . to study the d - particle or gravitational wave scattering by d6 , o6 or o6+d6 in m - theory
, we should analyze the propagation of these fields on the manifold @xmath70 where @xmath18 is one of @xmath3 , @xmath4 and @xmath7 .
we label the directions of @xmath5 by @xmath71 and that of @xmath18 by @xmath72 .
since the wave propagation along the @xmath5 direction is trivial , the problem is reduced to solving the wave equation on the manifold @xmath18 . on this
four dimensional manifold , @xmath68 and @xmath69 with @xmath73 are regarded as scalar fields .
once the problem is solved for these scalar fields , we can get the solution for other modes ( vector , tensor and spinor ) by means of the supersymmetry .
therefore we may focus only on the scalar modes .
the problem we consider in this section is mathematically identical to what is investigated in detail by n.s.manton et al . in the context of monopole scattering@xcite .
it is known that we can construct the complete system of complex scalar functions on the group manifold from the representation matrices @xmath29 as follows : @xmath74 where the index @xmath32 labels the representations of the group . in the @xmath75 case we shall consider
, @xmath32 is an integer or a half integer , and @xmath76 and @xmath77 satisfy @xmath78 it is worth giving the physical interpretation of these indices here .
the indices @xmath32 and @xmath76 are quantum numbers for @xmath22 .
because @xmath22 is identified with the rotation of @xmath79 in the target space , these indices represent the orbital angular momentum of the d - particle or the gravitational wave .
the index @xmath77 is the charge of @xmath80 ( i.e. , the shift along @xmath52 of the @xmath54 direction ) , and therefore it is regarded as the d - particle number . from eq.([r11is ] ) , the relation between @xmath77 and the d - particle number @xmath81 is @xmath82 any scalar function @xmath83 on @xmath18 is expanded as @xmath84 and its derivative with respect to local coordinate @xmath37 is @xmath85|_{\epsilon^a=0}\nonumber\\ & = & \sum_s\tr[c^{(s)}(\rho)u^{(s)}(g)t^{(s)}_a].\end{aligned}\ ] ] laplacian on the manifold @xmath13 is given by @xmath86\nonumber\\ & = & \sum_s\tr\left[c^{(s)}(\rho)u^{(s)}(g)a^{(s)}(\rho)\right ] , { \label{laponm}}\end{aligned}\ ] ] where @xmath87 in ( [ laponm ] ) , we have used the property of the spin connection @xmath88 .
the wave equation on @xmath13 for the scalar field @xmath89 is @xmath90 where @xmath91 is the square of the momentum along the @xmath5 direction . expanding the field @xmath92 by ( [ expand ] ) , we get the equation for the coefficient @xmath93 : @xmath94_{mk}c^{(s)}_{kn}(\rho)=0 . { \label{waveeq}}\ ] ]
if we have @xmath58 , the matrix @xmath95 is diagonalized .
@xmath96\delta_{m , n}.\ ] ] for non - relativistic approximation , we rewrite the energy @xmath97 in ( [ waveeq ] ) as @xmath98 where @xmath99 is the d - particle mass , and neglect the @xmath100 term to get @xmath101c_{nn}^{(s)}(\rho)=0 . { \label{nonrela}}\ ] ] the potential between d - particle and 6-plane can be read off from eqs.([nonrela ] ) , ( [ asym ] ) and ( [ ndis ] ) : @xmath102 where @xmath103 is magnetic charge of the 6-plane . @xmath104
in this section , we shall consider the gravitational wave and the d - particle propagation on the o6 or the o6+d6 background in the m - theory framework .
these backgrounds correspond to @xmath60 and @xmath61 , respectively . on these manifolds ,
as we mentioned in the last section , the @xmath53 symmetry is broken .
this implies that the charge with respect to this symmetry ( @xmath105d - particle number ) is not conserved near the 6-plane .
the purpose of this section is to study this d - particle number changing process .
unfortunately , we can not solve the wave equation ( [ waveeq ] ) exactly .
therefore , we calculate the amplitude by means of numerical methods .
the orientifold flip @xmath106 changes the sign of the d - particle charge and this fact prevents us from defining it globally .
even if we consider the d - particle wave packet , it convert into anti - d - particle when it goes around the 6-plane .
this d - particle number changing process is not the one we mentioned above . in the process
we consider here , not only the sign of the d - particle number but also its absolute value changes . in solving eq.([waveeq ] ) , we should take account of the projection due to the discrete group @xmath62 .
namely , only @xmath15-invariant modes should be kept . in the o6+d6 case , the group @xmath107 is generated by @xmath108 .
because @xmath108 flips the sign of @xmath76 , the following condition should be satisfied .
@xmath109 ( we omit the indices @xmath32 and @xmath76 of @xmath110 . )
this implies that a state with a definite orbital angular momentum is a superposition of a d - particle state and an anti - d - particle state . in the o6 case ,
the group @xmath111 is generated by two elements @xmath112 and @xmath108 .
the matrix element of @xmath112 is @xmath113_{nm}=(-1)^m\delta_{m , n}.\ ] ] therefore , in addition to the condition ( [ flip ] ) , we should impose the requirement that @xmath76 is an even integer and @xmath32 is an integer . * eq.([waveeq ] ) does not contain planck length @xmath117 , and the dimensional parameters contained in the amplitude are @xmath65 ( compactification radius ) and @xmath97 ( energy ) . therefore , the amplitude depends only on one parameter @xmath118 .
* in the high energy limit ( @xmath119 ) , the wave propagation is described by geodesic orbits of a classical particle .
therefore , the amplitude is independent of the wavelength and it approaches to a constant . * because the wave equation ( [ waveeq ] ) for o6 and that for o6+d6 are identical , we have @xmath120 where eqs.([r11is ] ) and ( [ ndis ] ) have been used .
* because the non - zero elements of the matrix @xmath121 in eq.([waveeq ] ) are only @xmath122 and @xmath123 , the change of @xmath77 is restricted to an even integer . using ( [ ndis ] ) , we get a selection rule for the d - particle number .
@xmath124 namely , @xmath125 for o6 and @xmath126 for o6+d6 .
this is consistent with the relation ( [ mandm ] ) .
hereafter , we focus on the case of @xmath127 , which is the simplest case where the d - particle number changing process occurs . in this case
, @xmath128 has only two independent components @xmath129 and @xmath130 .
the metric of atiyah - hitchin manifold is @xmath131 where the functions @xmath132 , @xmath133 , and @xmath134 are given in @xcite in the parametric form with parameter @xmath135 : @xmath136 where @xmath137 , @xmath138 and @xmath139 is the complete elliptic integral .
the relation between the radial parameter @xmath140 and geodesic distance @xmath16 follows from ( [ ahmetric ] ) as @xmath141 . near
the bolt ( @xmath17 ) , @xmath7(@xmath4 ) is a plane bundle over @xmath146(@xmath147 ) .
if we parameterize @xmath13(@xmath148 ) by @xmath149 then @xmath150 parameterizes the base manifold @xmath146(@xmath147 ) , and @xmath151 gives a polar coordinate on the plane fiber .
since @xmath152 near the bolt , @xmath153 is almost an isometry of @xmath4 and it rotates the @xmath16-@xmath154 plane . for the @xmath127 modes , the eigenvalue of @xmath155 is @xmath156 ( mode a ) or @xmath157 ( mode b ) .
the @xmath154-dependence of these modes are @xmath158 and @xmath159 , respectively . to avoid singularity at @xmath17 ,
the @xmath16-dependence of these modes should be @xmath158 and @xmath160 .
( the cartesian coordinate on the @xmath16-@xmath154 plane is given by @xmath161 and the wave function of the mode b is a linear function of this coordinate . ) therefore if we adopt the generators @xmath162{${\mbox{\tiny$\sqrt{\frac{3}{2}}$}}$ } } \\ & { \makebox[0.7em]{${\mbox{\tiny$\sqrt{\frac{3}{2}}$}}$ } } & 0 & { \makebox[0.7em]{${\mbox{\tiny$\sqrt{\frac{3}{2}}$}}$ } } \\ & & { \makebox[0.7em]{${\mbox{\tiny$\sqrt{\frac{3}{2}}$}}$ } } & 0 & 1 \\ & & & 1 & 0 \end{array}\right),\quad t_y^{(2)}=\left(\begin{array}{ccccc } 0 & -i \\ i & 0 & { \makebox[0.7em]{$-i\!{\mbox{\tiny$\sqrt{\frac{3}{2}}$}}$ } } \\ & { \makebox[0.7em]{$i\!{\mbox{\tiny$\sqrt{\frac{3}{2}}$}}$ } } & 0 & { \makebox[0.7em]{$-i\!{\mbox{\tiny$\sqrt{\frac{3}{2}}$}}$ } } \\ & & { \makebox[0.7em]{$i\!{\mbox{\tiny$\sqrt{\frac{3}{2}}$}}$ } } & 0 & -i \\ & & & i & 0 \end{array}\right),\quad t_z^{(2)}=\left(\begin{array}{ccccc } { \makebox[0.7em]{$2 $ } } \\ & { \makebox[0.7em]{$1 $ } } \\ & & { \makebox[0.7em]{$0 $ } } \\ & & & { \makebox[0.7em]{$-1 $ } } \\ & & & & { \makebox[0.7em]{$-2 $ } } \end{array}\right),\ ] ] the boundary conditions for these two modes at @xmath17 are @xmath163 starting with the initial conditions ( [ initial1 ] ) and ( [ initial2 ] ) , we can calculate the wave functions in the asymptotic region @xmath49 numerically . from its behavior
, we obtain the transition amplitude between incoming gravitational wave and outgoing d - particle .
the result is shown in fig.[result ] .
100 a. sen , hep - th/9707042 , + _ dynamics of multiple kaluza - klein monopoles in @xmath166- and string theory_. a. sen , hep - th/9707123 , + _ a note on enhanced gauge symmetries in @xmath166- and string theory_. e. bergshoeff and b. janssen , hep - th/9706117 , + _ kaluza - klein monopoles and gauged sigma - model_. a.sen , hep - th/9708002 , + _ strong coupling dynamics of branes from m - theory_. y. imamura , hep - th/9706144 , + _ born - infeld action and chern - simons term from kaluza - klein monopole in m - theory_. e. witten , _ nucl . phys . _ * b443 * ( 1995 ) 85 , hep - th/9503124 , + _ string theory dynamics in various dimensions_. n.seiberg and e.witten , hep - th/9607163 , + _ `` gauge dynamics and compactification to three dimensions''_. m. f. atiyah and n. hitchin , + _ the geometry and dynamics of magnetic monopoles _ , + ( princeton university press , 1988 ) .
n.s.manton , _ phys . lett . _ * 110b * ( 1982 ) 54 , + _ a remark on the scattering of bps monopoles_. g.w.gibbons and n.s.manton , _ nucl .
* b274 * ( 1986 ) 183 , + _ classical and quantum dynamics of bps monopoles_. | we study the propagations of gravitational wave and d - particle on d6-brane and orientifold 6-plane backgrounds in the m - theory framework . in the case of orientifold plane , d - particle number is not conserved and gravitational wave can convert into d - particle . for the simplest case ,
we calculate its amplitude numerically . 22.5cm17.5cm-0.8cm-.9 cm = 6pt plus 2pt minus 1pt plus 0.2pt minus 0.1pt |
@xmath17 for any graph @xmath1 , we denote the set of its vertices by @xmath18 , and the set of its edges by @xmath19 , @xmath20 .
all the graphs in this paper are undirected , so @xmath21 implies @xmath22 . unless otherwise specified , our graphs are finite and may contain loops . for two graphs @xmath1 and @xmath2 , a * graph homomorphism * from @xmath1 to @xmath2 is a map @xmath23 , such that if @xmath24 are connected by an edge , then @xmath25 and @xmath26 are also connected by an edge .
we denote the set of all homomorphisms from @xmath1 to @xmath2 by @xmath27 .
[ dfhom ] @xmath0 is a polyhedral complex whose cells are indexed by all functions @xmath28 , such that if @xmath21 , then @xmath29 .
the closure of a cell @xmath30 consists of all cells indexed by @xmath31 , which satisfy @xmath32 , for all @xmath33 .
the set of vertices of @xmath0 is precisely @xmath27 .
since all cells of @xmath0 are products of simplices , the geometric realization of @xmath0 is defined in a straightforward fashion .
on the intuitive level , one can think of each @xmath34 , satisfying the conditions of the definition [ dfhom ] , as associating non - empty lists of vertices of @xmath2 to vertices of @xmath1 with the condition on this collection of lists being that any choice of one vertex from each list will yield a graph homomorphism from @xmath1 to @xmath2 .
a direct geometric construction of @xmath0 is as follows .
consider the partially ordered set @xmath35 of all @xmath30 as in definition [ dfhom ] , with the partial order defined by @xmath36 if and only if @xmath32 , for all @xmath37 .
then the order complex @xmath38 is a barycentric subdivision of @xmath0 .
a cell @xmath39 of @xmath0 corresponds to the union of all the simplices of @xmath38 labeled by the chains with the maximal element @xmath39 . in this paper
we study properties of the complexes @xmath0 .
more specifically we compute the homotopy type of @xmath0 for several families of @xmath1 and @xmath2 and also derive some information about natural finite group actions on these complexes .
@xmath17 a particularly frequently studied special case of a graph homomorphism is that of a vertex coloring : for a graph @xmath1 a vertex coloring of @xmath1 with @xmath40 colors is simply a graph homomorphism from @xmath1 to @xmath41 . here
@xmath41 denotes an unlooped complete graph on @xmath40 vertices , that is @xmath42 $ ] , @xmath43 , x\neq y\}$ ] .
historically , one was especially interested in the question of existence of vertex colorings with a specified number of colors . from this point of view
, the minimal possible number of colors in a vertex coloring is of special importance .
it is called the _ chromatic number _ of the graph , and is denoted by @xmath44 .
the kneser conjecture was posed in 1955 , see @xcite , and concerned chromatic numbers of a specific family of graphs , later called _ kneser graphs_. for @xmath45 , @xmath46 , @xmath47 , the kneser graph @xmath48 is the graph whose vertices are all @xmath49-subsets of @xmath50 $ ] , and edges are all pairs of disjoint @xmath49-subsets ; here @xmath51 . in 1978
l. lovsz solved the kneser conjecture by finding geometric obstructions of borsuk - ulam type to the existence of graph colorings .
( kneser - lovsz , @xcite ) .
@xmath52 . to show the inequality @xmath53 lovsz associated a simplicial complex @xmath54 , called the _ neighborhood complex _ , to an arbitrary graph @xmath1 , and then used the connectivity information of the topological space @xmath54 to find obstructions to the colorability of @xmath1 .
[ lothm ] ( lovsz , @xcite ) .
_ let @xmath1 be a graph , such that @xmath54 is @xmath49-connected for some @xmath55 , @xmath6 , then @xmath56 . _ the main topological tool which lovsz employed was the borsuk - ulam theorem .
see also @xcite for the extension to hypergraphs , which used the generalization of the borsuk - ulam theorem from @xcite
. we shall define the complex @xmath54 in section [ s4 ] , where we shall also see that for any graph @xmath1 the complex @xmath54 is homotopy equivalent to @xmath57 .
this fact leads one to consider the family of @xmath58 complexes as a natural context in which to look for further obstructions to the existence of graph homomorphisms .
accordingly , lovsz has made the following conjecture , @xcite .
let @xmath59 be a cycle with @xmath60 vertices : @xmath61 , @xmath62 .
[ loconj ] ( lovsz ) .
_ let @xmath1 be a graph , such that @xmath63 is @xmath49-connected for some @xmath64 , @xmath65 , @xmath6 , then @xmath66 .
_ our proof of conjecture [ loconj ] was announced in @xcite . the full version of the proof consists of 2 parts : the study of the important general properties of @xmath58 complexes , appearing in this paper , and more detailed and specific spectral sequence calculations , appearing in @xcite .
however , the general study undertaken in this paper contains more than the results which we later use for our spectral sequence computations .
there was a more general conjecture , also due to lovsz .
[ loconj2 ] ( lovsz ) .
_ let @xmath67 , @xmath1 be two graphs , then @xmath68 _ here , @xmath69 is the connectivity of the topological space @xmath70 , i.e. , the maximum @xmath49 such that @xmath70 is @xmath49-connected .
note that conjecture [ loconj ] is the special case of the conjecture [ loconj2 ] corresponding to @xmath71 .
hoory and linial , @xcite , gave a counterexample to the conjecture [ loconj2 ] .
in their counterexample @xmath72 , and @xmath67 is a graph with 9 vertices and 22 edges .
furthermore , @xmath73 , but @xmath74 is connected , showing that the equation is false in general . in this paper
we show that conjecture [ loconj2 ] is true for @xmath75 ( lovsz himself proved for @xmath76 ) .
more specifically , let @xmath9 act on @xmath10 for @xmath5 , by swapping the vertices 1 and 2 and fixing the vertices @xmath77 .
since the graph homomorphism flips an edge , it induces a free @xmath9-action on @xmath78 , for an arbitrary graph @xmath1 without loops . for a cw complex @xmath70
on which @xmath9 acts freely let @xmath79 denote its _ first stiefel - whitney class _ , see @xcite . as a corollary of our computations
, we prove the following theorem .
[ cothm ] let @xmath1 be a graph , and let @xmath80 , such that @xmath5 , @xmath81 .
if @xmath7 , then @xmath8 . note that if a @xmath9-space @xmath70 is @xmath49-connected , then there exists a @xmath9-map @xmath82 .
the functoriality of stiefel - whitney classes and the fact that @xmath83 imply that @xmath84 .
therefore theorem [ cothm ] implies conjecture [ loconj2 ] for @xmath75 .
@xmath17 in section 2 we define notations , describe the category of graphs and graph homomorphisms , and give several examples of @xmath58 complexes
. furthermore , we list many simple , but fundamental properties of the @xmath58 construction . in section 3
we describe two results from topological combinatorics which we need for our arguments : a proposition from discrete morse theory , and a quillen - type result . in section 4
we see first that in general @xmath57 is homotopy equivalent to the neighborhood complex @xmath54 , implying in particular that @xmath85 .
we observe that in fact @xmath86 is a boundary complex of a polytope , on which the natural @xmath9-action on the first argument , induces an antipodal action . in the subsection [ ss4.3 ]
we prove the central result of this section , namely we show that @xmath3 is homotopy equivalent to a wedge of @xmath4-dimensional spheres , and provide an enumeration formula for the number of the spheres . as a corollary we derive the theorem [ cothm ] .
in section 5 we prove that a fold in the first argument of @xmath0 induces a homotopy equivalence . as a corollary
, we show for an arbitrary forest @xmath15 , that @xmath12 is homotopy equivalent to a direct product of @xmath13-dimensional spheres , while @xmath14 is homotopy equivalent to a wedge of spheres . for an arbitrary tree @xmath67 with @xmath9-action we describe the @xmath9-homotopy type of @xmath87 . to conclude the introduction , we refer the reader to a recent survey of the previous studies of other complexes related to graph colorings , see @xcite .
* acknowledgments .
* we would like to thank lszl lovsz for insightful discussions , sonja uki for a remark , and the anonymous referee for useful suggestions which helped us to improve the presentation significantly .
@xmath17 @xmath88 for a graph @xmath1 we distinguish between looped and unlooped complements , namely we let @xmath89 be the graph defined by @xmath90 while @xmath91 is the graph defined by @xmath92 @xmath88 for a graph @xmath1 and @xmath93 we denote by @xmath94 $ ] the graph on the vertex set @xmath95 induced by @xmath1 , that is @xmath96)=s$ ] , @xmath97)=(s\times s)\cap e(g)$ ] .
for @xmath93 we set @xmath98 to be the graph @xmath99 $ ] .
for @xmath33 we shall sometimes simply write @xmath100 instead of @xmath101 .
@xmath88 for a graph @xmath1 and @xmath102 , let @xmath103 denote the set of all common neighbors of the vertices of @xmath104 . in particular , @xmath105 , and @xmath106 is simply the set of all neighbors of @xmath107 , with the convention being that @xmath107 is its own neighbor if and only if @xmath108 .
if needed , we will also specify the graph by writing @xmath109 .
@xmath88 for two arbitrary graphs @xmath1 and
@xmath2 we let @xmath110 denote the _ direct product _ of @xmath1 and @xmath2 : @xmath111 @xmath88 for two arbitrary graphs @xmath1 and @xmath2 we let @xmath112 denote the _ disjoint union _ of @xmath1 and @xmath2 .
@xmath88 for @xmath113 , @xmath114 , we let @xmath115 denote the graph defined by @xmath116 $ ] , @xmath117 .
@xmath88 let @xmath118 be the graph defined by @xmath119 $ ] , @xmath120 .
@xmath88 for an arbitrary graph @xmath1 , we let @xmath121 denote the _
loop completion _ of @xmath1 , that is @xmath122 , @xmath123 .
@xmath88 for a polyhedral complex @xmath124 we let @xmath125 denote its _ face poset _ , that is a partially ordered set of the faces ordered by inclusion .
@xmath88 for any finite category @xmath126 ( in particular a finite poset ) we denote by @xmath127 the realization of the nerve of that category .
@xmath88 for a poset @xmath128 we let @xmath129 denote the barycentric subdivision of @xmath128 , that is the poset of all the chains in the given poset ordered by inclusion . for a polyhedral complex @xmath124 we let @xmath130 denote the barycentric subdivision of @xmath124 . clearly , @xmath131 , and @xmath132 .
@xmath88 for any finite poset @xmath128 , we let @xmath133 denote the finite poset which has the same set of elements as @xmath128 , but the opposite partial order . also , for any finite poset @xmath128 , whenever a subset of the elements of @xmath128 is considered as a poset , the partial order is taken to be induced from @xmath128 . @xmath88
* top * is a category having topological spaces as objects , and continuous maps as morphisms .
@xmath17 it is an easy check that a composition of two graph homomorphisms is again a graph homomorphism .
we denote a composition of @xmath134 and @xmath135 by @xmath136 .
since the composition is associative and since for any graph @xmath1 we have a unique identity homomorphism in @xmath137 we can define a category * graphs * as the one having graphs as objects , and graph homomorphisms as morphisms .
one can check that the direct product of graphs is a categorical _ product _ in * graphs * , while the disjoint union of graphs is a categorical _
coproduct _ in * graphs*. note that with the above notations @xmath138 is a graph consisting of one vertex and one loop , it is the _
terminal object _ of * graphs*. the empty graph is the _ initial object _ of * graphs*. @xmath17 [ ss2.3 ] to start with , we have various trivial cases : @xmath88 @xmath139 is a simplex with @xmath140 vertices ; @xmath88 @xmath141 , unless @xmath142 , in which case @xmath143 is a point ; @xmath88 more generally , @xmath144 if @xmath145 ; @xmath88 @xmath146 is a simplex with vertices indexed by the looped vertices of @xmath2 ; @xmath88 @xmath147 is a point , as mentioned above ; @xmath88 @xmath148 is a direct product of @xmath149 simplices , each simplex having @xmath40 vertices ; @xmath88 @xmath150 if @xmath1 is not bipartite ; it consists of @xmath151 points , if @xmath1 bipartite and has @xmath152 connected components ; @xmath88 @xmath153 if and only if @xmath154 ; @xmath88 @xmath155 is a disjoint union of @xmath156 points , for @xmath157 ; @xmath88 @xmath158 is a disjoint union of two cycles , each of length @xmath159 , for @xmath160 ; @xmath88 @xmath161 consists of 6 isolated points , 6 solid cubes and 18 squares connected as shown on the figure [ fig : homc6k3 ] .
the left part of figure [ fig : homc6k3 ] is incomplete for the purpose of visualizing , it shows the 6 points , 6 cubes and some of the squares .
the right part shows the link of each of the 6 vertices , where two of the cubes touch .
the closed star of such a vertex consists of 2 solid cubes and 3 squares .
@xmath88 @xmath162 is a disjoint union of @xmath163 points ; @xmath88 @xmath164 is the cayley graph of @xmath165 with the set of generators consisting of @xmath166 transpositions @xmath167 .
indeed , every vertex of @xmath164 is an injection @xmath168{\rightarrow}[n]$ ] , which can be identified with a permutation of @xmath50 $ ] by writing out the values of @xmath169 and then writing the missing element of @xmath50 $ ] in the last position .
an edge is a changing of one arbitrary value of @xmath169 , say @xmath170 , to the missing value , which is precisely the same as acting with the transposition @xmath171 on the corresponding permutation .
@xmath88 @xmath172 is the full 2-skeleton of the 3-cell depicted on figure [ fig : hom3 ] .
@xmath88 @xmath173 is homeomorphic to a disjoint union of two mbius bands .
the local structure of each mbius band is shown on the figure [ fig : hom3 ] .
the middle cycle which is painted bold has length 21 in each band , and all visible squares on the picture are filled with 2-cells .
@xmath88 it is not difficult to count the number of connected components of @xmath174 . denote this number @xmath175 , the general formula is @xmath176 for @xmath177 . the crucial fact for deriving this formula for @xmath175 is to notice that the connected components of @xmath174 are indexed with the number of times @xmath178 can wind around the triangle @xmath179 , with the sign encoding the direction . as the examples above show , for the considered values of @xmath60 and @xmath40 , the spaces @xmath180 are either empty or consist of several connected components , with each component either being a point , or homotopy equivalent to a circle .
recently , this fact has been proved for all values of @xmath60 and @xmath40 , see @xcite .
@xmath88 let @xmath181 denote the stiefel manifold of all orthonormal @xmath49-frames in @xmath182 .
csorba , @xcite , has made the following conjecture : ( csorba ) .
@xmath183 is homeomorphic to @xmath184 , for all @xmath114 .
the cases @xmath185 are tautological , as both spaces are empty .
the example above verifies the case @xmath186 : @xmath187 .
several cases , including @xmath188 have been recently verified by csorba and lutz , @xcite .
@xmath88 note that for an arbitrary @xmath1 , @xmath189 can be interpreted as a cubical cone over the independence complex of @xmath1 ; recall that the independence complex of @xmath1 is the simplicial complex consisting of all independent sets of @xmath1 . when saying _ cubical cone _ we mean the following construction : given an arbitrary simplicial complex @xmath190 , add an extra vertex @xmath191 , and for each simplex @xmath192 with @xmath193 vertices span a @xmath193-dimensional cube @xmath194 with @xmath191 being a vertex of @xmath194 and @xmath195 forming the link of @xmath191 in @xmath194 . note that @xmath196 is cubical for any graph @xmath1 having no isolated vertices .
by a theorem of gromov , see @xcite , @xmath196 allows metric with nonpositive curvature if and only if the link of every vertex is a flag complex ( which means that each link is the clique complex of its 1-skeleton ) . for any @xmath197
, we say that @xmath198 has a cubical neighborhood if @xmath198 does not belong to any simplex with more than 2 vertices .
[ cubprop ] if @xmath197 has a cubical neighborhood , then @xmath199 is a flag complex .
* set @xmath200 , i.e. , the simplicial complex whose face poset is @xmath201 . for @xmath33 ,
set @xmath202 since @xmath198 has a cubical neighborhood , we have @xmath203 , for any @xmath33 .
let @xmath204 be the set of all vertices @xmath107 with @xmath205 .
clearly @xmath206 has @xmath207 as the set of vertices .
furthermore , @xmath208 , such that @xmath209 , is a simplex in @xmath206 , if and only if , for any two @xmath210 , and any @xmath211 , @xmath212 , we have @xmath213 . since this is a local condition depending only on the pair @xmath214 , we conclude that @xmath206 is a flag complex .
it follows that the cubical complex @xmath196 always allows a metric with nonpositive curvature .
moreover , for any @xmath215 , the proof of the proposition [ cubprop ] yields that @xmath216 is the independence complex of @xmath217 $ ] .
@xmath17[ss2.4 ] \(1 ) for any two graphs @xmath1 and @xmath2 , @xmath0 is a regular cw complex .
\(2 ) cells of @xmath0 are direct products of simplices .
more specifically , each @xmath30 as in the definition [ dfhom ] is a product of @xmath149 simplices , having dimensions @xmath218 , for @xmath219 .
thus @xmath220 .
\(3 ) for any three graphs @xmath1 , @xmath2 , and @xmath124 , we have @xmath221 and , if @xmath1 is connected , and @xmath222 , then also @xmath223 where the equality denotes isomorphism of polyhedral complexes .
the first formula is obvious . to see the second one , note that for @xmath224 , and @xmath24 , such that @xmath21 , if @xmath225 , then @xmath226 , which under assumptions on @xmath1 implies that @xmath227 .
\(4 ) @xmath228 is a covariant , while @xmath229 is a contravariant functor from * graphs * to * top*. if @xmath230 , then we shall denote the cellular maps induced by composition as @xmath231 and @xmath232 .
\(5 ) the map induced by composition @xmath233 is a topological map .
\(6 ) obviously , it is difficult to decide in general whether @xmath234 is non - empty , let alone @xmath49-connected .
it is certainly non - empty if the valency of each vertex is at most @xmath166 .
the following fact is true in general .
[ pr : conn ] let @xmath1 be any graph
. if the maximal valency of @xmath1 is equal to @xmath193 , then @xmath234 is connected , for all @xmath235 . * proof . * assume @xmath234 is not connected .
choose @xmath236 , such that @xmath237 and @xmath238 belong to different connected components , and @xmath239 for the maximal possible number of vertices .
pick @xmath240 , such that @xmath241 .
if @xmath242 can not be changed to @xmath243 , that is , if @xmath244 , defined by @xmath245 for @xmath246 , @xmath247 , is not a graph homomorphism , then there exists a vertex @xmath248 , such that @xmath249 , and @xmath250 .
since the valency of @xmath248 is at most @xmath251 , we can change @xmath252 to something else , without changing the number of vertices on which @xmath237 and @xmath238 coincide .
once this is done for each such neighbor of @xmath240 , we can change @xmath242 to @xmath243 , thereby increasing the number of vertices on which @xmath237 and @xmath238 coincide , hence obtaining a contradiction to the choice of @xmath237 and @xmath238 .
this result motivates the following conjecture .
let @xmath1 be any graph .
if the maximal valency of @xmath1 is equal to @xmath193 , then @xmath234 is @xmath49-connected , for all integers @xmath6 , @xmath253 .
proposition [ pr : conn ] corresponds to the case @xmath254 ; the case @xmath255 is obviously true since the graph of maximal valency @xmath193 can be colored with @xmath256 colors .
@xmath17 for a poset @xmath128 with the covering relation @xmath257 , we define a _ partial matching _ on @xmath128 to be a set @xmath258 , and an injective map @xmath259 , such that @xmath260 , for all @xmath261 .
the elements of @xmath262 are called critical .
the next proposition is a special case , which will be sufficient for our purposes , of a more general result proved by r. forman , see @xcite .
[ dmt ] let @xmath190 be a regular cw complex and @xmath263 a subcomplex of @xmath190 , then the following are equivalent : \a ) there is a sequence of collapses leading from @xmath190 to @xmath263 ; \b ) there is a partial matching @xmath264 on @xmath265 with the set of critical cells being @xmath266 , such that there is no sequence @xmath267 , @xmath268 , such that @xmath269 ( such matching is called acyclic ) .
* see ( * ? ? ?
* proposition 5.4 ) .
proposition [ dmt ] is a part of the discrete morse theory ; @xcite are just some of the references where it has been studied and used .
@xmath17 in this subsection we prove a quillen - type result which , given a poset map @xmath238 satisfying certain conditions , provides us with some topological information about the induced simplicial map @xmath270 . [ prop_abc ] let @xmath271 be a map of finite posets .
consider a list of possible conditions on @xmath238 .
condition @xmath272 .
for every @xmath273 , @xmath274 is contractible .
condition @xmath275 .
for every @xmath276 and @xmath273 with @xmath277 the poset @xmath278 has a maximal element . in this case
we denote this maximal element by @xmath279 . condition @xmath280 .
let @xmath281 be the poset map induced by @xmath238 .
we require that @xmath282 satisfies condition @xmath283 . in this case
we denote the minimal element of @xmath284 by @xmath285
. then 1 . if @xmath238 satisfies @xmath272 and either @xmath275 or @xmath280 , then @xmath286 is a homotopy equivalence .
if @xmath238 satisfies @xmath275 and @xmath280 , and @xmath287 is connected , then for any @xmath288 we have @xmath289 .
furthermore , we have a fibration homotopy long exact sequence : @xmath290 * proof . *
consider the poset map @xmath291 , which maps @xmath292 , @xmath293 to @xmath294 . since @xmath238 is order - preserving , the last set is totally ordered , and thus can be interpreted as a chain in @xmath287 .
we set @xmath295 and view it as a subposet of p. note that @xmath296 first we show ( 1 ) . because of the symmetry , we restrict our consideration to the case when @xmath238 satisfies conditions @xmath272 and @xmath280 . by quillen s theorem
a , see @xcite , it is enough to show that @xmath297 is contractible for any @xmath298 . by it
is enough to show that @xmath299 is contractible for any @xmath298 .
we use induction on the length of the chain @xmath300 .
when @xmath301 , this is precisely condition @xmath272 , so we assume that @xmath268 .
define @xmath302 , by @xmath303 , for @xmath304 .
this is well - defined since @xmath305 .
note that 1 .
@xmath306 , since @xmath307 ; 2 .
@xmath308 , by the definition of @xmath309 ; 3 .
@xmath310 is order - preserving .
indeed , take @xmath311 , such that @xmath312 .
then , on one hand @xmath313 , on the other hand @xmath314 , hence , by the definition of @xmath315 , we have @xmath316 .
this means that @xmath310 is a closure map , hence @xmath317 is homotopy equivalence , see ( * ? ? ?
* corollary 10.12 ) .
it follows by induction that @xmath299 is contractible for any @xmath298 .
next we prove ( 2 ) .
let @xmath318 , such that @xmath319 .
we want to show that the inclusion map @xmath320 induces a homotopy equivalence of the order complexes .
. then @xmath322 , @xmath323 , and @xmath324 .
consider the sequence of inclusion maps @xmath325 , and let @xmath326 be the map defined above . by the argument for the part ( 1 ) we know that pairs @xmath327 and @xmath328 induce homotopy equivalences of the order complexes .
it follows that the pair @xmath329 also induces a homotopy equivalence , since @xmath330 and @xmath331 by a symmetric argument the inclusion map
@xmath332 induces a homotopy equivalence as well . composing
, we get that @xmath333 is a homotopy equivalence . in the special case
@xmath334 we get that @xmath335 hence , since @xmath287 is connected as a poset , we get @xmath336 for any @xmath288 . finally , the existence of the fibration homotopy long exact sequence ( [ eq : piseq ] ) follows from ( [ eq : bdgam ] ) and quillen s theorem b , see @xcite
. we shall not use proposition [ prop_abc ] ( 2 ) in this paper .
we have proved it here as a result which is interesting on its own right and might be useful for other computations .
a more general version of proposition [ prop_abc ] ( 1 ) was proved in @xcite , see also @xcite .
@xmath17 we are now ready to define the neighborhood complex @xmath54 and show that it is homotopy equivalent to @xmath57 .
the natural advantage to working with the polyhedral complex @xmath57 instead of the simplicial complex @xmath54 is that @xmath57 possesses a natural free cellular @xmath9-action induced from the swapping @xmath9-action on @xmath337 . for an arbitrary graph @xmath1 the simplicial complex @xmath54 is defined as follows :
its vertices are all non - isolated vertices of @xmath1 , and its simplices all the subsets of @xmath18 which have a common neighbor . in other words ,
the maximal simplices of @xmath54 are @xmath338 , for @xmath33 .
[ propn ] @xmath57 is homotopy equivalent to @xmath54
. * proof .
* let @xmath339 and @xmath340 .
consider @xmath341 mapping the element @xmath342 to @xmath343 .
clearly , the vertices in @xmath344 have all the vertices in @xmath345 as their neighbors , hence , since @xmath346 , @xmath238 is well - defined .
let us show that @xmath238 induces homotopy equivalence @xmath347 .
first , let @xmath348 .
we see that @xmath349 is the set of all pairs @xmath350 , @xmath351 , such that for all @xmath352 , and @xmath353 , we have @xmath21 .
clearly , @xmath349 has a maximal element @xmath354 , so @xmath355 is a cone , hence contractible .
second , let us check the condition @xmath275 of the proposition [ prop_abc ] .
let @xmath348 and @xmath356 , such that @xmath357 .
. then @xmath359 .
this poset has a maximal element @xmath360 , since @xmath361 . in the notations of the proposition [ prop_abc ] we have @xmath362 . since conditions @xmath272 and @xmath275 are satisfied
, @xmath286 is a homotopy equivalence by proposition [ prop_abc ] .
this shows that @xmath363 , hence the result .
as proposition [ propn ] shows , the original complexes @xmath54 correspond to @xmath337-type obstructions to colorability .
the lovsz idea behind his conjecture [ loconj ] was that the next natural class of obstructions should come from the maps from odd cycles @xmath364 to our graph .
@xmath17 let @xmath365 denote the minkowski sum @xmath366^n+[(-1/2,-1/2,\dots,-1/2),(1/2,1/2,\dots,1/2)],\ ] ] where @xmath367^n$ ] denotes the cube in @xmath182 with vertices are all points whose coordinates have the absolute value @xmath368 .
@xmath365 is a zonotope in @xmath182 .
its dual , @xmath369 , is the polytope associated to the hyperplane arrangement @xmath370 defined by @xmath371 in the proof of the next proposition we identify each cell @xmath372 with the ordered pair @xmath350 of non - empty subsets of @xmath50 $ ] , by taking @xmath373 and @xmath374 .
[ polyprop ] @xmath375 is isomorphic as a cell complex to the boundary complex of @xmath369 .
the @xmath9-action on @xmath375 , induced by the flip action of @xmath9 on @xmath337 , corresponds under this isomorphism to the central symmetry .
* proof . *
set @xmath376 .
we shall see that @xmath128 is isomorphic to the face poset of @xmath365 , which we denote by @xmath377 .
we shall denote the future isomorphism by @xmath378 .
first , note that faces of the cube @xmath367^n$ ] are encoded by @xmath40-tuples of @xmath368 , @xmath379 , and @xmath380 , where @xmath380 denotes the coordinate where the value can be chosen arbitrarily from the interval @xmath367 $ ] .
for an arbitrary @xmath40-tuple @xmath381 , we let @xmath382 $ ] denote the set of the indices of coordinates which are either non - zero , or are denoted with a @xmath380 . additionally , for an arbitrary number @xmath49 , we let @xmath383 $ ] denote the set of the indices of the coordinates which are equal to @xmath49 ( in particular , they can not be denoted with a @xmath380 ) .
vertices of @xmath365 are labeled by all @xmath40-tuples of @xmath384 , @xmath385 , and @xmath386 , such that @xmath384 and @xmath385 are not present simultaneously , and not all the coordinates are equal to @xmath386 , that is @xmath107 is a vertex of @xmath365 if and only if @xmath387 , or @xmath388 , and @xmath389 .
these vertices correspond to atoms in @xmath128 as follows : @xmath390{\setminus}{\text{\rm supp}\,}(v)),&\text { if } v\in\{0,1\}^n ; \\ ( [ n+1]{\setminus}{\text{\rm supp}\,}(v),{\text{\rm supp}\,}(v)),&\text { if } v\in\{0,-1\}^n . \end{cases}\ ] ] clearly , restricted to atoms , @xmath378 is a bijection .
those faces of @xmath365 which are contained in the closed star of @xmath391 can be indexed by @xmath392 , where @xmath393 .
symmetrically , those faces of @xmath365 which are contained in the closed star of @xmath394 can be indexed by @xmath395 , where @xmath396 .
for these faces @xmath378 can be defined as follows : @xmath397 finally , we consider the faces of @xmath365 which are not in @xmath398 .
each such face is a convex hull of the union of two faces , @xmath399 , such that @xmath400 , @xmath401 , with the condition that @xmath402 , @xmath403 .
the element of @xmath128 associated to such a face under @xmath378 is @xmath404 .
it is an easy exercise to check that @xmath378 defines a poset isomorphism between @xmath128 and @xmath287 , which in turn induces the required cell complex isomorphism .
finally , a brief scanning through the definition of @xmath378 in different cases reveals that @xmath378 is equivariant with respect to the described @xmath9-actions on both sides .
hence the last part of the proposition follows .
the cellular map @xmath238 defined in the proposition [ propn ] , is in this case going from the boundary of an @xmath13-dimensional polytope @xmath369 to the boundary of an @xmath13-dimensional simplex
. it would be interesting to see whether it has interesting additional properties in the context of zonotopes and also to find out what other graphs @xmath1 provide a connection to polytopes .
@xmath3 can be viewed as a deleted product of simplices , see @xcite . in this context
it is well - known , probably due to van kampen , that for @xmath405 it is a boundary of a polytope .
@xmath17 [ ss4.3 ] we can still get a fairly detailed information about the topology of the spaces of homomorphisms between complete graphs in general . [ pr_chom ] @xmath3 is homotopy equivalent to a wedge of @xmath4-dimensional spheres . * proof .
* we use induction on @xmath60 and on @xmath406 .
the base is provided by the cases @xmath407 , which is a simplex with @xmath40 vertices , hence contractible , and @xmath162 , which consists of @xmath163 points , that is a wedge of @xmath408 spheres of dimension @xmath386 .
we assume now that @xmath5 and @xmath409 .
for @xmath410 $ ] let @xmath411 be the subcomplex of @xmath3 defined by : @xmath412{\rightarrow}2^{[n]}{\setminus}\{\emptyset\}\,|\ , n\notin\eta(j ) , \text { for } j\in[m],j\neq i\}.\ ] ] since any two vertices of @xmath10 are connected by an edge , @xmath40 can not be in @xmath413 , for @xmath414 .
this implies that @xmath415 .
clearly , for any @xmath416 , @xmath417 $ ] , we have @xmath418{\rightarrow}2^{[n]}{\setminus}\{\emptyset\}\,|\ , n\notin\eta(k),\text { for all } k\in[m]\},\ ] ] so @xmath419 is isomorphic to @xmath420 , hence , by induction , it is @xmath421-connected .
we shall now see that each @xmath411 is @xmath422-connected .
since all @xmath411 s are isomorphic to each other , it is enough to consider @xmath423 .
let us describe a partial matching on @xmath424 . for @xmath425 , such that @xmath426 , we set @xmath427 , defined by : @xmath428
obviously , this is an acyclic matching and the critical cells form a subcomplex @xmath429 defined by : @xmath430 in and only if @xmath431 . thus @xmath432 .
since , by the proposition [ dmt ] @xmath433 is homotopy equivalent to @xmath423 , and @xmath433 is @xmath422-connected by the induction assumption , we conclude that @xmath411 is @xmath422-connected for any @xmath434 .
it follows from ( * ? ? ?
* theorem 10.6(ii ) ) that @xmath3 is @xmath422-connected .
since dimension of @xmath3 is @xmath406 , it follows from ( * ? ? ? * ( 9.19 ) ) that @xmath3 is homotopy equivalent to a wedge of spheres .
one can use the construction in the proof of the proposition [ pr_chom ] to count the number of the spheres in the wedge .
let us say that @xmath3 is homotopy equivalent to a wedge of @xmath435 spheres .
let @xmath436 denote the stirling numbers of the second kind , and @xmath437 denote the generating function for these numbers .
it is well - known that @xmath438 for @xmath439 , let @xmath440 be the generating function for the number of the spheres .
clearly , @xmath441 , and @xmath442 . the numbers @xmath435 satisfy the following recurrence relation @xmath443 for @xmath444 ; with the boundary values @xmath445 , @xmath446 for @xmath114 , and @xmath447 for @xmath448 . the generating function @xmath449 is given by the equation : @xmath450 as a consequence , the following non - recursive formulae are valid : @xmath451 and @xmath452 for @xmath453 .
* let @xmath454 denote the non - reduced euler characteristics of the complexes @xmath3 , and , for @xmath455 , let @xmath411 be as in the proof of the proposition [ pr_chom ] . since @xmath456 , @xmath457 , for all @xmath416 , and @xmath458 , for @xmath410 $ ] , by simple inclusion - exclusion counting we conclude that @xmath459 for @xmath444 , additionally @xmath460 , @xmath461 , for @xmath114 .
since @xmath462 , a simple computation shows the validity of the relation . for @xmath439 ,
let @xmath463 .
multiplying each side of the equation by @xmath464 and summing over all @xmath40 yields @xmath465 , implying @xmath466 for @xmath439 , and hence , since @xmath467 , we get @xmath468 for @xmath469 . by multiplying the identity @xmath470 with @xmath464 and summing over all @xmath471 , we get @xmath472 follows from comparing the coefficients in . to prove we see that it fits the boundary values and satisfies the recurrence relation .
verifying is straightforward , as is checking for @xmath473 and @xmath405 .
finally , is seen for @xmath474 by expanding the expression @xmath475 by the binomial theorem and comparing the coefficient of @xmath464 on both sides of the expansion . in particular , we have @xmath476 , for @xmath46 , @xmath477 , for @xmath478 , @xmath479 , for @xmath480 , @xmath481 , for @xmath482 .
we are now ready to prove the result announced in the beginning of this paper .
* proof of the theorem [ cothm ] .
* if the graph @xmath1 is @xmath483-colorable , then there exists a homomorphism @xmath484 .
it induces a @xmath9-equivariant map @xmath485 by the proposition [ pr_chom ] the space @xmath486 is homotopy equivalent to a wedge of @xmath487-spheres , hence , by dimensional reasons , @xmath488 .
since the stiefel - whitney classes are functorial , the existence of the map @xmath489 implies that @xmath490 , which is a contradiction to the assumption of the theorem .
@xmath17 the next proposition , coupled with propositions [ polyprop ] and [ pr_chom ] , will be our workhorse for computing concrete examples .
[ prop5.1 ] if @xmath1 and @xmath2 are graphs and @xmath240 and @xmath107 are distinct vertices of @xmath1 , such that @xmath491 , then the inclusion @xmath492 , resp .
the homomorphism @xmath493 mapping @xmath107 to @xmath240 and fixing other vertices , induce homotopy equivalences @xmath494 , resp .
@xmath495 . * proof .
* let us apply the proposition [ prop_abc ] ( 1 ) for the cellular map @xmath494 .
take @xmath496 , @xmath497 .
we have @xmath498 an element in @xmath499 is determined by its value on @xmath107 .
take @xmath500 such that @xmath501 clearly , @xmath39 is the maximal element of @xmath499 , hence @xmath502 is contractible , so the condition @xmath272 is satisfied .
let us now check the condition @xmath275 .
take @xmath503 , @xmath496 , such that for any @xmath504 we have @xmath505 .
the set @xmath506 consists of all @xmath507 , such that for any @xmath219 we have @xmath508 , and for any @xmath504 we have @xmath509 .
thus , it has a maximal element defined by : @xmath510 conditions @xmath272 and @xmath275 being satisfied , we now get that @xmath511 , hence also @xmath512 , is a homotopy equivalence . to see that @xmath513 is also a homotopy equivalence note first that @xmath514 .
let @xmath515 be the homotopy inverse of @xmath512 , then @xmath516 .
if @xmath1 is a graph , and @xmath517 , @xmath518 , such that @xmath491 , then we say that @xmath1 _ reduces _ to @xmath100 .
we shall also say that @xmath240 _ dominates _
@xmath107 , or that @xmath107 is _ dominated _ by @xmath240 . if in addition @xmath519 we say that @xmath240
_ strongly dominates _
we call @xmath240 and @xmath107 _ equivalent _ if @xmath520 . the strong domination defines a partial order @xmath521 on the set of equivalence classes
. we call a graph _ irreducible _ if it does not reduce to any subgraph .
we note a simple , but useful property of the vertex domination : if @xmath522 , @xmath518 , and @xmath240 dominates @xmath107 in @xmath1 , then @xmath240 dominates @xmath107 in @xmath94 $ ] .
if @xmath240 strongly dominates @xmath107 in @xmath1 , it is not true in general that @xmath240 strongly dominates @xmath107 in @xmath94 $ ] . as already the example of the tree shows , the minimal subgraph of @xmath1 to which it reduces is not unique .
however the following weaker version of uniqueness is true .
let @xmath1 be a graph and @xmath523 , such that @xmath1 reduces both to @xmath94 $ ] and to @xmath524 $ ] , and both @xmath94 $ ] and @xmath524 $ ] are irreducible , then @xmath94 $ ] is isomorphic to @xmath524 $ ] .
* we prove the statement by induction on the number of vertices in @xmath1 .
if @xmath525 , then @xmath526 , so the result is trivially true .
assume now that @xmath527 .
choose @xmath528 containing exactly one vertex from each maximal equivalence class in @xmath521 , and no other vertices .
if @xmath529 , then @xmath1 is irreducible , so we can assume that @xmath530 .
let us show that there exists @xmath531 , such that @xmath1 reduces to @xmath532 $ ] , and @xmath94 $ ] is isomorphic to @xmath532 $ ] .
assume that no such @xmath533 exists .
consider all the reduction sequences @xmath534 leading from @xmath1 to a graph isomorphic to @xmath94 $ ] .
set @xmath535 , and choose the sequence which minimizes @xmath536 .
denote this sequence by @xmath537 .
set @xmath538 , and @xmath539 .
if each vertex of @xmath532 $ ] is either in @xmath540 or is dominated in @xmath532 $ ] by some vertex in @xmath540 , then , since @xmath532 $ ] is irreducible , we conclude that @xmath531 , yielding a contradiction
. thus we may pick the _
@xmath434 , such that there exists @xmath541 , which is not dominated by any vertex of @xmath542 in @xmath543 . by the choice of @xmath544 , and
what is said above , we have @xmath545 $ ] .
clearly , since @xmath107 was dominated by some vertex of @xmath542 in @xmath546 , we have that @xmath547 , and @xmath548 is the only vertex of @xmath542 which dominates @xmath107 in @xmath549 . in particular , @xmath548 itself is not dominated by any other vertex of @xmath542 in @xmath549 . by the choice of @xmath434 , every vertex in @xmath549 , which is not in @xmath542 ,
is dominated by some vertex in @xmath542 , hence @xmath548 is not strongly dominated by any other vertex . since @xmath550 is a legal reduction , there must exist a vertex @xmath248 equivalent to @xmath548 in @xmath549 .
we have @xmath551 , since either @xmath552 , or @xmath248 dominates @xmath107 .
consider a graph isomorphism @xmath553 , which swaps the vertices @xmath548 and @xmath248 , and fixes every other vertex .
it is easy to see that @xmath554 is a legal reduction sequence leading from @xmath1 to @xmath555 $ ] , such that @xmath555 $ ] is isomorphic to @xmath94 $ ] .
furthermore , since removal of @xmath547 was either replaced by or swapped with the removal of @xmath551 , the invariant , which we minimized over the sequences , is actually smaller for this sequence than for @xmath537 .
this is again a contradiction . finally , consider the case @xmath556 .
since @xmath557 , we can use the induction assumption to prove the theorem , as long as we can show that @xmath558 $ ] reduces to @xmath94 $ ] and to @xmath524 $ ] . by the argument above , we can choose @xmath95 so that , if @xmath537 is the reduction sequence leading to @xmath94 $ ] , and @xmath559 , then , for any @xmath560 , every vertex in @xmath561 is dominated by some vertex from @xmath542 in @xmath562 .
it is then immediate that @xmath563 is the reduction sequence from @xmath558 $ ] to @xmath94 $ ] , where @xmath564 , @xmath565 .
indeed , for any @xmath566 , @xmath548 is dominated by some vertex in @xmath567 , hence it is dominated by some vertex from @xmath568 in @xmath567 .
it follows that @xmath548 is dominated by some vertex in @xmath569 $ ] , allowing to reduce the latter graph to @xmath570 $ ] . for future reference we explicitly state the following consequence of the proposition [ prop5.1 ] .
[ equicrl ] let @xmath1 be a graph , and @xmath93 , such that @xmath1 reduces to @xmath94 $ ] .
assume @xmath95 is @xmath571-invariant for some @xmath572 .
then the inclusion @xmath573{\hookrightarrow}g$ ] induces a @xmath571-invariant homotopy equivalence @xmath574,h)$ ] for an arbitrary graph @xmath2 .
note also , that the proposition [ prop5.1 ] can not be generalized to encompass arbitrary graph homomorphisms @xmath238 of @xmath1 onto @xmath2 , where @xmath2 is a subgraph of @xmath1 , and @xmath238 is identity on @xmath2 . as an example in the subsection [ ss2.3 ] showed @xmath575 despite of the existence of the folding map of @xmath576 onto @xmath337 .
@xmath17 next , we use the proposition [ prop5.1 ] to compute homotopy types of the complexes of maps from finite forests to complete graphs .
[ homtree ] if @xmath67 is a tree with at least one edge , then the map @xmath577 induced by any inclusion @xmath578 is a homotopy equivalence , in particular @xmath579 . as a consequence ,
if @xmath15 is a forest , and @xmath580 are all its connected components consisting of at least 2 vertices , then @xmath581 .
* let @xmath67 be a tree with @xmath49 vertices , @xmath582 .
note the general fact , that if @xmath107 is a leaf of a tree , @xmath240 is the vertex adjacent to @xmath107 , and @xmath583 is a vertex adjacent to @xmath240 , then @xmath584 , hence @xmath67 reduces to @xmath585 . let us now number the vertices @xmath586 so that for any @xmath587 $ ] , @xmath588 is a leaf in @xmath589 . by the previous observation @xmath590=k_2\ ] ] is a valid reduction sequence .
thus the first part of the statement follows by the proposition [ prop5.1 ] .
that @xmath579 follows from the proposition [ polyprop ] . finally , the formula for the homotopy type of @xmath12 follows from ( 3 ) in the subsection [ ss2.4 ] .
let @xmath591 denote the @xmath40-dimensional sphere equipped with the antipodal action of @xmath9 ; in the same way @xmath592 denotes the @xmath40-dimensional sphere equipped with the trivial action of @xmath9
. given two spaces @xmath70 and @xmath593 with @xmath9-action , we let @xmath594 denote the @xmath9-equivariant homotopy equivalence . [ homtree2 ] let @xmath67 be a tree with at least one edge and a @xmath9-action determined by an invertible graph homomorphism @xmath595 . if @xmath596 flips an edge in @xmath67 , then @xmath597 , otherwise @xmath598 .
* proof . *
assume @xmath596 flips an edge , that is there exist @xmath599 , such that @xmath600 , @xmath601 , and @xmath602 . by the corollary [ equicrl ] the inclusion map @xmath603{\hookrightarrow}t$ ]
induces a @xmath9-equivariant homotopy equivalence @xmath604 , where the last space has the natural @xmath9-action induced by the flipping @xmath9-action on @xmath337 . by the proposition [ polyprop ] we get @xmath597 .
assume now , there is no edge flipped by @xmath596 .
since @xmath67 is a contractible finite cw complex ( the topology is generated by fixing homeomorphisms between edges of @xmath67 and a standard unit interval ) it follows from @xcite that @xmath596 must have a fixed point .
denote this point by @xmath381 . clearly , either @xmath381 is a vertex of @xmath67 or @xmath381 is the middle - point of some edge @xmath605 $ ] . in the latter case ,
if the edge is not fixed pointwise , then it is flipped , which contradicts our assumptions on @xmath596 .
thus we found a vertex @xmath33 fixed by @xmath596 .
if there exists @xmath606 , such that @xmath607 , @xmath608 , then @xmath603{\hookrightarrow}t$ ] induces a @xmath9-equivariant homotopy equivalence @xmath604 , where the @xmath9-action on the last space is the trivial one .
it follows that @xmath598 .
finally consider the case when there is no edge in @xmath67 which is fixed by @xmath596 pointwise .
let @xmath240 be any vertex of @xmath67 adjacent to @xmath107 , and let @xmath609 .
since the set @xmath610 is @xmath596-invariant , we see by the corollary [ equicrl ] that the inclusion map @xmath611{\hookrightarrow}t$ ] induces a @xmath9-equivariant homotopy equivalence @xmath612,k_n)={\text{\tt hom}\,}(l_3,k_n)$ ] , where the @xmath9-action on the last space is induced from the @xmath9-action on @xmath613 which swaps @xmath240 and @xmath248 .
let @xmath614 be any of the two elements of @xmath615 .
clearly , @xmath238 is @xmath9-equivariant with the @xmath9-action on @xmath337 being trivial .
this shows that @xmath616 is a @xmath9-equivariant homotopy equivalence , which finishes the proof .
since taking the unlooped complement reverses neighbor set inclusions , we see that @xmath1 reduces if and only if @xmath91 reduces .
the next proposition describes what happens if @xmath1 is a forest .
[ homcompf ] if @xmath15 is a forest , then @xmath617 , where @xmath60 is the maximal cardinality of an independent set in @xmath15
. * proof .
* we use induction on the number of edges in @xmath15 .
if @xmath618 , then @xmath619 , the maximal cardinality of an independent set in @xmath15 is @xmath620 , and the statement is obvious .
so assume @xmath621 .
let @xmath622 be an arbitrary leaf , and let @xmath623 be the vertex adjacent to @xmath107 .
we have @xmath624 .
hence @xmath16 reduces to @xmath625 .
clearly , @xmath626 , so by combining the induction assumption with the proposition [ prop5.1 ] we get @xmath627 , where @xmath628 is the maximal cardinality of an independent set in @xmath629 .
let @xmath630 be an independent set in @xmath15 of maximal cardinality .
either @xmath240 or @xmath107 must be in @xmath630 , since otherwise @xmath631 is independent , and larger than @xmath630 .
if @xmath632 , then @xmath633 is also an independent set in @xmath15 of maximal cardinality .
either way , we have an independent set @xmath634 in @xmath15 of maximal cardinality containing @xmath107 . since any independent set in @xmath629 is also independent in @xmath15 , and @xmath634 is independent in @xmath629 , we can conclude that @xmath635 , hence the result .
it follows from the proposition [ pr_chom ] that @xmath14 is homotopy equivalent to a wedge of @xmath4-dimensional spheres .
after the appearance of the initial series of papers , @xcite , and this paper , there has been a lot of further research activities pertaining to @xmath58-complexes .
we refer the reader to the recent survey @xcite . *
the conjecture [ conj1 ] has been proved , see @xcite . *
the proposition [ propn ] has been strengthened .
it has been proved that @xmath57 and @xmath54 have the same simple homotopy type .
see @xcite for the proof , and a description of an explicit formal deformation .
* the proposition [ prop5.1 ] has been strengthened .
it has been proved that the folds are admissible on both sides of the @xmath636 , and that one can replace the homotopy equivalences by formal deformations , see @xcite for an explicit construction .
bridson , a. haefliger , _ metric spaces of non - positive curvature _ , grundlehren der mathematischen wissenschaften [ fundamental principles of mathematical sciences ] , * 319 * , springer - verlag , berlin , 1999 . d.n .
kozlov , _ chromatic numbers , morphism complexes , and stiefel - whitney characteristic classes _ , invited contribution to _ geometric combinatorics _ , ias / park city mathematics series * 14 * , american mathematical society , providence , ri ; institute for advanced study ( ias ) , princeton , nj . | @xmath0 is a polyhedral complex defined for any two undirected graphs @xmath1 and @xmath2 .
this construction was introduced by lovsz to give lower bounds for chromatic numbers of graphs . in this paper
we initiate the study of the topological properties of this class of complexes .
we prove that @xmath3 is homotopy equivalent to a wedge of @xmath4-dimensional spheres , and provide an enumeration formula for the number of the spheres . as a corollary we prove that if for some graph @xmath1 , and integers @xmath5 and @xmath6 , we have @xmath7 , then @xmath8 ; here @xmath9-action is induced by the swapping of two vertices in @xmath10 , and @xmath11 is the first stiefel - whitney class corresponding to this action .
furthermore , we prove that a fold in the first argument of @xmath0 induces a homotopy equivalence .
it then follows that @xmath12 is homotopy equivalent to a direct product of @xmath13-dimensional spheres , while @xmath14 is homotopy equivalent to a wedge of spheres , where @xmath15 is an arbitrary forest and @xmath16 is its complement . |
this work was supported by the department of energy under grant doe / er/40561 .
n. bohr and b. r. mottelson , _ nuclear structure _ vol .
i , benjamin ( new york , 1969 ) p. ring and p. schuck , _ the nuclear many - body problem _ springer ( new york , 1980 ) g. f. bertsch and r. a. broglia , _ oscillations in finite quantum systems _ cambridge university press , ( new york , 1994 ) t. papenbrock , _ collective and chaotic motion in self - bound many - body systems _ , e - print chao - dyn/9905007 t. papenbrock and t. prosen , _ quantization of a billiard model for interacting particles _ , e - print chao - dyn/9905008 r. l. liboff , _ kinetic theory _ , wiley & and sons ( new york , 1988 ) | we extend a recent billiard model of the nuclear @xmath0-body hamiltonian to consider a finite two - body interaction .
this permits a treatment of the hamiltonian by a mean field theory , and also allows the possibility to model reactions between nuclei .
the density and the mean field potential can be accurately described by a scaling function which shows the qualitative features of the liquid drop picture of the nucleus . the nuclear many - body problem is often approximated by phenomenological mean field or liquid drop models which serve as good starting points to study the shape and the spectrum of low lying single particle and collective excitations @xcite .
this well - known approach is a consequence of the inadequate knowledge of the microscopic nuclear hamiltonian , and the difficulty in dealing with the many - body problem .
the mean field approximation is commonly employed to make the many - body problem tractable .
therefore it is of interest to question of how well mean - field theory works for a given microscopic hamiltonian that can be solved exactly . a promising model for such a program
may readily be obtained from a recent billiard model of the nuclear @xmath0-body hamiltonian .
this is an interacting @xmath0-body system with two - body interactions that is rich enough to show various features of self - bound many - body systems yet simply enough to allow for practical calculations and an understanding @xcite . generalizing the model of ref .
@xcite , we consider the hamiltonian [ ham ] h=_i=1^np_i^22 m + _ i < jv(|_i-_j| ) , where @xmath1 is a three - dimensional position vector of the @xmath2-th nucleon and @xmath3 is its conjugate momentum and the interaction given by [ int ] v(r)=\ { ll -v_0 & , + 0 & , .
where @xmath4 is a positive constant .
total energy @xmath5 , momentum @xmath6 and angular momentum @xmath7 are conserved quantities .
interactions alter the momenta of particles whenever they are distance @xmath8 apart , and otherwise they move freely . for @xmath9
one obtains a billiard model , which was recently studied in ref .
@xcite . due to
the simple form of the two - body interaction , the numerical integration of the equations of motion is quite easy , scaling with particle number as @xmath10 rather than the usual @xmath11 typical of the classical @xmath0-body problem .
it was found that the dynamics is dominantly chaotic and ergodic . in the center of mass system one
finds a constant single - particle density inside a circle of diameter @xmath8 that quickly drops to zero within a thin surface region .
the phase space structure does not depend on energy in such billiards .
finite values of @xmath4 introduce an energy scale and a finite binding energy of the system .
the hamiltonian then permits reactions like particle capture or emission , giving it important features for a model of the nucleus and its dynamics . in this note
we will examine the mean field theory of of the @xmath0-body hamiltonian ( [ ham],[int ] ) .
to this purpose we introduce the single - particle phase space density @xmath12 .
let @xmath13 be the mean field hamiltonian .
as is well known , any density which is a functional of the mean field hamiltonian solves the corresponding vlasov equation @xcite .
however , we are interested in the situation corresponding to thermal equilibrium and choose the canonical distribution function f ( , ) , where @xmath14 denotes the temperature . obviously , the phase space distribution function is a product of the normalized momentum space density @xmath15^{-1 } \exp{\left(-p^2/2mkt\right)}$ ] and the number density [ ndens ] n()=c .
the normalization @xmath16 is given by @xmath17^{-1}\ ] ] the mean field potential @xmath18 depends on the number density and the two - body interaction [ vmean ] w()=d^3xv(|-|)n ( ) . eq.([ndens ] ) together with eq.([vmean ] ) are the mean field equations to be solved self - consistently .
we next observe that the solution of the mean field equations ( [ ndens]-[vmean ] ) will have an interesting scaling property .
let @xmath18 and @xmath19 denote the mean field potential and density for parameters @xmath20 and @xmath21 .
then @xmath22 is the mean field potential for parameters @xmath23 and @xmath24 , while @xmath25 is the density for parameters @xmath26 and @xmath27 .
thus , the normalized quantities @xmath28 and @xmath29 depend only on the ratio nv_0/kt of parameters . this scaling behavior comes directly from eqs .
( [ ndens],[vmean ] ) and does not depend on the specific form of the two - body interaction ( [ int ] ) .
it is well known that the density of an open system can be represented as a thermal distribution function only for @xmath30 . in this case
the density is exponentially small outside the nucleus and can be omitted . in what follows
we restrict ourselves to this low temperature regime .
we are interested in spherical symmetric solutions of the mean field equations . in this case
eq.([vmean ] ) may be written as w(r)=_0^dx n(x ) v(r , x ) , where v(r , x)=-2v_0\ { ll 0 & , + 2x^2 & , + x2r & . .
for computational purposes it is useful to rewrite the mean field potential as [ compot ] w(r)=-2v_0where @xmath31 is the unit step function .
we solve the mean field equations numerically by iteration .
we take an initial density @xmath32 that is constant inside a sphere of diameter @xmath8 and zero outside . upon several iterations
one obtains a converged solution .
fig .
[ fig1 ] shows the resulting number density @xmath33 and mean field potential @xmath34 for different values of the parameter @xmath35 . for large @xmath35 the mean
field potential is flat for @xmath36 and increases until @xmath37 while the density is constant for @xmath38 and quickly drops to zero around @xmath39 . decreasing values of @xmath35 ( i.e. decreasing values of @xmath4 and/or increasing temperatures @xmath14 while the number of particles @xmath0 is kept fixed ) lead to a thicker surface region and to a more extended mean field potential .
however , for @xmath40 the potential is still a flat bottom potential and the density is roughly constant .
thus , the system exhibits qualitative features of the liquid drop model .
however , our simple model does not yield a saturation of the density or the binding energy with increasing @xmath0 .
this is due to the absence of a repulsive potential core . in the limit
@xmath41 the density approaches that of a sharp - edged liquid drop , constant inside a sphere of radius @xmath42 and zero outside . for finite values of @xmath35 ,
the surface thickness can be estimated assuming a trapezoidal density distribution .
taking the width parameter as @xmath43 and demanding self - consistency in the neighborhood of @xmath44 , one finds the relation /a=(3)^-1/2 .
[ eps / a ] it is also interesting to compare the mean field results with the interacting system ( [ ham],[int ] ) .
agreement may only be expected for sufficiently large numbers of particles @xmath45 . in what follows we set @xmath46 .
the initial conditions of the @xmath45-body system are chosen such that the positions are drawn at random inside a sphere of diameter @xmath8 while the momenta are drawn from a maxwell - boltzmann distribution with temperature @xmath14 .
we follow the time evolution of the many - body system for roughly @xmath47 interactions to allow for equilibration . at the end of the evolution
we take the positions in the center of mass system and compute the integrated density n(r)=4_0^r dxx^2 n(x ) by counting the number of particles inside a sphere of diameter @xmath48 .
fig .
[ fig2 ] compares the average of ten runs with the mean field result for various values of the temperature .
the agreement is rather good confirming the validity of the mean field approximation .
finally , we ask what values the parameters should have to correspond to physical properties of nuclei .
let us consider a nucleus of mass @xmath49 .
then the radius is given by @xmath50 fm .
the liquid - drop surface thickness is roughly @xmath51 fm , yielding from eq .
( [ eps / a ] ) @xmath52 .
we would like to set the energy scale by the depth of the mean field potential , which should be @xmath53 mev to correspond to a typical woods - saxon potential of nuclear physics . turning to eq .
( 7 ) , we see that the two parameter values can be obtained taking the temperature as @xmath54 mev .
equilibrated nuclei at such a temperature can be easily produced in heavy ion reactions , so the model might have some applicability to nuclear reactions .
however , in detail the fermionic nature of the nuclear many - body problem will make considerable differences from the present classical model .
in particular , the chemical potentials at a given temperature are very different , which would be important for the nucleon evaporation rates . |
since the configuration space of the robotic arm @xmath1 is cat(0 ) , we are able to navigate it thanks to the following theorem . as shown in @xcite
, there is also a ( more complicated ) algorithm to move the robot optimally in terms of the euclidean metric in the configuration space .
however , this metric is less relevant , as it does not seem to have a natural interpretation in terms of the robot .
the algorithms of theorem [ th : alg ] are described in detail in @xcite ; they are based on the _ normal cube paths _ of niblo and reeves @xcite .
we have implemented these algorithms in python for the robotic arms discussed in this paper . to use
the program , the user inputs the width of the tunnel and two positions of the robotic arm of the same length , expressed as sequences of steps of the form : r ( right ) , u ( up ) , d ( down ) steps .
the program outputs the distance between the two states in the edge metric and in the cube metric , and an animation taking the robot between the two states .
the downloadable code , instructions , and a sample animation are at http://math.sfsu.edu/federico/articles/movingrobots.html[`http://math.sfsu.edu/federico/articles/movingrobots.html ` ] .
* our robotic arms @xmath1 can never have links facing left .
now consider the robotic arm of length @xmath2 in a tunnel of width @xmath3 whose links may face up , down , right , * or left * , starting in the fully horizontal position , and subject to the same kinds of moves : switching corners and flipping the end . is the configuration space of this robot also cat(0 ) ?
in @xcite , we show the answer is affirmative for width @xmath10 .
we believe this is also true for larger @xmath3 , but proving it in general seems to require new ideas .
* theorem [ thm : diameter ] gives the diameter of the configuration space @xmath7 in the edge metric , that is , the largest number of moves separating two positions of the robot @xmath1 .
what is the diameter of @xmath7 in the cube metric , that is , the largest number of steps separating two positions of the robot @xmath1 , where a step may consist of several physically independent moves ?
this paper includes results from hb s undergraduate thesis at universidad del valle in cali , colombia ( advised by fa and cc ) and jg s undergraduate project at san francisco state university in san francisco , california ( advised by fa ) .
we thank the sfsu - colombia combinatorics initiative , which made this collaboration possible .
we would also like to thank the pacific ocean for bringing us the beautiful coral that inspired definition [ def : snakepip ] and the proof of our main theorem [ conj : pip ] .
f. ardila , h. bastidas , c. ceballos , and j. guo . the configuration space of a robotic arm in a tunnel of width 2 . to appear in the proceedings of formal power series and algebraic combinatorics ( fpsac 2016 ) , _ discrete mathematics and theoretical computer science_. | we study the motion of a robotic arm inside a rectangular tunnel .
we prove that the configuration space of all possible positions of the robot is a @xmath0 cubical complex .
this allows us to use techniques from geometric group theory to find the optimal way of moving the arm from one position to another .
we also compute the diameter of the configuration space , that is , the longest distance between two positions of the robot .
[ [ section ] ] we consider a robotic arm @xmath1 of length @xmath2 moving in a rectangular tunnel of width @xmath3 without self - intersecting .
the robot consists of @xmath2 links of unit length , attached sequentially , and facing up , down , or right .
the base is affixed to the lower left corner .
figure [ fig : armingrid ] illustrates a possible position of the arm of length 8 in a tunnel of width @xmath4 .
the robotic arm may move freely using two kinds of local moves : @xmath5 _ switch corners : _ two consecutive links facing different directions interchange directions .
@xmath5 _ flip the end : _ the last link of the robot rotates @xmath6 without intersecting itself .
we study the following fundamental problem .
[ prob : moverobot ] move the robotic arm @xmath1 from one given position to another optimally .
when we are in a city we do not know well and we are trying to get from one location to another , we will usually consult a map of the city to plan our route .
this is a simple but powerful idea .
our strategy to approach problem [ prob : moverobot ] is pervasive in many branches of mathematics : we will be to build and understand the map " of all possible positions of the robot ; we call it the _ configuration space _ @xmath7 .
such spaces are also called _ state complexes _ , _ moduli spaces _ , or _ parameter spaces _ in other fields .
following work of abrams
ghrist @xcite in applied topology and reeves @xcite in geometric group theory , ardila , owen , and sullivant @xcite and ardila , baker , and yatchak @xcite showed how to solve problem [ prob : moverobot ] for robots whose configuration space is _
cat(0 ) _ ; this is a notion of non - positive curvature defined in section [ sec : cat(0 ) ] . this is the motivation for our main result .
[ th : cat(0 ) ] the configuration space @xmath7 of the robotic arm @xmath1 of length @xmath2 in a tunnel of width @xmath8 is a @xmath0 cubical complex . of the robotic arm @xmath9.,height=151 ] in light of @xcite , theorem [ th : cat(0 ) ] provides a solution to problem [ prob : moverobot ] .
there is an explicit algorithm , implemented in python , to move the robotic arm @xmath1 optimally from one given position to another
. the reader may visit http://math.sfsu.edu/federico/articles/movingrobots.html[`http://math.sfsu.edu/federico/articles/movingrobots.html ` ] to watch a video preview of this program or to download the python code .
abrams and ghrist showed that cubical complexes appear in a wide variety of contexts where a discrete system moves according to local rules such as domino tilings , reduced words in a coxeter group , and non - intersecting particles moving in a graph and that these cubical complexes are sometimes cat(0 ) @xcite .
the goal of this paper is to illustrate how the techniques of @xcite may be used to prove that configuration spaces are cat(0 ) , uncover their underlying combinatorial structure , and navigate them optimally . to our knowledge ,
theorem [ th : cat(0 ) ] provides one of the first infinite families of configuration spaces which are cat(0 ) cubical complexes , and the first one to invoke the full machinery of _ posets with inconsistent pairs _ of @xcite .
it is our hope that the techniques employed here will provide a blueprint for the study of cat(0 ) cubical complexes in many other contexts .
we now describe the structure of the paper . in section [ sec : prelim ] we define more precisely the configuration space @xmath7 of the robotic arm @xmath1 .
section [ sec : euler ] is devoted to collecting some preliminary enumerative evidence for our main result , theorem [ th : cat(0 ) ] .
it follows from very general results of abrams and ghrist @xcite that the configuration space @xmath7 is a cubical complex . also , we know from work of gromov @xcite that @xmath7 will be @xmath0 if and only if it is contractible . before launching into the proof that @xmath7 is cat(0 ) , we first verify in the special case @xmath10 that this space has the correct euler characteristic , equal to @xmath11 .
we do it as follows .
[ th : faceenum ] if @xmath12 denotes the number of @xmath13-dimensional cubes in the configuration space @xmath14 for a robot of width @xmath4 , then @xmath15 the euler characteristic of @xmath14 is @xmath16 .
substituting @xmath17 above we find , in an expected but satisfying miracle of cancellation , that the generating function for @xmath18 is @xmath19.we obtain : [ th : euler ] the euler characteristic of the configuration space @xmath14 equals @xmath11 . in section [ sec :
cat(0 ) ] we collect the tools we use to prove theorem [ th : cat(0 ) ] .
ardila , owen , and sullivant @xcite gave a bijection between @xmath0 cubical complexes @xmath20 and combinatorial objects @xmath21 called _ posets with inconsistent pairs _ or pips .
the pip @xmath21 is usually much simpler , and serves as a
remote control " to navigate the space @xmath20 .
this bijection allows us to prove that a ( rooted ) cubical complex is @xmath0 by identifying its corresponding pip .
we use this technique to prove theorem [ th : cat(0 ) ] in section [ sec : ourpip ] . to do this ,
we introduce the _ coral pip _
@xmath22 of _ coral tableaux _
, illustrated in figure [ fig : coralpip ] . by studying the combinatorics of coral tableaux , we show the following : [ conj : pip ] the _ coral pip _ @xmath22 is the pip that certifies that the configuration space @xmath7 is a cat(0 ) cubical complex . in section [ sec
: diameter ] we use the combinatorics of coral tableaux to find a combinatorial formula for the distance between any two positions of the robotic arm .
this allows us to find the longest of such distances that is , the diameter of the configuration space @xmath7 in the @xmath23-metric answering a question of the first author . @xcite the diameter of the transition graph of the robot @xmath1 of length @xmath2 in a tunnel of width @xmath24 is @xmath25 where @xmath26 is the _ left justified _ position , @xmath27 and @xmath28 is the _ fully horizontal _ position ( see figure [ fig : robot_left ] ) .
a precise formula is stated in theorem [ thm : diameter ] .
finally , in section [ sec : algorithm ] we discuss our solution to problem [ prob : moverobot ] .
as explained in @xcite , we use the coral pip as a remote control for our robotic arm .
this allows us to implement an algorithm to move the robotic arm @xmath1 optimally from one position to another .
[ [ sec : prelim ] ] recall that @xmath1 is a robotic arm consisting of @xmath2 links of unit length attached sequentally , and facing up , down , or right .
the robot is inside a rectangular tunnel of width @xmath3 and pinned down at the bottom left corner of the tunnel .
it moves inside the tunnel without self - intersecting , by switching corners and flipping the end as illustrated in figure [ fig : mov ] . in this section
we describe the configuration space @xmath29 of all possible positions of the robot @xmath1 .
we begin by considering the _ transition graph _
@xmath30 whose vertices are the possible states of the robot , and whose edges correspond to the allowable moves between them .
figure [ fig : grafodetrancisionn=4 ] illustrates the transition graph @xmath31 of a robotic arm of length @xmath32 .
as these examples illustrate , each one of these graphs is the @xmath11-skeleton of a cubical complex .
for example , consider a position @xmath33 which has two legal moves @xmath34 and @xmath35 occuring in disjoint parts of the arm .
we call @xmath34 and @xmath35 _ physically independent _ or _ commutative _ because @xmath36 . in this case
, there is a square connecting @xmath37 and @xmath38 in @xmath30 .
more generally , we obtain the configuration space by filling all the cubes of various dimensions that we see in the transition graph .
let us make this precise .
[ def : configuration space]the _ configuration space _ or _ state complex _
@xmath29 of the robot @xmath1 is the following cubical complex .
the vertices correspond to the states of @xmath1 .
an edge between vertices @xmath33 and @xmath39 corresponds to a legal move which takes the robot between positions @xmath33 and @xmath39 .
the @xmath40-cubes correspond to @xmath40-tuples of commutative moves : given @xmath40 such moves which are applicable at a state @xmath33 , we can obtain @xmath41 different states from @xmath33 by performing a subset of these @xmath40 moves ; these are the vertices of a @xmath40-cube in @xmath29 .
we endow @xmath29 with a euclidean metric by letting each @xmath40-cube be a unit cube .
figure [ fig : gft6 ] shows the configuration space @xmath42 ; every square and cube in the diagram is filled in . |
recently , much attention has been focused on uncertainties associated with the parton distributions that are determined in the next - to - leading order ( nlo ) global analyses of a wide range of deep inelastic and related scattering data .
there are many sources of uncertainty , but they can be divided into two classes : those which are associated with the _ experimental _ errors on the data that are fitted in the global analysis and those which are due to what can loosely be called _ theory _ errors . in this latter category
we have uncertainties due to ( i ) nnlo and higher - order dglap contributions , ( ii ) effects beyond the standard dglap expansion , such as extra @xmath11 and @xmath12 terms , higher twist and saturation contributions , ( iii ) the particular choice of the parametric form of the starting distributions , ( iv ) heavy target corrections , ( v ) model assumptions , such as @xmath13 . in order to estimate some of these ` theory ' errors , we can also look at the uncertainties arising from different choices of the data cuts ( @xmath14 , @xmath15 , @xmath16 ) , defined such that data with values of @xmath0 , @xmath17 or @xmath18 below the cut are excluded from the global fit . this approach indicates where the current theory is struggling to fit the data compared to other regions . here
we study the uncertainties due to the errors on the data , and leave the discussion of the ` theory ' uncertainties to a second paper .
other groups @xcite@xcite have also concentrated on the experimental errors and have obtained estimates of the uncertainties on parton distributions within a nlo qcd framework , using a variety of competing procedures . of course , the parton distributions are not , themselves , physical quantities .
however , using the standard approach of the linear propagation of errors , these uncertainties of the parton distributions can be translated into uncertainties on observables .
therefore , we first follow the general approach in @xcite and @xcite , the hessian method , and diagonalize the error matrix , parameterizing an increase in @xmath19 of the fit in terms of a quadratic function of the variation of the parameters away from their best fit values .
this gives us a number of sets of partons with variations from the minimum in orthogonal directions which can be used in a simple manner to calculate the uncertainty on any physical quantity .
however , this approach depends for its reliability on the assumption that the quadratic dependence on the variation of the parton parameters is very good .
we find that this approximation , with some modifications of the precise framework , i.e. the elimination of some parameters and rescaling of others , can be made to work well .
we make available 30 sets of partons 2 for each of the 15 eigenvector directions in parton parameter space which can be used to calculate the uncertainties on any physical quantity . despite its convenience
, the hessian approach does suffer from some problems if one looks at it in detail , and if one tries to extrapolate results , in particular if we consider large increases in @xmath19 .
it is also not , in principle , the most suitable method when allowing @xmath20 to vary as one of the free parameters in the fit .
hence , in this paper we also investigate the uncertainties on observables directly . in order to do this
we apply the lagrange multiplier method @xcite to the observables themselves , therefore avoiding some of the approximations involved in the linear propagation of errors from partons to the observables , and confirming that these approximations do not usually cause serious problems . when using this lagrange multiplier approach , the resulting sets of parton distributions , which correspond to the extreme values of each observable , can to a certain extent be thought of as the maximum allowed variation of the dominant contributing partons in the relevant kinematic ( @xmath3 ) domain .
we select observables which are particularly relevant for experiments at present and future colliders , and which illustrate the uncertainties on specific partons in a variety of kinematic ( @xmath3 ) domains . in order to determine the true uncertainty on quantities we also let @xmath6 vary along with the parameters describing the parton distributions directly , which is easy to implement in this approach .
some quantities are then far more sensitive to @xmath6 than others .
fortunately our global fit @xcite produces a value of @xmath6 which is consistent with the world average , with the same type of error , i.e. , @xmath21 .
hence , it is completely natural to simply let @xmath6 vary as a free parameter in the fit in the same way as all the other parameters when determining uncertainties .
however , we also perform an investigation of the uncertainties with @xmath6 fixed at 0.119 in order to study more directly which variations in the parton distributions are responsible for extreme variations in given physical quantities , and to compare with the results of the hessian approach . the physical observables that we select as examples in this introductory study are , first , the charged - current structure functions @xmath22 for deep inelastic scattering at high @xmath17 and @xmath18 at hera . these observables almost directly represent the @xmath23 , @xmath24 valence quarks at high @xmath17 and @xmath18 , where deep inelastic data do exist @xcite@xcite , but have errors of @xmath25 or more at present .
the precision on these data is expected to increase dramatically in the near future .
second , we determine the uncertainties on the cross - sections @xmath4 and @xmath5 , for @xmath0 boson production and for the production of a higgs boson of mass domains . ]
@xmath26 gev by gluon fusion respectively , at tevatron and lhc energies .
the cross section @xmath4 is sensitive to the sea quarks ( and also , at the tevatron energy , weakly sensitive to the valence quarks ) in a range of rapidity centered about @xmath27 , and for @xmath28 .
similarly , @xmath5 is sensitive to the gluon distribution in the domain @xmath29 and @xmath30 . as a third example
we determine the uncertainty on the ratio of @xmath1 to @xmath2 production at the lhc energy .
this ratio is expected to be extremely accurately measured in the lhc experiments .
other relevant examples , which we study , are the uncertainties of the moments of the non - singlet ( @xmath24@xmath23 ) quark distributions .
these are quantities for which lattice qcd predictions are becoming available , see , for example , refs .
@xcite .
the same techniques can be easily and quickly applied to a wide variety of other physical processes sensitive to different partons and different domains . besides giving a direct evaluation of the uncertainties on the observables , we can , in principle , unfold this information to map out the uncertainties on nlo partons over the whole kinematic domain where perturbative qcd is applicable .
the plan of the paper is as follows . in section 2
we discuss the hessian method , and outline our extraction of different parton distribution sets using this approach .
in particular we highlight the problems encountered , and how they are dealt with in order to obtain reliable results .
we make the sets of partons obtained publicly available . in section 3
we briefly recall the elements of the lagrange multiplier method . in the following four sections we determine the uncertainties of the observables that we have mentioned above .
this will involve a series of global fits in which the observables are constrained at different values in the neighbourhood of their values obtained in the optimum global fit . in each case
we explore , and discuss , the allowed variation of the dominantly contributing partons . using this more rigorous method
we also confirm the general appropriateness of the hessian approach , but discuss where it can start to break down .
finally , in section 8 , we summarize and briefly investigate the uncertainties associated with the initial assumptions made in performing the global fit . in order to do this
we compare the @xmath0 and @xmath31 boson predictions with those obtained using both a slightly updated set of our own partons , mrst2002 , and using the cteq6 partons @xcite .
( all the results in sections 27 are based on mrst2001 partons @xcite . )
we find that for the comparison with cteq some of the variations in predictions are surprisingly large .
we also illustrate the same result for the extractions of @xmath6 by various different groups .
this implies that uncertainties involved with initial assumptions and also with theoretical corrections can be more important than those due to errors on the data .
the basic procedure involved in this approach is discussed in detail in @xcite , but we briefly introduce the important points here . in this method
one assumes that the deviation in @xmath19 for the global fit from the minimum value @xmath32 is quadratic in the deviation of the parameters specifying the input parton distributions , @xmath33 , from their values at the minimum , @xmath34 . in this case
we can write @xmath35 where @xmath36 is an element of the hessian matrix , and @xmath37 is the number of free input parameters . in this case
the standard linear propagation of errors allows one to calculate the error on any quantity @xmath38 using the formula @xmath39 where @xmath40 is the covariance , or error matrix of the parameters , and @xmath41 is the allowed variation in @xmath19 .
hence , in principle , once one has either the hessian or covariance matrix ( and a suitable choice of @xmath42 ) one can calculate the error on any quantity .
however , as demonstrated in @xcite , it more convenient and more numerically stable to diagonalize either the hessian or covariance matrix , and work in terms of the eigenvectors .
since the hessian and covariance matrices are symmetric they have a set of orthogonal eigenvectors defined by @xmath43 moreover , because variations in some directions in parameter space lead to deterioration in the quality of the fit far more quickly than others , the eigenvalues @xmath44 span several orders of magnitude .
hence it is helpful to work in terms of the rescaled eigenvectors @xmath45 .
then the parameter displacements from the minimum may be expressed as @xmath46 or using the orthogonality of the eigenvectors @xmath47 i.e. , the @xmath48 are the appropriately normalized combinations of the @xmath49 which define the orthogonal directions in the space of deviation of parton parameters . in practice a @xmath48
is often dominated by a single @xmath49 .
the error determination becomes much simpler in terms of the @xmath48 .
the increase in @xmath19 is @xmath50 i.e. , the surface of constant @xmath19 is a hyper - sphere of given radius in @xmath51-space .
similarly the error on the quantity @xmath38 is now @xmath52^{1/2}. \label{eq : covariancediag}\ ] ] thus it is convenient to introduce parton sets @xmath53 for each eigenvector direction , i.e. , from eq .
( [ eq : component ] ) we define @xmath54 where the tolerance @xmath55 is defined by @xmath56 and @xmath41 is the allowed deterioration in fit quality for the error determination . then , assuming the quadratic behaviour of @xmath38 about the minimum , ( [ eq : covariancediag ] ) becomes the simple expression @xmath57^{\frac{1}{2}}. \label{eq : covdiag}\ ] ] if everything were ideal this framework would provide us with a simple and efficient method for calculating the uncertainty due to experimental errors on any quantities , where we would use the standard choice of @xmath58 .
however , the real situation is not so simple , and there are two major complications we must overcome in order to obtain reliable results .
although , in principle , the @xmath59 uncertainty in any cross - section should be given by @xmath60 , the complicated nature of the global fitting procedure , where a large number of independent data sets are used , results in this being an unrealistically small uncertainty @xcite .
this is undoubtedly due to some failure of the theoretical approximation to work absolutely properly over the full range of data , which introduces the type of theoretical errors outlined in the introduction , and also due to some sources of experimental error not being precisely quantified .
both problems are in practice extremely difficult to surmount
. we shall implicitly ignore the potential theoretical error in this paper , but account for the lack of ideal behaviour in the framework by determining the uncertainties using a larger @xmath41 .
we estimate @xmath61 to be a conservative uncertainty ( perhaps of the order of a @xmath62 confidence level or a little less than @xmath63 ) due to the observation that an increase of 50 in the global @xmath19 , which has a value @xmath64 for @xmath65 data points , usually signifies that the fit to one or more data sets is becoming unacceptably poor .
we find that an increase @xmath42 of 100 normally means that some data sets are very badly described by the theory .
though this estimation does not rely on any real mathematical foundation we do not think it is any less valid than the approaches used in e.g. @xcite or @xcite , both of which ultimately appeal to some value judgment rather than using all available information in a statistically rigorous manner , and ultimately give similar results .
the approaches @xcite do use @xmath60 but either rely on much smaller and more internally compatible data sets , or in some cases have rather small errors . the second complication is the breakdown of the simple quadratic behaviour in terms of variations of the parameters , i.e. , the fact that eq .
( [ eq : hessian ] ) may receive significant corrections and the simple linear propagation of errors is therefore not accurate .
of course , we expect some deviations from this simple form for very large @xmath41 , but unfortunately very significant deviations can occur for relatively small @xmath66 , as outlined below . due to the very large amount of data in our global fit
, we have a lot of parameters in order to allow sufficient flexibility in the form of the parton distributions .
each of the valence quarks and the total sea quark contribution are parameterized in the form @xmath67 where for the valence quarks the normalization @xmath68 is set by the number of valence quarks of each type .
because we find it necessary to have a negative input gluon at low @xmath17 the gluon parameterization has been extended to @xmath69 where @xmath70 is determined by the momentum sum rule , and @xmath71 can be set to some fixed large value , e.g. 10 or 20 , so that the second term only influences large @xmath17 .
the combination @xmath72 has a slightly different parameterization , i.e. , @xmath73 overall , this gives 24 free parameters . in our standard
fits we allow all these parameters to vary .
however , when investigating in detail the small departures from the global minimum we notice that a certain amount of redundancy in parameters leads to potentially disastrous departures from the behaviour in eq .
( [ eq : hessian ] ) .
for example , in the negative term in the gluon parameterization very small changes in the value of @xmath74 can be compensated almost exactly by a change in @xmath75 and ( to a lesser extent ) in the other gluon parameters over the range of @xmath17 probed , and therefore changes in @xmath74 lead to very small changes in @xmath19 . however , at some point the compensation starts to fail significantly and the @xmath19 increases dramatically .
hence , this certain degree of redundancy between @xmath74 and @xmath75 leads to a severe breaking of the quadratic behaviour in @xmath41 .
essentially the redundancy between the parameters leads to a very flat direction in the eigenvalue space ( a very large / small eigenvalue of the covariance / hessian matrix ) which means that cubic , quartic _ etc_. terms dominate . during the process of diagonalization
this bad behaviour feeds through into the whole set of eigenvectors to a certain extent .
therefore , in order that the hessian method work at all well we have to eliminate the largest eigenvalues of the covariance matrix , i.e. , remove the redundancy from the input parameters . in order to do this
we simply fix some of the parameters at their best fit values so that the hessian matrix only depends on a subset of parameters that are sufficiently independent that the quadratic approximation is reasonable .
in fact we finish up with 15 free parameters in total 3 for each of the 5 different types of input parton . in particular , fixing the other parameters at the best fit values we find that @xmath76 , @xmath77 and @xmath74 are sufficient for the gluon
one for high @xmath17 , one for medium @xmath17 and one for low @xmath17 . however , we emphasize that we can not simply set the other parameters to zero .
for example @xmath75 must be of a size as to allow a sufficiently negative input gluon at low @xmath17 with a sensible value of @xmath74 , but we can not allow it to vary simultaneously with @xmath74 .
we could possibly allow one or two more parameters to be free , but judge that the deterioration in the quality of the quadratic approximation does not outweigh the improvements due to increased flexibility in the parton variations .
we note that this problem seems to be a feature of the full global fits obtained by cteq and mrst , and that the other fitting groups have not yet needed to introduce enough parameters to notice such redundancy .
it has clearly been noticed by cteq though , since in @xcite they only have 16 free parameters out of a possible 22 , and in @xcite , where they use a significantly altered type of parameterization , they have only 20 free parameters out of a possible 26 . hence , we produce 30 sets of parton distributions labeled by @xmath53 to go along with the central best fit ; that is 15 `` + '' sets corresponding to each eigenvector direction , and 15 `` '' sets , mrst2001c ) are available at http://durpdg.dur.ac.uk/hepdata/mrs@xmath78 . ] .
even though we have limited the number of free parameters in the calculation of the hessian matrix , we note that we still have significant departure from the ideal quadratic behaviour . for the 10 or so lowest eigenvalues of the covariance matrix
the quadratic approximation is very good the distance needed to go along one of the @xmath48 to produce @xmath79 being the expected @xmath80 to good accuracy in both `` + '' and `` '' directions .
however , for 4 or 5 of the largest eigenvalues of the covariance matrix , corresponding mainly to the large-@xmath17 @xmath23 quark , large-@xmath17 gluon and @xmath81 distributions , the absolute scaling and symmetry break down somewhat . in the very worst case of the largest eigenvalue ,
the scale factors to produce @xmath82 are 9.5 and 4.5 in the two opposite directions . in order to produce the sets corresponding to @xmath83 we have to multiply the parton deviations required for @xmath84 by these scale factors rather than the expected 7.07 .
( in fact we do this for all 30 sets , but in most cases the scale factor is in the range 6.57.5 . ) hence , as in @xcite , this necessitates the supply of both `` + '' and `` ''
sets , whereas in the quadratic approximation one could easily be obtained from the other . indeed from fig .
9 of @xcite it is clear that cteq encounter a breakdown of the quadratic behaviour of much the same type that we do . using the 30 parton sets @xmath85 corresponding to the 15 eigenvector directions for variations of the partons about the minimum @xmath19
, one can use eq .
( [ eq : covariancediag ] ) to calculate the error for any quantity , assuming an allowed @xmath86 .
in fact it has been proposed @xcite that one may also account for some of the asymmetry due to departures from quadratic behaviour by replacing eq .
( [ eq : covdiag ] ) by the slightly more sophisticated form ( f)_+ & = ^ + ( f)_- & = ^ , [ eq : covdiagasymm ] where @xmath87 represents the best fit set of partons . in @xcite and @xcite
examples are discussed where the use of eq .
( [ eq : covdiagasymm ] ) instead of eq .
( [ eq : covdiag ] ) leads to not only an asymmetric error , but also a larger uncertainty overall .
we find only fairly minor effects , with no real evidence that eq .
( [ eq : covdiagasymm ] ) leads to markedly more reliable results than eq .
( [ eq : covdiag ] ) , so we use the simpler eq .
( [ eq : covdiag ] ) in this paper . as an example of the use of the hessian method we show in figs .
14 the uncertainty on some of the parton distributions at various values of @xmath18 , namely the @xmath88 distribution , the @xmath89 distribution and the gluon distribution respectively .
as one sees , the @xmath88 distribution is very well determined , and the uncertainty shrinks slightly with increasing @xmath18 .
the lowest uncertainty is in the region of @xmath90 where there are very accurate data which mainly constrain the valence quarks . at lower @xmath17
the direct constraint is on the sum of valence and sea quarks .
the @xmath89 distribution is also well determined in general , but is rather more uncertain as we go to the highest @xmath17 values .
the gluon distribution is known less well , but at the highest @xmath18 has an uncertainty of as little as @xmath91 for @xmath92 where it is constrained by both @xmath93 of the hera data and the lowest-@xmath94 tevatron jet data .
it becomes very uncertain for @xmath95 where only the relatively imprecise highest-@xmath94 jet data provide any information .
the fractional uncertainty at very small @xmath17 decreases very rapidly as @xmath18 increases because much of the small-@xmath17 gluon at higher @xmath18 is generated from that at higher @xmath17 via evolution .
we also show the gluon at @xmath96 explicitly in fig . 4 .
at this low scale the central gluon is negative at @xmath97 , but we see that the gluon may be positive within the uncertainty .
this just about persists if we go to as low as @xmath98 at this @xmath18 , but at our input scale @xmath99 the gluon would be negative for @xmath17 less than 0.0005 , outside the level of uncertainty chosen . also shown on the plots are the cteq6 m partons . for the @xmath89 distribution
the agreement is excellent . for the @xmath88 distribution
the agreement at @xmath100 is very good , but there is a discrepancy below this value .
however , in this range , the valence quarks become very small indeed and the data only really constrain the total @xmath24 distribution which is completely dominated by the sea . this apparent discrepancy
is probably due to parameterization effects , and is irrelevant in practice .
however , in fig . 3 we see that the mrst and cteq gluons show a genuine and significant level of incompatibility .
we will comment on this more in section 8 .
one might worry that the fixing of some of the parameters , that determine the input parton distributions , will cast some doubt on the error obtained .
however , we stressed that these are largely redundant parameters , and we have checked that the errors obtained ( when using @xmath83 ) are indeed compatible with the errors obtained using more rigorous means , i.e. , the lagrange multiplier method , in the following sections . ) instead of eq .
( [ eq : covdiag ] ) in these comparisons . in all cases
the former only introduced a relatively small asymmetry in the uncertainty , with the average being very close indeed to the result using the latter . also , the asymmetry was of the same sign as that found using the lagrangian approach only half the time , i.e. the use of eq .
( [ eq : covdiagasymm ] ) did not reliably predict the direction of steeper increase of @xmath41 , even when the asymmetry was quite large .
we find this surprising , and have no good explanation .
however , it illustrates the semi - qualitative nature of the hessian approach compared to the more rigorous lagrange multiplier method . ] nevertheless , it is a sign of the breakdown of the quadratic approximation .
of more practical concern is the fact that this breakdown is also exhibited in a non - trivial manner in some of the eigenvectors used particularly those eigenvectors associated with the least known partons , e.g. the high-@xmath17 down quark and gluon .
the scaling has been designed to give correct results if @xmath83 is used .
however , one can not simply extrapolate to different choices of @xmath41 .
for example if @xmath101 were deemed a more suitable choice , in principle the error would just be that using eq .
( [ eq : covariancediag ] ) divided by @xmath102 , but the breakdown of quadratic behaviour does not guarantee this , especially for some directions in parameter space .
also , if one wished to be very conservative in the estimation of an uncertainty , simple extrapolation can not reveal when @xmath66 might start to increase rapidly .
we will see examples of this later .
also we note that we have performed this analysis for a fixed value of the coupling constant : @xmath103 .
one can in principle include this as another free parameter .
indeed we then find that the behaviour obeys the quadratic approximation quite well and that @xmath83 gives an error of about @xmath104 , corresponding well to our error of @xmath105 obtained in @xcite using @xmath106 .
we will discuss extractions of @xmath6 again in section 8 .
however , for the hessian approach there is a slight difference between varying @xmath6 and varying the parton parameters .
when @xmath6 is fixed the maximum error on any quantity is obtained from some linear combination of our different parton sets , and in principle one could reproduce the particular parton set which corresponds to this linear combination , which would be a perfectly well - defined set itself . however , a linear combination of @xmath107 coming from contributions with different @xmath6 does not actually correspond to one particular choice of @xmath6 ( each contribution has a branch point at a different value of @xmath18 , so a linear combination will have multiple branch points ) , so one can not precisely define a particular set of partons corresponding to a particular @xmath6 for the extreme .
hence , although the 30 parton sets obtained using the hessian approach provide the most convenient framework for calculating the uncertainties on a physical observable , for the reasons described above we would also like to study an alternative approach , partially just to check how well our adapted hessian approach really works . a more robust method , which also allows us to directly investigate the partons , and @xmath20 , corresponding to the extreme variations of a given physical quantity is the lagrange multiplier method .
we study this in detail below .
it is much more rigorous to investigate the allowed variation of a specific observable by using the lagrange multiplier method .
this was also one of the approaches used by the cteq collaboration @xcite . in this , one performs a series of global fits while constraining the values @xmath108 of one , or more , physical quantities in the neighbourhood of their values @xmath109 obtained in the unconstrained global fit . to be precise ,
we minimize the function ( _ i , a ) = _
global^2(a ) + _ i_i_i(a)[eq : psi ] with respect to the usual set @xmath110 of parameters , which specify the parton distributions and the coupling @xmath6 .
this global minimization is repeated for many fixed values of the lagrange multipliers @xmath111 . at the minima , with the lowest @xmath112 ,
the observables have the values @xmath113 and the value of @xmath114 is the minimum for these particular values of @xmath108 .
these optimum parameter sets @xmath115 depend on the fixed values of @xmath111 . clearly , when @xmath116 , we have @xmath117 and @xmath118 . in this way we are able to explore how the global description of the data deteriorates as the @xmath113 move away from the unconstrained best fit values @xmath109 .
thus by spanning a range of @xmath111 we obtain the @xmath119 profile for a range of values of @xmath108 about the best fit values , @xmath109 . in this study
we take the best fit values corresponding to the mrst2001 partons @xcite .
this procedure involves none of the approximations involved in the hessian approach .
we can use the full set of parameters in the fit , obtaining maximum flexibility in the partons without having to worry about the large correlations or anticorrelations between some parameters .
we never make any assumption about quadratic dependence on the parameters , and indeed , by using different values of the lagrange multipliers , we can map out precisely how the quadratic approximation breaks down in the uncertainty for any physical quantity .
also , one produces a particular set of partons with a particular value of @xmath6 at every point in the space of cross - sections for the physical quantities mapped , so the interpretation of the extremes is more obvious and natural . hence , in principle , this is a far superior method of obtaining uncertainties to the hessian approach .
however , it suffers from the large practical disadvantage that a series of global fits must be done every time one considers a new quantity . as examples
we investigate a number of interesting physical cases below .
the @xmath42 contour plot for the variation of @xmath121 and @xmath122 about their predicted values from the unconstrained global fit is shown in fig . 5 , where we allow @xmath20 to be a free parameter ( unstarred labels ) or fix it at the best fit value of @xmath103 ( starred labels ) .
we show the contours for @xmath61 , 100 , etc .
overall , the ellipses one would expect from the quadratic approximation for @xmath41 in section 2 are more or less what one sees , but there is a certain asymmetry in that @xmath19 increases rather more rapidly for an increase in both @xmath123 and @xmath122 than for a corresponding decrease in both .
thus , from fig . 5 , we see that the uncertainties of the @xmath124 and @xmath122 structure functions at @xmath125 and @xmath126 ( due to the experimental errors on the data in the global fit ) are about % and @xmath127 respectively . in comparison ,
the values using the hessian approach are @xmath128 and @xmath129 respectively , in good agreement , although slightly smaller for @xmath124 . at this value of @xmath17 the uncertainties in @xmath124 and
@xmath122 have a particularly simple interpretation since @xmath124 is almost exactly proportional to the valence down quark distribution , @xmath89 , and @xmath122 is almost exactly proportional to the @xmath88 distribution .
this can clearly be seen in fig .
6 , which shows the @xmath24 and @xmath23 distributions for the extreme sets ( t*,u*,v * and w * ) corresponding to maximum and minimum @xmath124 and @xmath122 ( for the case of fixed @xmath6 ) .
rather obviously the @xmath23 distribution maximises at large @xmath17 for the case of maximum @xmath124 and minimises for minimum @xmath124 , with similar behaviour for the @xmath24 distribution and @xmath122 .
note however that in each case the extreme in the parton distribution is not precisely at @xmath125 , but at slightly higher @xmath17 , where the data are less constraining .
there are also sum rules on the partons which must be satisfied .
it is also clear that there is a strong inverse correlation between the @xmath24 and @xmath23 distributions .
this is because the data which constrain the relevant partons are the structure function measurements @xmath130 , @xmath131 and @xmath132 which are essentially proportional to @xmath133 , @xmath134 and @xmath134 respectively , where @xmath135 at @xmath125 .
this constrains @xmath24 far more than @xmath23 as we have seen , but means that for maximum variation in the partons a change in @xmath24 must be compensated by a much larger opposite change in @xmath23 .
the result that the major axis of the ellipse for given change in @xmath66 is approximately aligned along @xmath136 ( i.e. , @xmath137 ) is therefore not at all surprising .
the rate of quickest increase in @xmath19 is then along @xmath138 , where the changes in the partons add in such a way as to maximise changes in the measured structure functions .
we see that allowing @xmath6 to also vary allows the error ellipses to grow slightly , mainly in width . now
the maximum and minimum allowed values of @xmath122 ( or @xmath24 ) correspond to @xmath139 and @xmath140 and to parton sets t and v respectively .
most of the constraining data are for @xmath141 , and must be well fit , but smaller @xmath20 means slower evolution of the quarks and thus greater values of @xmath24 and @xmath122 at @xmath126 .
opposite considerations lead to the maximum @xmath122 .
since the extrema of @xmath124 and @xmath23 are more involved , due to the negative correlation with the @xmath24 distribution , they are less altered by allowing @xmath20 to vary ; see sets u and w. we see that the axes of the ellipse are essentially unchanged when @xmath20 is left free . thus fig .
6 is much the same except that the variations for parton sets t and v are a little greater than for t@xmath142 and v@xmath142 .
it is , of course , the fixed target data which constrain these cross - sections and the high-@xmath17 quarks .
it is very largely the bcdms @xmath143 measurements which are responsible for the upper extremum in @xmath124 .
the best fit tends to overshoot these data in the region of @xmath125 , and a large increase in @xmath23 makes the fit to these measurements very poor . for the extrema in @xmath122 and @xmath24 ,
the deterioration is more evenly spread over pretty much all the fixed target data at @xmath144 ( with the exception that the description of the bcdms @xmath145 measurements improves slightly ) , but the cumulative result is a very poor fit .
one of the worst instances of deterioration is for the nmc @xmath146 ratio .
the @xmath42 contour plot for the variation of @xmath4 and @xmath5 about their predicted values at the lhc energy from the unconstrained global fit is shown in fig . 7 , where again we allow @xmath20 to be a free parameter or fix it at @xmath103 .
again we show the contours for @xmath61 , 100 , etc .
this time the hessian approach should work well , although the ellipses start becoming a little rectangular . allowing @xmath6
to vary , we see that the uncertainties of the @xmath0 and @xmath31 cross - sections at the lhc ( due to the experimental errors on the data in the global fit ) are about % and @xmath147 respectively , and are positively correlated .
again this analysis also gives information on the uncertainties of particular parton distributions . to be specific ,
the parton sets which correspond to the points a , b , c , d , on the @xmath61 contour in fig . 7 , give the uncertainties in the parton distributions that dominantly determine @xmath4 and @xmath5 in the kinematic domain @xmath148 , @xmath149 gev@xmath150 relevant to @xmath0 and @xmath31 production at the lhc . the extrema in @xmath4 , represented by a and c , correspond to variations in the sea quark distributions , while the extrema in @xmath5 , represented by b and d , correspond to variations in the gluon distribution _ and _ @xmath6 .
the values of @xmath20 for sets a and c are 0.119 and 0.118 respectively , both very close to the default mrst2001 value , showing that @xmath4 , which begins at zeroth order , is insensitive to @xmath20 .
however , the values of @xmath20 for fits b and d are 0.120 and 0.117 respectively , reflecting the fact that @xmath151 .
this is well illustrated by repeating the entire analysis with @xmath20 fixed at the default value ( 0.119 ) obtained in the unconstrained global fit @xcite .
the @xmath61 and 100 contours for this additional analysis are shown by the smaller shapes in fig
we can see that the uncertainty on @xmath4 is almost unchanged , while that for @xmath5 is reduced to about @xmath129 .
the corresponding values using the hessian approach are @xmath152 and @xmath152 , in good agreement but slightly smaller in each case . the up quark distribution for each ` extrema ' set with fixed @xmath6 is shown in fig .
8(a ) and the gluon distribution in fig .
we see that indeed the parton distributions do reflect the extrema in the cross - sections in a fairly simple manner .
the quark densities at high @xmath17 show almost no variation between fits since they are well constrained at high @xmath17 and because the @xmath0 and @xmath31 production cross - sections are sensitive to the partons at an @xmath17 range centered at a few @xmath153 .
indeed the maximum and minimum @xmath0 cross - sections correspond to the maximum and minimum sea quarks for @xmath154 at @xmath155 .
the maximum and minimum higgs cross - sections correspond to the maximum and minimum gluon distributions in the same sort of range , although the large @xmath17 gluon must now decrease for increases in the small @xmath17 partons , and _
vice versa _ , in order to maintain the momentum sum rule .
the strong correlation between the two cross - sections is due to the fact that at high @xmath18 the size of the quark distribution at small @xmath17 is mainly determined by evolution , and the larger the small @xmath17 gluon the stronger the quark evolution ( and _ vice versa _ ) .
when @xmath20 is left free the resulting partons at the extrema are similar to the fixed @xmath20 results . however , in this case , their variation is a little larger at smaller @xmath18 , since the slight changes in @xmath20 lead to different rates of evolution . for the case of fixed @xmath20
the main contributions to @xmath66 come from the hera small-@xmath17 structure function data and , because of the changes in the high @xmath17 gluon , also from the tevatron jet data .
for the upper extrema in @xmath4 and @xmath5 the slope @xmath93 tends to be too large for @xmath156 , while for the lower extrema the slope is too small . in both cases
the fit to jet data deteriorates due to the shape of the high-@xmath17 gluon becoming wrong .
when @xmath6 is allowed to vary the data which are particularly sensitive to this also play a role , for example the bcdms data are fitted less well when @xmath157 in fit b , and the nmc data are described less well when @xmath139 in fit d. the corresponding @xmath42 contour plot for the tevatron is shown in fig . 9 , where again we either allow @xmath20 to be a free parameter or fix it at @xmath103 .
we see that the uncertainty of the @xmath0 cross - section at the tevatron ( due to the experimental errors on the data in the global fit ) has decreased to about @xmath158 while that for the higgs has increased to about @xmath159% for varying @xmath6 , and that the correlation has disappeared . for @xmath6
fixed at 0.119 @xmath4 is again largely unaffected , but the uncertainty of @xmath5 now more than halves to about % , reflecting the fact that this time the maximum and minimum higgs cross - sections for variable @xmath20 correspond to @xmath160 and @xmath161 respectively . with
@xmath6 fixed there is now even a very slight anti - correlation between the cross - sections .
the extrema in @xmath4 , represented by p and r , correspond roughly to variations in the quark distributions at @xmath162 , while the extrema in @xmath5 , represented by q and s , correspond to variations in the gluon distribution at @xmath163 _ and _ @xmath6 .
the values of @xmath17 sampled at the tevatron are an order of magnitude greater than at the lhc .
this , coupled with the fact that it is a proton antiproton collider , rather than a proton
proton collider , complicates the interpretation of the extremes of the cross - sections in terms of partons . the up quark distribution for each extrema set with fixed @xmath6 is shown in fig .
10(a ) and the gluon distribution in fig .
the corresponding distributions obtained when @xmath6 is allowed to vary are shown in fig .
we first consider the cases of the maximum and minimum @xmath0 cross - sections , which are insensitive to whether @xmath20 is left to vary or not .
for discussion purposes , let us consider only the @xmath24 and @xmath23 quark flavour contributions .
then the @xmath0 cross - section is roughly proportional to u(x_1)d(x_2 ) + d(x_1)u(x_2 ) + |u(x_1)|d(x_2 ) + |d(x_1 ) @xmath164 .
hence the average value of @xmath165 .
this is sufficiently large that there is a distinct difference between the quark and antiquark distributions , and the contribution to the cross - section from the quark contribution is the greater .
hence , one can decrease the cross - section by replacing a quark by its antiquark at @xmath166 , or _
vice versa_. of course , there is a fundamental constraint in doing this due to the sum rule for each valence quark .
however , the only real experimental constraint is from the ccfr @xmath167 data , all other structure function data being insensitive to the distinction between quark and antiquark . in the optimum global fit
most data would like there to be more quarks at high @xmath17 , while the ccfr @xmath167 data would prefer more valence quarks at @xmath168 .
this leads to a compromise where for the best fit the ccfr @xmath167 data at low @xmath17 are undershot .
the minimum @xmath4 is therefore achieved mainly by this exchange of quark for antiquark , which most data are happy with , and hence the deterioration in @xmath19 at p ( and p@xmath142 ) is almost entirely from the description of the ccfr @xmath167 data .
hence , both the gluon and quark distribution for p ( and p@xmath142 ) are hardly changed , as seen in figs .
10 and 11 , but @xmath169 and @xmath170 decrease for @xmath92 .
going in the other direction , an increase in @xmath171 and the consequent decrease in the valence quarks at higher @xmath17 causes a large penalty in @xmath19 and the maximum @xmath4 is achieved in a different manner . at @xmath172
the quark evolves much more slowly than at @xmath172 and the density at @xmath173 is determined largely by the input value , and modified by the rate of evolution .
hence the maximum @xmath4 is achieved by having a large quark distribution at @xmath172 at low @xmath18 and also by having an enhanced gluon at @xmath92 to increase evolution .
these are displayed in figs .
10 and 11 .
the deterioration in @xmath19 then comes mostly from the low @xmath18 quarks causing an overshooting of nmc structure function data , but there is also a contribution due to the enhanced gluon at @xmath174 causing it to be smaller for @xmath175 and hence fitting the tevatron jet data less well .
the extrema of the higgs cross - section are also slightly complicated .
it is not possible to simply increase or decrease the gluon in a range centered on @xmath172 because this is precisely the @xmath17 region where the majority of the gluon s momentum is carried , and this total is very well constrained by the momentum sum rule and the accurate high @xmath17 quark determination .
therefore , for fixed @xmath6 the change in @xmath5 is largely reliant on the fact that this total cross - section actually probes quarks within about an order of magnitude either side of the central production value of @xmath176 . hence , as we see from fig .
10 the maximum cross - section is obtained from the gluon in set q@xmath142 which is slightly reduced for @xmath177 and more enhanced for @xmath178 and the minimum cross - section is obtained from the gluon in set s@xmath142 which is slightly increased for @xmath177 and more reduced for @xmath178 . in both cases those data sets sensitive to the small @xmath17 and large @xmath17 gluon , i.e. , hera structure function data and tevatron jet data respectively , are those for which the description deteriorates .
when @xmath6 is allowed to go free it varies by about @xmath104 and there is a large increase in the variation of @xmath5 .
this is not only because @xmath179 but also because the hera data anti - correlate @xmath20 and the small @xmath17 gluon .
therefore , in set q , for example , the increased value of @xmath6 allows the small @xmath17 gluon to get much smaller , and the high @xmath17 gluon much larger , compared to set q@xmath142 .
this compensation between @xmath20 and the small @xmath17 gluon also means that hera data remains well fit , and it is the jet data ( particularly cdf ) , sensitive to large @xmath17 , and the large @xmath20-phobic bcdms data , for which the description deteriorates .
similar considerations apply to set s as compared to s@xmath142 . here
it is the d0 jet data and the small @xmath20-phobic slac and nmc data that are badly fit . for @xmath83
the hessian approach gives an uncertainty of @xmath180 for @xmath4 and @xmath147 for @xmath5 , at the tevatron energy .
in simplistic terms this is in good agreement , but a little smaller for the gluon - sensitive higgs cross - section . however , in this case we see from fig .
9 a very marked asymmetry on the contour plot . for fixed @xmath6 the ellipses
are certainly not centered on the best fit values , and for varying @xmath6 we see that @xmath19 is clearly increasing far more rapidly for increases in the predicted @xmath0 cross - section than for corresponding decreases . thus , it is clear that within the framework of this fit , increases of the cross - section of much more than @xmath181 are completely ruled out , whereas decreases of the same amount are much more acceptable .
this information would be largely lost in the hessian approach , and for these quantities the lagrange multiplier method does supply some important additional information .
the ratio of the @xmath1 to the @xmath2 production cross - sections at hadron colliders is a particularly interesting observable .
the measurement is expected to be quite precise ( better than @xmath182 at the lhc , see e.g. @xcite ) , since many of the experimental uncertainties cancel in the ratio .
the uncertainty in the prediction of the ratio at the lhc can be deduced from the @xmath42 profile shown in fig .
taking , as before , the @xmath183 measure , we obtain @xmath184 , and the hessian approach is in very good agreement with this .
since the @xmath185 ratio is sensitive to the ratio of the @xmath23 and @xmath24 quark distributions , it is not surprising that the increase in @xmath19 is almost entirely due to the nmc @xmath146 data @xcite .
a detailed discussion of the @xmath185 ratio may be found in ref .
consider , for instance , the ratio as a function of the @xmath0 rapidity @xmath186 [ eq : ratio_func_y ] , where @xmath187 at the lhc . in eq .
( [ eq : ratio_func_y ] ) we have ignored , for simplicity , the contributions involving strange and heavier quarks .
thus a measurement of the ratio at large @xmath186 would provide a direct determination of @xmath188 at large @xmath17 .
for example , for @xmath189 , we measure @xmath188 at @xmath190 at the lhc . of course
, it is the decay lepton rapidity that is measured , rather than the parent @xmath0 rapidity , and so the ratio in a given rapidity bin will have a greater uncertainty than that for @xmath191 .
the parton distribution functions of the nucleon are fundamental quantities that should , in principle , be calculable from first principles in qcd . in particular
, the @xmath17 _ moments _ of parton distributions at a given scale @xmath18 are related , by the operator product expansion , to a product of perturbatively calculable wilson coefficients and non - perturbative matrix elements of quark and gluon operators .
the latter can be computed using lattice qcd and , indeed , in recent years the precision of the lattice calculations has improved significantly .
although in principle the lattice results can be related to moments of physical structure functions , in practice it is more efficient to use parton distributions determined in a global fit to represent the physical ` data ' .
comparisons of recent lattice moment calculations with the predictions of earlier mrs parton distributions are encouraging , see for example @xcite . in order to quantify the agreement between the lattice calculations and the parton distribution predictions
it is obviously important to know the uncertainties in the latter .
it is straightforward to apply the lagrange multiplier method used in previous sections to determine the uncertainties in observable cross - sections to the moments of parton distributions . to avoid contamination from gluon contributions ,
the lattice calculations focus on the moments of non - singlet quark operators .
for example , lattice results are available for the first three moments of the combination @xmath192 , i.e. , [ eq : m_n ] m_n^u - d(q^2 ) = _ 0 ^ 1 dx x^n-1 [ u(x , q^2)-d(x , q^2 ) ] with @xmath193 .
the predictions of the mrst2001 set ( at @xmath194 ) for these moments are given in table 1 .
the @xmath41 contour plot for the ( percentage ) variation of the second and third moments about their predicted values is shown in fig . 13 .
we again show the @xmath195 and @xmath196 contours corresponding to the fixed @xmath20 analysis , but there is evidently little difference between the fixed and variable coupling results in this case .
as expected , there is a strong positive correlation between the two moments . using the @xmath195 , varying @xmath20 criterion for defining a conservative error , we obtain errors of @xmath197 , @xmath198 and @xmath199 for the second , third and fourth moments respectively .
the corresponding predictions for the errors on the moments are also given in table 1 .
the increasing relative error with increasing moment is to be expected higher moments probe the @xmath200 region where there are fewer dis structure function data .
again we notice that there is a small asymmetry in the contours the increase in @xmath19 when both moments increase being less severe than when both moments decrease .
.the moments and their errors of the ( @xmath24@xmath23 ) distribution , eq . ( [ eq : m_n ] ) , predicted at @xmath201 using mrst2001 partons @xcite . [ cols="^,^,^",options="header " , ] [ tab1 ] we also find a large variation in the values of @xmath6 extracted from the fits of the different collaborations : cteq6 @xcite , zeus @xcite , mrst2001 @xcite , h1 @xcite , alekhin @xcite and giele et al .
( gkk ) @xcite .
the resulting values of @xmath6 are listed in table 2 , together with the determination of this work ( mrst2002 ) , in order of decreasing tolerance ( @xmath202 ) , which is reflected in the size of the corresponding experimental error .
not all are presented as determinations of @xmath6 , but all are extracted using the same criteria as for the uncertainty on partons in the respective fit , and hence should be as reliable .
clearly there is a very large variation , with some very low values .
the uncertainties due to experimental errors are determined in different fashions in each case , and a summary can be found in @xcite .
we use @xmath203 for the zeus determination @xcite , because they use the offset method for determining uncertainties which for @xmath84 gives a larger uncertainty than the more common hessian method .
zeus estimate that this is equivalent to @xmath204 if they were to use the same treatment of errors as cteq .
we also use @xmath203 for the gkk value @xcite , because the uncertainties are obtained using confidence limits , but the error quoted corresponds to the one sigma usually associated with @xmath58 .
the model errors incorporate such effects as the heavy quark prescription and masses , parameterizations , changes in the starting scale of evolution _ etc_. the theory error is often determined by variation of renormalization and factorization scales , though mrst use an estimate appealing to current knowledge of nnlo and resummations , which we feel is more reliable .
since each fit is centered on nlo qcd with scales equal to @xmath18 , the `` theory errors '' are very strongly correlated , and can not therefore be responsible for the differences .
these discrepancies are undoubtedly due to the assumptions going into the fits , mainly on which data sets are included and which cuts on @xmath18 and @xmath205 are used .
mrst , who obtain the largest value of @xmath6 , use the widest range of data sets and also the least conservative cuts . as compared to mrst2001 @xcite , as seen in table 2
. we still use @xmath106 for our one - sigma uncertainty , since if @xmath83 corresponds to 90% confidence level , or 1.65 sigma , simple scaling implies that one sigma corresponds to @xmath206 , i.e. @xmath106 to a good approximation . ]
cteq use only a slightly smaller number of data sets but also cut data below @xmath201 , as described previously .
they also use a definition of the nlo coupling which truncates the solution of the renormalization group equation , whereas most other groups use the full solution of the nlo equation . both approaches are equally correct , but the truncation of the solution leads to a slightly higher value of @xmath107 at scales below @xmath207 , for the same value of @xmath6 , than the other method , and thus tends to yield a lower @xmath6 .
cteq also have a very conservative estimate of the error , though it is meant to be somewhat more than a one - sigma error .
zeus and alekhin use a similar selection of data sets , i.e. , hera dis data ( only zeus data in the former case ) and a number of fixed target dis data sets .
hence , it is unsurprising that they obtain similar central values of @xmath6 , with respective errors which are easily explained by their choices of @xmath41 .
h1 and gkk both use a small number of sets of data : the former collaboration uses h1 dis data @xcite and bcdms fixed - target proton dis data @xcite , while gkk use older h1 dis data @xcite together with bcdms and e665 @xcite fixed - target proton dis data .
both determinations are heavily influenced by the bcdms proton data set which prefers rather small from bcdms data alone @xcite . ]
@xmath6 , and this feeds into the final values .
also , both are strict in their statistical interpretation , obtaining small uncertainties , even with relatively small data samples .
finally we note that only cteq and mrst include the tevatron jet data in their analyses .
this is relevant because of the @xmath20gluon correlation .
from the discussion of the previous two subsections , it is clear that different ideas about the best way to perform a nlo fit can lead to a wide variation in both the central values and the errors of @xmath6 as well as in predictions for physical quantities such as @xmath4 and @xmath5 . the fact that the various ` nlo ' fits can yield such different outputs is disturbing , and is indicative of the uncertainty arising from theoretical assumptions .
indeed , we have always believed that ` theory ' , rather than experiment , will provide the dominant source of error @xcite .
we have already produced approximate nnlo parton distributions and predictions @xcite ( based on the approximate splitting functions @xcite obtained from the known nnlo moments @xcite ) , and find , for example , that the nnlo @xmath0 cross - section at the tevatron is @xmath208 higher than at nlo , and believe that this result is reliable .
this change is somewhat larger than the uncertainty due to experimental errors shown in fig .
moreover , @xmath0 production is likely to be subject to smaller theoretical uncertainty than many other observables particularly those directly related to the gluon .
our estimates for the uncertainty in @xmath209 at small @xmath17 are @xmath210 or more even at @xmath126 , and significantly larger at lower @xmath18 , for example .
hence , an understanding of theoretical uncertainties is clearly a priority at present , and a preliminary attempt at this will be the subject of a future publication @xcite .
we would like to thank mandy cooper - sarkar , vladimir chekelian , dan stump and wu - ki tung for useful discussions .
rst would like to thank the royal society for the award of a university research fellowship .
rgr would like to thank the leverhulme trust for the award of an emeritus fellowship .
the ippp gratefully acknowledges financial support from the uk particle physics and astronomy research council .
p. m. nadolsky and z. sullivan , `` pdf uncertainties in @xmath211 production at tevatron , '' in _ proceedings of snowmass 2001 : the future of particle physics _ ,
edited by n. graf ( stanford linear accelerator center , stanford , 2002 ) , econf c010630 , p510 , hep - ph/0110378 ; + z. sullivan and p. m. nadolsky , in _ proceedings of snowmass 2001 : the future of particle physics _ , edited by n. graf ( stanford linear accelerator center , stanford , 2002 ) , econf c010630 , p511 , hep - ph/0111358 . and @xmath23 quark distributions ( at @xmath212 and @xmath213 ) of the extrema fits which lie on the @xmath61 contour of fig .
5 for fixed @xmath103.,title="fig:",width=359 ] and @xmath23 quark distributions ( at @xmath212 and @xmath213 ) of the extrema fits which lie on the @xmath61 contour of fig . 5 for fixed @xmath103.,title="fig:",width=359 ] and @xmath213 in the extrema global fits on the @xmath61 contour of the @xmath214(lhc ) plot of fig .
7 for @xmath6 fixed at 0.119.,title="fig:",width=359 ] and @xmath213 in the extrema global fits on the @xmath61 contour of the @xmath214(lhc ) plot of fig . 7 for @xmath6 fixed at 0.119.,title="fig:",width=359 ] and
@xmath213 found in the extrema global fits on the @xmath61 contour of the @xmath214(tevatron ) plot of fig . 9 with @xmath6 fixed at 0.119.,title="fig:",width=359 ] and @xmath213 found in the extrema global fits on the @xmath61 contour of the @xmath214(tevatron ) plot of fig . 9 with @xmath6 fixed at 0.119.,title="fig:",width=359 ] ,
@xmath5 at the lhc and tevatron energies , shown on the @xmath41 contour plots centered on the mrst2001 partons @xcite .
the @xmath42 contours are taken from figs .
7 and 9 respectively , for the case in which @xmath6 is a free parameter . the inner contour with @xmath61 is taken to represent the error on the observables @xmath4 and @xmath5 arising from the _ experimental _ errors of the data that are used in the global fit.,title="fig:",width=302 ] , @xmath5 at the lhc and tevatron energies , shown on the @xmath41 contour plots centered on the mrst2001 partons @xcite .
the @xmath42 contours are taken from figs . 7 and 9
respectively , for the case in which @xmath6 is a free parameter . the inner contour with @xmath61 is taken to represent the error on the observables @xmath4 and @xmath5 arising from the _ experimental _ errors of the data that are used in the global fit.,title="fig:",width=302 ] | we determine the uncertainties on observables arising from the errors on the experimental data that are fitted in the global mrst2001 parton analysis . by diagonalizing the error matrix
we produce sets of partons suitable for use within the framework of linear propagation of errors , which is the most convenient method for calculating the uncertainties . despite the potential limitations of this approach we find that it can be made to work well in practice .
this is confirmed by our alternative approach of using the more rigorous lagrange multiplier method to determine the errors on physical quantities directly .
as particular examples we determine the uncertainties on the predictions of the charged - current deep - inelastic structure functions , on the cross - sections for @xmath0 production and for higgs boson production via gluon
gluon fusion at the tevatron and the lhc , on the ratio of @xmath1 to @xmath2 production at the lhc and on the moments of the non - singlet quark distributions .
we discuss the corresponding uncertainties on the parton distributions in the relevant @xmath3 domains . finally , we briefly look at uncertainties related to the fit procedure , stressing their importance and using @xmath4 , @xmath5 and extractions of @xmath6 as examples . as a by - product of this last point we present a slightly updated set of parton distributions , mrst2002 . plus 2 mm minus 2 mm 23.0 cm 17.0 cm -1.0 in -42pt ippp/02/49 + dcpt/02/98 + cavendish - hep-2002/10 + 5 november 2002 + * uncertainties of predictions from parton distributions + i : experimental errors * a.d .
martin@xmath7 , r.g .
roberts@xmath8 , w.j .
stirling@xmath7 and r.s .
thorne@xmath9 + @xmath7 institute for particle physics phenomenology , university of durham , dh1 3le , uk + @xmath8 rutherford appleton laboratory , chilton , didcot , oxon , ox11 0qx , uk + @xmath10 cavendish laboratory , university of cambridge , + madingley road , cambridge , cb3 0he , uk |
in the field of heavy hadrons , it is noticed that some experimental observations obviously deviate from our theoretical predictions , so definitely , such anomalies " need to be clarified .
supposing the experimental measurements are right , there must be some loopholes in our present understanding of the nature , either there exists a contribution from new physics beyond the standard model ( sm ) , or the concerned hadrons have exotic structure such as hybrid , multi - quark etc .
it is believed that radiative decays on this aspect provide an ideal lab to testify the hadron structures and as well help to identify the quantum numbers for a newly observed resonances .
even though , the reaction mechanism for radiative decays is free of strong interaction ( at tree level ) , the non - perturbative qcd effects are still involved in the hadronization of emerging hadrons , as well as the wavefunction of the parent hadron . to search for new physics or possible exotic components of the hadron , one must thoroughly study how the non - perturbative qcd affects the decay rates in a reasonable theoretical framework . in this work
, we are going to derive the formulas for the radiative decays of @xmath0 and @xmath9 heavy mesons which are supposed to have regular @xmath10 structures , in the light front quark model ( lfqm ) . through a comparison of the theoretical predictions with data ,
the consistency degree would reveal if the concerned hadron possesses the regular @xmath10 structure or has an exotic component . during recent years several new particles such as @xmath5@xcite , @xmath11@xcite , @xmath12@xcite , @xmath13@xcite and @xmath14@xcite
have been observed in experiments and re - confirmed as new resonances .
however it is not easy to identify their inner structures , i.e. if the constituents of the resonances are indeed just @xmath15 and @xmath16 or with something else ( gluon or @xmath17 pair etc . ) .
some of them are speculated as exotic states .
the reasons may be twofold : one is their peculiar decay modes whereas the other is that the theoretical expectations on their excited states are not well consistent with data or even missing .
for example when @xmath5 was announced it was considered as a molecular state or a tetraquark , however it may just be a regular @xmath18wave @xmath19 meson with @xmath20 as it appears in particle data book@xcite . because radiative decay is fully governed by the electromagnetic interaction ,
the reaction mechanism is relatively simple and the emitted photon can be well measured in experiment , it may offer a good opportunity to justify the quantum numbers and constituent structure of the newly observed resonance .
namely , one may check whether with the simple @xmath10 assignment the theoretical prediction agrees with data . for @xmath18wave particles
there are three degenerate states @xmath0 , @xmath21 and @xmath22 with the total intrinsic spin @xmath23 and one singlet @xmath9 with @xmath24 .
it is well known that @xmath25 and @xmath26 and @xmath27 are triplets and singlet @xmath18wave charmonia respectively . in this work
we will calculate the rates of @xmath28 and @xmath29 in lfqm @xcite which has been successfully applied to evaluate rates of semileptonic and non - leptonic decays of @xmath30wave heavy mesons . for the @xmath18wave mesons ,
the wave functions of @xmath0 , @xmath31 , @xmath9 and @xmath32 have been constructed and the leading feynman diagrams are simple , with them we are able to calculate the corresponding transition amplitudes . since the branching ratio of @xmath33 is well measured , we first calculate @xmath34 in lfqm to fix the model parameters and check the validity degree of this approach , then with the formulas for the transition @xmath35 we estimate the rates of @xmath36 and @xmath37 .
these formulas can also be applied to study decays of other @xmath0 states .
recently the spin singlets @xmath27 and @xmath38 attract intensive interests of both experimentalists and theorists@xcite . by the formulas for @xmath29 we calculate the widths of @xmath6 and @xmath7 .
the results can be compared with the data which will be available soon at the bes ii and future b - factory .
this comparison definitely assures us if they are pure @xmath18wave heavy quarkonia . after the introduction we derive the formulas for the transition @xmath35 and @xmath39 in section ii . then in sec .
iii , we numerically evaluate the decay widths of @xmath33 , @xmath36 , @xmath37 , @xmath6 and @xmath7 and make some discussions . in the last section we give a brief summary and discussion .
some notations and definitions of relevant quantities are collected in the attached appendix .
in ref.@xcite chung studied various transitions and derived the corresponding amplitudes , for example , he determined the amplitude structure of @xmath40 . for transition @xmath29 @xcite
a photon exists in the final states , thus the gauge invariance demands the transition amplitude to be in the form @xmath41 where @xmath15 and @xmath42 represent the momenta of the photon and daughter meson , @xmath43 and @xmath44 are the polarizations of the vector - meson and photon respectively , and @xmath45 is the form factor which is what we are going to derive and numerically compute in this work . though the @xmath46 was not discussed in ref.@xcite , its amplitude structure is the same as eq.([p1 ] ) and it can be seen from analyzing the @xmath47 characters of the involved mesons .
below , following the schemes given in literature , we will calculate the transition amplitudes of interest in lfqm .
the vertex functions of @xmath0 and @xmath31 are respectively @xcite @xmath48\end{aligned}\ ] ] where @xmath49 , @xmath50 and @xmath51 are defined in ref.@xcite and @xmath52 and @xmath53 are the momenta of constituents of the corresponding meson .
ccc the transition amplitude corresponding to the left diagram of fig.1 is written as @xmath54 where @xmath55(\slash\!\!\!p'_1+m_1)\gamma_\mu(\slash\!\!\!p_1+m_1)\},\ ] ] @xmath56 , @xmath57 , @xmath58 , @xmath59 is the electric charge of the quark of u- or d - types . in the light front frame , @xmath60 is decomposed as ( @xmath61 ) . integrating out @xmath62 with the methods given in ref.@xcite one has @xmath63 with @xmath64 where @xmath65 and @xmath66 represent the masses of decaying and produced mesons and the relation between @xmath67 and @xmath68 can be found in the appendix of ref.@xcite . to include the contributions from the zero mode , @xmath69 , @xmath70 , @xmath71 and @xmath51 in @xmath72 should be replaced by appropriate expressions as discussed in ref.@xcite , for example @xmath73 with @xmath74 and @xmath75 is the momentum of decaying meson .
the definitions of @xmath76 and @xmath77 are presented in the appendix , as more details about the derivations and relevant notations can be found in ref.@xcite . in this framework
, @xmath78 replaces @xmath72 and is written into the form @xmath79 with @xmath80 - \nonumber\\&&\frac{4\,{a12}\,\left ( 2\,{{m_1}}^2 + 4\,{m_1}\,{m_2 } + 2\,{{m_2}}^2 - { { m}}^2 - { { m'}}^2 + { q^2 } - 2\,{m_1}\,{w_v } - 2\,{m_2}\,{w_v } + 2\,{z2 } \right ) } { { w_v}},\nonumber\\ { f_2}=&&\frac{4\,k\cdot q\,{a_2^{(1)}}\,\left ( -{\hat{n}_1 } - { \hat{n}_1 ' } + { q^2 } \right ) } { { w_v } } + \frac{4\,k\cdot q\,\left ( { \hat{n}_1 } + { \hat{n}_1 ' } - { q^2 } + { m_1}\,{w_v } \right ) } { { w_v } } + \nonumber\\ & & \frac{4\,k\cdot q\,{a_2^{(2)}}\,\left\{-2\,\left [ 2\,{\left ( { m_1 } + { m_2 } \right ) } ^2 - { { m}}^2 - { { m'}}^2 + { q^2 } \right ] + 4\,{m_1}\,{w_v } + 4\,{m_2}\,{w_v } \right\ } } { { w_v } } + \nonumber\\ & & \frac{4\,k\cdot q\,{a_3^{(2)}}\,\left\ { -2\,\left [ 2\,{\left ( { m_1 } + { m_2 } \right ) } ^2 - { { m}}^2 - { { m'}}^2 + { q^2 } \right ] + 4\,{m_1}\,{w_v } + 4\,{m_2}\,{w_v } \right\ } } { { w_v } } + \nonumber\\ & & \frac{4\,k\cdot q\,{a_1^{(1)}}\,\left [ 4\,{\left ( { m_1 } + { m_2 } \right ) } ^2 - 2\,{{m}}^2 - 2\,{{m'}}^2 - { \hat{n}_1 } - { \hat{n}_1 ' } + 3\,{q^2 } - 2\,\left ( 3\,{m_1 } + { m_2 } \right ) \,{w_v } \right ] }
{ { w_v}},\end{aligned}\ ] ] and @xmath81 are presented in the attached appendix .
we define the form factors as following @xmath82 where @xmath83 and @xmath84 are respectively the wavefunctions of the initial and final mesons .
these form factors will be numerically evaluated in next section . with these form factors the amplitude corresponding to the left diagram of fig.1
is obtained as @xmath85 the right diagram is just the charge conjugation of the left one of fig.1 , so that one can immediately write it down @xmath86 the total amplitude is simply a sum of @xmath87 and @xmath88 : @xmath89 comparing eq .
( 1 ) with eq .
( 12 ) we determine the full form factor as @xmath90.\end{aligned}\ ] ] in principle one can calculate @xmath91 and @xmath92 separately and fix the form factor , then go on obtaining the decay rate .
the vertex functions of @xmath32 and @xmath9 are presented in ref.@xcite , they are respectively @xmath93\gamma_5,\end{aligned}\ ] ] where @xmath94 , @xmath95 and @xmath96 are some relevant functions which are slightly lengthy and can be found in ref.@xcite , for saving space , we do not repeat them here . the transition amplitude corresponding to the left diagram of fig.1 is written as @xmath97 where @xmath98(\slash\!\!\!p'_1+m_1)\gamma_\mu(\slash\!\!\!p_1+m_1)\gamma_5\}.\ ] ] integrating out @xmath62 we have @xmath99 where @xmath100 with a replacement similar to eq.([p6 ] ) , @xmath78 is re - written into the form of eq .
( [ p7 ] ) with @xmath101 .\end{aligned}\ ] ]
in this section we first test the formula by comparing our theoretical evaluation on the rate of a well measured decay mode with the data and confirm its validity , then apply it to predict the rates of radiative decays of @xmath0 and @xmath9 which will be experimentally measured soon .
we select transition @xmath33 as a probe to check the approach since its branching ratio is well measured . setting @xmath102gev and the model parameters @xmath103gev@xcite10% range to estimate the corresponding errors . ] , we get the form factors @xmath104 and @xmath105 for the transition @xmath33 which are presented in tab . i. one can immediately notice that @xmath104 is sensitive to the variation of the parameters , especially @xmath106 , but @xmath105 is insensitive .
the problem originates from the fact that @xmath107 is proportional to a cancelation of two large numbers .
even though @xmath107 resides in both @xmath108 and @xmath109 , its contribution is suppressed by @xmath110 or @xmath111 in @xmath109 , so that @xmath109 is insensitive to the parameters . for reducing uncertainties of our theoretical computations
, we will calculate @xmath45 in terms of the relation between @xmath105 and @xmath45 in eq.([c1 ] ) .
our theoretical estimate of the width of @xmath33 is ( 152@xmath11231)kev and the corresponding branching ratio is ( [email protected])% which is consistent with the data @xmath113%@xcite within a tolerable error range .
then we study the transition @xmath3 using the parameter @xmath114gev and @xmath115gev which were fixed in ref.@xcite by fitting other well measured channels .
our estimate of @xmath116 is ( [email protected])kev . with the branching ratio ( [email protected])%@xcite , the total width of @xmath117 is estimated to be ( [email protected] ) mev .
.the form factors @xmath104 and @xmath105 [ cols="^,^,^,^,^,^",options="header " , ] supposing @xmath5 to be a regular bound state of @xmath19 and setting @xmath118gev , @xmath119gev@xcite , we can calculate the rate of @xmath8 using the above formulas .
we obtain its partial width as ( [email protected])kev .
namely , if @xmath5 is a @xmath0 regular meson , possessing the @xmath18wave @xmath19 structure , our numerical prediction should be consistent with the data which will be available at bes ii or b - factory soon , therefore the consistency degree with data can help to confirm or negate its assignment @xmath120 .
in fact these formulas can be applied to study radiative decays of other heavy mesons such @xmath121 , @xmath122 , @xmath123 and @xmath124 .
once @xmath27 and @xmath38 were experimentally measured , their special behaviors draw intensive interests of theorists . the main focus is if they are regular quarkonia or have exotic components .
using the formulas for @xmath125 , we calculate the branching ratios of @xmath6 and @xmath126 and the numerical result are presented in tab .
i , which can be used to analyze their characters . in terms of the measured upper limit of the total width of @xmath27 we estimate the branching ratio of @xmath127 which is a little larger than the data @xmath128 .
if @xmath27 ( @xmath38 ) is a pure charmonium ( bottomonium ) our estimate indicates that its total width should be around 1 mev ( 80kev ) . to further identify the inner structures of @xmath27 and @xmath38 ,
more precise experiments are needed .
it is noted that the masses of initial and finial mesons we used in our numerical computations are taken from ref.@xcite except the mass of @xmath129 which is 9402.4mev@xcite . some parallel researches on these decays
can be found in ref.@xcite and a comparison of their results with ours is made and listed in tab .
[ tab : com ] . because of large errors in the inputs for our numerical computations , our predictions are of relatively large uncertainties . to further testify the model or constrain the model parameters ,
much more precise experiments are needed .
in this work we derive the formulas for the radiative decay of @xmath0 and @xmath9 heavy mesons and numerically compute the rates . we formulate the transition matrix elements and extract the form factors in the lfqm . to check the validity degree of this approach where the model parameters were fixed by fitting other physical processes in previous works , we calculate the decay widths of @xmath33 and compare it with data . considering both theoretical and measurement uncertainties , the theoretical result on the branching ratio of @xmath33 ( [email protected])% is satisfactorily consistent with data @xmath130 . with the same scenario
, we predict the width of @xmath3 to be about ( [email protected])kev and the width of @xmath8 to be ( [email protected])kev . comparing those results with the data which will be available at bes and b - factory , one can further confirm the inner structure of @xmath117 and @xmath5 . in terms of the data
@xmath131 we estimate the total width of @xmath117 as ( [email protected])mev which would be easier to be experimentally checked . as for the @xmath9 decay
, we evaluate the decay rates of @xmath132 and @xmath126 which are @xmath133kev and @xmath134 kev respectively .
these formulas deduced in this work can also be applied to study radiative decays of other @xmath0 and @xmath9 particles .
this work is supported by the national natural science foundation of china ( nnsfc ) under the contract no .
11075079 and no . 11005079
; the special grant for the ph.d .
program of ministry of eduction of p.r .
in the conventional light - front model , a meson containing a quark @xmath135 and an antiquark @xmath136 with its total momentum @xmath137 and spin @xmath138 can be expressed@xcite @xmath139 where @xmath140 is wave function in momentum - space , @xmath141 and @xmath142 denote helicities , @xmath143 are the on - mass - shell light - front momenta defined by @xmath144 and @xmath145 from the eigen - equation@xcite @xmath146 one can deduce a light - front bethe - salpeter equation @xmath147 in principle by solving eq.([bse ] ) one can obtain the momentum distribution amplitude @xmath148 .
however the two - body interaction kernel @xmath149 is complicate so some phenomenological amplitudes are chosen in practicable theoretical calculations . in this work we use the gaussian - type amplitudes , for example @xmath150 where @xmath151 is the model parameter to be fixed by calculating the decay constant .
a parallel approach about lfqm was done in refs.@xcite where the gaussian - type amplitude is adopted as a trial function for the variational computations of the qcd - motivated effective hamiltonian .
here we list some variables and notations appearing in the context .
the incoming meson in fig .
[ t1 ] has momentum @xmath152 where @xmath153 and @xmath154 are the momenta of the off - shell quark and antiquark and @xmath155 with @xmath156 and @xmath157 are internal variables and @xmath158 .
99 b. aubert _ et al . _ [ babar collaboration ] , phys .
lett . * 90 * , 242001 ( 2003 ) [ hep - ex/0304021 ] .
d. besson _ et al .
_ [ cleo collaboration ] , phys . rev .
d * 68 * , 032002 ( 2003 ) [ erratum - ibid .
d * 75 * , 119908 ( 2007 ) ] [ hep - ex/0305100 ] .
p. krokovny _ et al .
_ [ belle collaboration ] , phys .
lett . * 91 * , 262002 ( 2003 ) [ hep - ex/0308019 ] .
s. k. choi _ et al . _ [ belle collaboration ] , phys .
lett . * 91 * , 262001 ( 2003 ) [ arxiv : hep - ex/0309032 ] .
k. abe _ et al . _ [ belle collaboration ] , phys . rev
. lett .
* 98 * , 082001 ( 2007 ) [ arxiv : hep - ex/0507019 ] . h. w. ke , x. q. li and z. t. wei , phys .
d * 77 * , 014020 ( 2008 ) [ arxiv:0710.1927 [ hep - ph ] ] ; z. t. wei , h. w. ke and x. q. li , phys .
d * 80 * , 094016 ( 2009 ) [ arxiv:0909.0100 [ hep - ph ] ] ; h. w. ke , x. q. li and z. t. wei , phys .
d * 80 * , 074030 ( 2009 ) [ arxiv:0907.5465 [ hep - ph ] ] ; h. w. ke , x. q. li and z. t. wei , eur .
j. c * 69 * , 133 ( 2010 ) [ arxiv:0912.4094 [ hep - ph ] ] ; h. w. ke , x. h. yuan and x. q. li , int
a * 26 * , 4731 ( 2010 ) , arxiv:1101.3407 [ hep - ph ] . h. w. ke and x. q. li , eur .
j. c * 71 * , 1776 ( 2011 ) [ arxiv:1104.3996 [ hep - ph ] ] ; h. w. ke and x. q. li , eur .
j. c * 71 * , 1776 ( 2011 ) [ arxiv:1104.3996 [ hep - ph ] ] .
w. jaus , phys .
d * 41 * , 3394 ( 1990 ) ; d * 44 * , 2851 ( 1991 ) ; w. jaus , phys . rev .
d * 60 * , 054026 ( 1999 ) . c. r. ji ,
p. l. chung and s. r. cotanch , phys .
d * 45 * , 4214 ( 1992 ) .
h. y. cheng , c. y. cheung and c. w. hwang , phys .
d * 55 * , 1559 ( 1997 ) [ arxiv : hep - ph/9607332 ] .
j. -z .
li , y. -q .
ma and k. -t .
chao , arxiv:1209.4011 [ hep - ph ] .
s. fleming and t. mehen , phys .
d * 58 * , 037503 ( 1998 ) [ hep - ph/9801328 ] .
s. godfrey , `` production of the h(c ) and h(b ) and implications for quarkonium spectroscopy , '' j. phys .
* 9 * , 123 ( 2005 ) [ hep - ph/0501083 ] .
f. de fazio , phys .
d * 79 * , 054015 ( 2009 ) [ erratum - ibid .
d * 83 * , 099901 ( 2011 ) ] [ arxiv:0812.0716 [ hep - ph ] ] ; p. colangelo , f. de fazio and a. ozpineci , phys .
d * 72 * , 074004 ( 2005 ) [ hep - ph/0505195 ] . c. -y .
cheung , w. -m .
zhang and g. -l .
lin , phys .
d * 52 * , 2915 ( 1995 ) [ hep - ph/9505232 ] .
h. -m . choi and c. -r .
ji , phys .
d * 59 * , 074015 ( 1999 ) [ hep - ph/9711450 ] .
h. -m .
choi and c. -r .
ji , phys . lett .
b * 460 * , 461 ( 1999 ) [ hep - ph/9903496 ] . | the radiative decay is believed to be an ideal lab to study hadronic structure of newly observed resonances because the reactions are governed by only the electromagnetic interaction ( tree level ) . however , to obtain correct theoretical values , one has to properly deal with the non - perturbative qcd effects in the wavefunction and hadronization . in this work
we derive the formulas for the radiative decays of @xmath0 and @xmath1 heavy mesons in the light front quark model ( lfqm ) .
because @xmath2 is well measured , the theoretical evaluation of the transition rate can be used to test our approach . within this theoretical framework ,
the width of @xmath3 is evaluated .
the formulas can be applied to identify the inner structures of new resonances , for example the isospin of @xmath4 and the structure of @xmath5 , via processes @xmath6 , @xmath7 and @xmath8 . |
we derive the vulf correlation function @xmath26 by generalizing the formalism of quantum optics to massive spin-0 bosons and quasi - maxwellian velocity distribution of dm fields .
we use the natural units and start the derivation in the galactic reference frame and then transform the result into the moving device frame .
@xmath153 , where @xmath154 is the density matrix and @xmath155 are the field operators , @xmath156 . here
@xmath157 is the quantization volume , @xmath158 and @xmath159 are 4-momentum and 4-position vectors , @xmath160 and @xmath161 are bosonic creation and annihilation operators .
the summation is carried out over the field modes with frequencies @xmath162 .
taking into account that the mode occupation numbers @xmath163 are macroscopic , @xmath164 .
further , the mode occupation numbers is related to the local dm energy density @xmath165 and dm velocity distribution @xmath166 as @xmath167 .
the requirement that the occupation numbers are macroscopic , @xmath168 , leads to @xmath169 .
this explains the upper limit on vulf masses of @xmath170 used in the main text .
the resulting two - point correlation function reads ( restoring fundamental constants ) @xmath171 given the dm velocity distribution @xcite @xmath172 this expression can be evaluated numerically .
analytical result can be obtained by taking the maxwellian distribution of the standard halo model , @xmath173 , where @xmath40 is the virial velocity and @xmath174 is the earth s velocity in the galactic reference frame .
we further take the galactic escape velocity cutoff to be infinite and neglect the non - relativistic kinetic energy correction in the denominator .
the resulting correlation function is given by eq .
( [ eq : fieldcorrfuncanalyl ] ) of the main text .
28ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ]
+ 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty link:\doibase 10.1146/annurev - astro-082708 - 101659 [ * * , ( ) ] link:\doibase 10.1016/j.physrep.2004.08.031 [ * * , ( ) ] , link:\doibase 10.1103/physrevlett.115.221801 [ * * , ( ) ] , link:\doibase 10.1103/physrevd.91.015015 [ * * , ( ) ] , link:\doibase 10.1103/physrevd.93.075029 [ * * , ( ) ] , http://arxiv.org/abs/1605.04048 [ ( ) ] , link:\doibase 10.1103/physrevx.4.021030 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.113.161801 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.114.161301 [ * * , ( ) ] , link:\doibase 10.1103/physrevlett.116.031102 [ * * , ( ) ] , link:\doibase 10.1063/pt.3.2896 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.110.021803 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1088/1742 - 6596/378/1/012012 [ * * , ( ) ] , http://arxiv.org/abs/1308.1703 http://stacks.iop.org/0004-637x/784/i=2/a=161 [ * * , ( ) ] @noop * * , ( ) , @noop _ _ , , link:\doibase 10.1088/1475 - 7516/2010/08/004 [ * * , ( ) ] , link:\doibase 10.1103/physrevlett.115.011802 [ * * , ( ) ] link:\doibase 10.1002/andp.201300061 [ * * , ( ) ] link:\doibase 10.1007/s10714 - 014 - 1730 - 2 [ * * , ( ) ] , link:\doibase 10.1103/revmodphys.83.331 [ * * , ( ) ] link:\doibase 10.1103/physrevx.6.021016 [ * * , ( ) ] http://arxiv.org/abs/1511.07735 [ ] link:\doibase 10.1063/1.1149844 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.110.133001 [ * * , ( ) ] link:\doibase 10.1126/science.1227460 [ * * , ( ) ] link:\doibase 10.1111/j.1365 - 2966.2009.14630.x [ * * , ( ) ] , | virialized ultra - light fields ( vulfs ) while being viable cold dark matter candidates can also solve the standard model hierarchy problem .
direct searches for vulfs due to their non - particle nature require low - energy precision measurement tools . while the previous proposals have focused on detecting coherent oscillations of the measured signals at the vulf compton frequencies , here we exploit the fact that vulfs are essentially dark matter _ waves _ and as such they carry both temporal and spatial phase information .
thereby the discovery reach can be improved by using distributed networks of precision measurement tools .
we find the expected dark - matter signal by deriving the spatio - temporal two - point vulf correlation function .
based on the developed understanding of coherence properties of dark - matter fields , we propose several experiments for dark matter wave detection . in the most basic version , the modifications to already running experiments are minor and only require gps - assisted time - stamping of data . we also derive the expected dark matter line profile for individual detectors . exacting
the microscopic nature of dark matter ( dm ) is one of grand challenges of modern physics and cosmology @xcite .
the challenge lies in extrapolating from the observational galactic to the laboratory or microscopic scales .
it was recently realized @xcite that ultralight fields , while being cold dm candidates , can also solve the hierarchy problem elevating such fields to the important class of `` economic '' dm models such as weakly - interacting massive particles and axions @xcite .
we will refer to such ultralight fields as vulfs ( virialized ultra - light fields ) .
direct searches for vulfs due to their non - particle nature require low - energy precision measurements .
such measurements , with their exquisite precision , have been historically important in powerfully constraining new physics beyond the standard model ( sm ) .
individual vulf dm search proposals cover a broad range of experiments @xcite : atomic clocks , magnetometers , accelerometers , interferometers , cavities , resonators , permanent electric - dipole and parity - violation measurements , and extend to gravitational wave detectors .
the sought dm signature in these proposals is dm - induced oscillations of the measured signals at the vulf compton frequencies at the device location .
an important point is that vulfs are _ waves _ , and while they do induce an oscillating in time signal at a given spatial location , dm signals at different locations have a fixed phase relation ( see fig . [
fig : mapsatsearth](a ) ) , i.e. , the signals are correlated .
based on this observation , here we argue that a wider discovery reach can be gained by sampling the dm wave at several locations via a network of precision measurement tools .
further , the vulf signal is composed out of interfering waves traveling at different velocities and in different directions ; thereby the problem of relating signals at different space - time locations requires computations of dark - matter correlation functions , derived here .
based on these ideas and derivations , we propose a number of dm wave detection experiments . in the most basic version , the modifications to already running experiments are minor and only require simple gps - assisted time - stamping of data taking @xcite .
previously a network of precision measurement devices have been proposed for detecting topological defects sweeping through the networks @xcite ; here we show that such networks can be also used as discovery tools for vulf dark matter . ( a ) dark - matter wave observatory based on a global network of existing low - energy precision measurement laboratories ( red dots ) around the globe ; ( b ) satellite mission for probing vulf dm correlation function ; both the distance between the satellites and the angle between galactic velocity @xmath0 and separation @xmath1 vectors can be varied .
( c ) terrestrial experiment with fixed nodes utilizes the daily variation of the angle between galactic velocity and two - node separation vector . ] in the vulf models , dark matter is composed of ultralight spin 0 bosonic fields , oscillating at their compton frequency @xmath2 , where @xmath3 is the boson mass ( see review @xcite ) .
the frequencies can span many orders of magnitude : @xmath4 for @xmath5 . here
the lower bound comes from requiring that the compton wavelength is smaller than the galactic size and the upper limit from requiring that number of particles per de broglie volume is macroscopic ( see supplementary materials ) .
the proposals @xcite have focused on searching for an oscillating signal at the compton frequency . unfortunately , in a laboratory environment , an observation of an oscillating signal could be ascribed to some mundane ambient noise and it desirable to establish additional dm signatures .
here we derive vulf spatiotemporal correlation functions to probe the dm coherence .
qualitatively , the coherence times and coherence lengths of the signal are related to dm properties . indeed , in the standard halo model ( see , e.g. ,
@xcite ) , during the galaxy formation , as dm constituents fall into the gravitational potential , their velocity distribution in the galactic reference frame becomes quasi - maxwellian with a characteristic dispersion ( virial ) velocity @xmath6 , @xmath7 and a cut - off at the galactic escape velocity .
this velocity distribution leads to spectral broadening of oscillations ( dephasing ) characterized by the coherence time @xmath8 .
the velocity distribution also results in a spatial dispersion of individual wave packets , leading to the coherence length @xmath9 .
additional commonality of all the enumerated proposals is the coupling of dm fields to particles and fields in terms of the so - called portals , when the gauge - invariant operators of the sm fields @xmath10 are coupled to the operators involving dm fields @xcite .
one of possibilities is the linear coupling ( higgs or linear portal ) , @xmath11 here @xmath10 are various pieces from the sm lagrangian , @xmath12 such as the fermion rest mass energies , electromagnetic field tensor contribution , gluon field contribution , etc .
@xmath13 are the energy scales , characterizing the strength of the coupling .
quite naturally this portal when combined with the sm lagrangian , leads to variation of fundamental constants , e.g. , the electron rest mass @xmath14 is modulated by dm field as @xmath15 @xmath16 or the fine structure constant @xmath17 @xmath18 , where @xmath14 and @xmath19 are unperturbed quantities .
vulf fields oscillate at compton frequencies , leading to oscillating corrections to fundamental constants @xcite . a typical apparatus measures a signal @xmath20 associated with a piece @xmath10 of the sm lagrangian ( or their combination ) . for example , an optical clock measures transition frequencies that depend on @xmath19 .
then the measured quantity has a dm - induced admixture @xmath21 that is proportional to the field value @xmath22 at the device location .
thereby in the assumption of higgs portal , the correlation between two devices dm signals can be expressed in terms of the two - point dm field correlation function @xmath23 : @xmath24 we could also introduce a correlation function for space - time variation of fundamental constants , e.g. @xmath25 ; it is expressed in terms of dm field correlation function .
we derive the vulf correlation function @xmath26 by generalizing the formalism of quantum optics to massive spin-0 bosons and quasi - maxwellian velocity distribution of dm fields .
the derivation is given in supplementary materials .
the resulting correlation function reads@xmath27 here @xmath28 is the doppler - shifted value of the compton frequency @xmath29 and @xmath30 is the `` galactic '' wave vector associated with the apparatus motion through the dark matter halo .
the effective field amplitude @xmath31 , and correlation amplitude @xmath32 and phase @xmath33 are defined as @xmath34 where coherence time @xmath35 and length @xmath36 are expressed in terms of the virial velocity @xmath37 .
thereby the correlation function encodes the priors on vulfs and dm halo , such as the dm energy density in the vicinity of the solar system @xcite , @xmath38 , motion through the dm halo at @xmath39 and the virial velocity @xmath40 . as such it provides an improved statistical confidence in the event of an observation of dm signal .
now with the correlation function at hand , we explore the vast vulf parameter landscape .
typical values of various parameters are compiled in table [ tab : vulf - params ] .
coherence time is roughly a million of compton periods .
the coherence length can be interpreted as the de broglie wavelength of a particle moving at virial velocity and it is thousand times larger than the compton wavelength .
the number density of vulf particles @xmath41 ranges from @xmath42 to @xmath43 for the indicated masses , i.e. , a typical device interacts with a macroscopic number of dm particles .
compton frequencies range from nhz to phz .
notice that one oscillation per year corresponds to @xmath44 ( @xmath45 ) . as points of reference for the coherence length ,
the size of our galaxy is @xmath46 and the earth diameter @xmath47 .
llll + @xmath48 & @xmath49 & @xmath50 & + + @xmath51 & @xmath52 & @xmath53 & @xmath54 + @xmath55 & @xmath56 & @xmath57 & @xmath58 + @xmath59 & @xmath60 & @xmath61 & @xmath62 + @xmath63 & @xmath64 & @xmath65 & @xmath66 + @xmath67 & @xmath68 & @xmath69 & @xmath70 + @xmath71 & @xmath72 & @xmath73 & @xmath74 + in the limit when both coherence length @xmath75 and time @xmath76 are infinitely large one recovers the fully coherent wave correlation function @xmath77 for a single geographic location the associated power spectral density ( psd ) in frequency space is a spike at the doppler - shifted compton frequency .
searching for such resonances has been proposed @xcite for atomic clocks and the first spectroscopic results were reported @xcite . for a global network ,
the spatial correlations come from the dot product @xmath78 .
a natural question is how to probe various distances . in galactic coordinates
, the sun moves through the halo in the @xmath79 direction , roughly towards the cygnus constellation .
the most straightforward approach is to place two satellites with , for example , two precision clocks onboard in the vicinity of a lagrange point , orient their relative position vector @xmath1 towards cygnus constellation ( see fig . [
fig : mapsatsearth](b ) ) and carry out a series of measurements at various distances .
this would map out the spatial part of the correlation function .
more practical approach is to carry out terrestrial experiments ( see fig .
[ fig : mapsatsearth](c ) ) with fixed positions of network nodes and rely on the angular dependence of the @xmath80 of the phase .
the angle @xmath81 would change due to the earth rotation .
the inclination angle of the cygnus constellation ( i.e. direction of motion through the halo ) as seen from the earth is about @xmath82 . as an example , consider two geographically remote labs located along the 45th parallel ( see fig .
[ fig : mapsatsearth](c ) ) . their two - node correlation function would sample @xmath83 in the range from zero to @xmath84 over the course of one day .
it is worth emphasizing that the coherence length @xmath85 , thereby two sites have the most sensitivity for @xmath86 . the geographic arrangement in fig .
[ fig : mapsatsearth](c ) serves as an illustrative example as simply time - synchronized measurements in existing laboratories @xcite , fig .
[ fig : mapsatsearth](a ) , should be sufficient for the proposed global data analysis .
gnome network @xcite infrastructure can serve as a natural host for the vulf dm observatory .
one could also use clocks on numerous navigational satellites , such as gps , to search for correlation patterns @xcite ; these have an advantage of half - a - day orbits ( gps ) and a large @xmath8750,000 km aperture .
elisa gravitational wave mission @xcite ( a network of three satellites ) can also be used for vulf detection .
another point is that the network does not need to be homogenous , and various precision measurement tools can be included in the network . indeed , as long as the dm portal is linear , eq.([eq : linearportal ] ) , such a global observatory can `` cast a much wider net '' on possible dm couplings .
compared to an individual device , the statistical sensitivity of a correlated network is improved by the factor @xmath88 , where @xmath89 is the number of nodes .
it is further improved through the sampling of the @xmath83 phase by @xmath90 ( @xmath91 measurements of correlation function ) .
all individual detector proposals @xcite searching for compton frequency oscillations can substantially constrain yet unexplored parameter space .
there is no need to reproduce projected sensitivities of these proposals here ; one could simple rescale them by the factor @xmath92 . for co - located devices ( or the same apparatus ) , the `` local '' temporal correlation function is@xmath93 notice the presence of the coherence length in the combination @xmath94 ; it arise due to our motion through dm halo over the lag time @xmath95 thereby sampling dm fields @xmath96 distance apart . for @xmath97
this translates into the @xmath98 distance . considering that @xmath99 .
the signal primary oscillation frequency also depends on @xmath100 since the earth velocity changes seasonally , annual velocity modulations are imprinted in the correlation function . in practice
, one would obtain a time series of measurements @xmath101 at @xmath102 , compute auto - correlation function @xmath103 and see if it fits eq .
( [ eq : localcorrelation ] ) .
another approach is to work in the frequency space and examine the power spectral density of @xmath104 given by the fourier transform of @xmath105 . to facilitate a comparison with @xmath105 .
we can define the _ dm - induced line - shape _ as a fourier transform of the correlation function @xmath106 , where @xmath107 is normalized as @xmath108 . with @xmath109 and @xmath110
the resulting `` dm line - shape '' reads @xmath111 this holds for detunings @xmath112 , otherwise @xmath113 .
the profile is shown in fig.[fig : dmprofile ] .
as expected the line - width is @xmath114 .
the profile is strongly asymmetric due to parabolic dispersion relation for massive non - relativistic bosons . for @xmath115 ,
the maximum @xmath116 value of @xmath117 is attained at @xmath118 and the width at half - maximum is @xmath119 .
it is broad in the spectroscopic sense as @xmath120 . .
dashed vertical line marks the position of the doppler - shifted compton frequency @xmath121 and the line - width at half maximum in the frequency space is @xmath122 .
[ fig : dmprofile ] ] now we examine a practical example of a network : a network of atomic clocks . as an illustration we focus on two nodes each hosting identical devices in the configuration of fig .
[ fig : mapsatsearth](b ) , i.e. , orientation with respect to galactic velocity @xmath0 is static yet we can vary the distance @xmath123 between the nodes . extension to rotating reference frame for terrestrial networks is straightforward . in our preceding discussion
we assumed that the measurements were instantaneous ; in practice there is always a finite interrogation time @xmath124 for a single measurement .
we assume that the next measurement is taken right after the previous one was completed ( no `` dead '' times ) and that the measurements at different nodes are synchronized .
we form a time series of fractional frequency excursions @xmath125 taken at @xmath126 for a fixed distance @xmath123 , with @xmath127 labeling the node and @xmath128 being the nominal clock frequency ( strictly speaking one needs to form frequency ratios of two co - located clocks @xcite ) . the vulf contribution to @xmath129
can be expressed in terms of sensitivity coefficients @xmath130 and single - measurement accumulated clock phase @xmath131 notice the integral of the vulf field time evolution history over the interrogation duration .
the two - node correlator @xmath132 reduces to a double integral over the field correlation function @xmath133 : @xmath134 , with @xmath135 . while @xmath136 can be numerically evaluated with eq .
( [ eq : fieldcorrfuncanalyl ] ) , we demonstrate the essential features in the limit of infinite coherence lengths and times ( @xmath137 , i.e. with @xmath138 , eq . ( [ eq : gcoherent ] ) : @xmath139 the dependence on the distance @xmath123 is transparent .
the `` window '' function @xmath140 emphasizes the dependence on the ratio of interrogation time to the period of vulf oscillation .
if vulf oscillations are slow compared to @xmath124 , then @xmath141 and if they are fast , the effect tends to average out , @xmath142 . for a typical @xcite @xmath143
, the separation between the two regimes occurs for @xmath144 ( @xmath145 ) .
another consideration is synchronization accuracy : for fast oscillations , the uncertainly comes from @xmath146 in @xmath147 : if @xmath124 is uncertain @xmath148 , the @xmath147-factor in eq .
( [ eq : g - network ] ) has uncontrolled oscillatory behavior invalidating the entire analysis .
ref .
@xcite has demonstrated open - air synchronization of optical clocks at the level of fs .
thereby we expect that a network should remain sensitive to @xmath149 ; at such frequencies the smallness of coherence time on the interrogation time - scale becomes more important .
similar considerations apply to fiber networks @xcite and gps time transfer @xcite .
so far we have discussed large - scale experiments : a satellite mission where the distance @xmath123 between the devices can be in the order of @xmath150 km and terrestrial networks where the separations can be as large as @xmath151 km . finally , it is worth pointing out that small - scale experiments can probe a complementary parameter space of large masses ( see table [ tab : vulf - params ] ) .
for example , the coherence length of sub - mev vulfs is of a meter scale , while ev vulfs have @xmath152 .
these lengths suggest table - top experiments or measurements with optical tweezers @xcite where the distance between individual atoms can be varied .
additional advantage of table - top experiments is a better synchronization as the common laser source can be used to probe both locations and the possibility of using rotating tables @xcite that would compensate for the earth rotation with respect to the galactic wave vector . _
acknowledgments _ i would like to thank a. arvanitaki , d. budker , a. geraci , r. guyer , d. kimball , m. lukin , and m. pospelov for discussions and n. rodrigue for help with the figures .
this work was supported in part by the u.s
. national science foundation . |
the short distance structure of conventional geometry can be considered experimentally confirmed up to the order of 1 tev , see e.g. @xcite . in string theory and quantum gravity certain corrections to the short distance structure and the uncertainty relations
have been suggested to appear at smaller scales ( the latest at the planck scale ) , see e.g. @xcite-@xcite and for a recent review @xcite . here
we continue a series of articles @xcite-@xcite in which are studied the quantum theoretical consequences of small corrections to the canonical commutation relations = i(_ij + _
ijkl*x*_k*x*_l + _ ijkl*p*_k*p*_l + ... ) [ xp - ndim ] including the possibility that also @xmath0 \ne 0$],@xmath1\ne 0 $ ] .
a crucial feature of this ` noncommutative geometric ' ansatz , which was first studied in @xcite , is that for appropriate matrices @xmath2 and @xmath3 , eq.[xp - ndim ] implies the existence of finite lower bounds to the determination of positions and momenta .
these bounds take the form of finite minimal uncertainties @xmath4 and @xmath5 , obeyed by all physical states . in fact , the approach covers the case of those corrections to the uncertainty relations which we mentioned above , see @xcite . a framework with a finite minimal uncertainty @xmath4 can as well be understood to describe effectively nonpointlike particles , than as describing a fuzzy space .
as discussed in @xcite-@xcite the approach , with appropriately adjusted scales , could have therefore more generally a potential for an effective description of nonpointlike particles , such as e.g. nucleons or quasi - particles in solids .
analogously , on large scales a minimal uncertainty @xmath5 may offer new possibilities to describe situations where momentum can not be precisely determined , in particular on curved space @xcite . using the path integral formulation
it has been shown in @xcite that such noncommutative background geometries can ultraviolet and infrared regularise quantum field theories in arbitrary dimensions through minimal uncertainties @xmath6 . however ,
a complete analysis of the modified short distance structure , and in particular the calculation of the states of maximal localisation , has so far only been carried out for the special case of the commutation relations
@xmath7 = i\hbar ( 1 + \beta { { \bf p}}^2)$ ] , in @xcite .
the reason is that those cases are representation theoretically much easier to handle in which either @xmath2 or @xmath3 vanish , i.e. with minimal uncertainties in either positions or in momenta only .
we now solve the more general , though still one - dimensional problem , involving both minimal uncertainties in positions and in momenta .
we define the associative heisenberg algebra @xmath8 with corrections parametrised by small constants
@xmath9 = i(1 + * x*^2 + * p*^2 ) [ xp ] or , in a notation which will prove more convenient ( @xmath10 ) = i(1 + ( q^2 - 1 ) ( + ) ) [ xpq ] where the constants @xmath11 carry units of length and momentum and are related by : 4 kl=(1+q^2 ) while the first correction term contributes for large @xmath12 , which is the definition of the infrared , the second correction term contributes for large @xmath13 , i.e. in the ultraviolet . the corresponding uncertainty relation x p ( 1 + ( ( x)^2 + * x*^2 ) + ( ( p)^2 + * p*^2 ) ) holds in all @xmath14-representations of the commutation relations and reveals these infrared and ultraviolet modifications as minimal uncertainties in positions and momenta @xcite : ( x_min)^2 = l^2 ( 1 + ( q^2 - 1 ) ( + ) ) ( p_min)^2 = k^2 ( 1 + ( q^2 - 1 ) ( + ) ) in particular , for all physical states i.e. for all @xmath15 with @xmath16 being any @xmath14-representation of the commutation relations of @xmath8 in a hilbert space @xmath17 , there are finite absolutly smallest uncertainties : ( x_v ) = ( * x*-*x*)^2 ^ 1/2 l vd [ nmu - x ] ( p_v ) = ( * p*-*p*)^2 ^ 1/2 k vd [ nmu - p ] we will here only deal with the kinematical consequences of possible corrections to the commutation relations .
arbitrary systems can be considered and studies on dynamical systems , including the calculation of the spectra of hamiltonians and integral kernels such as green functions have been carried out for example systems in @xcite .
compare also with the features of the discretised quantum mechanics studied e.g. in @xcite .
a very interesting canonical field theoretical approach with a similar motivation is focusing on generalising the uncertainty relations among the coordinates @xcite .
a crucial consequence of eqs.[nmu - x],[nmu - p ] is that there are no eigenvectors to @xmath18 nor to @xmath19 in any space of physical states i.e. in any @xmath14-representation @xmath20 of the generalised commutation relations .
as is clear from the definition of uncertainties , e.g. @xmath21 , eigenvectors to @xmath18 or @xmath19 could only have vanishing uncertainty in position or momentum .
in particular , the commutation relations of @xmath22 no longer find spectral representations of @xmath18 nor of @xmath19 .
in the situation of @xmath23 ( or @xmath24 ) , i.e. with @xmath25 ( or @xmath26 ) there is still the momentum ( or position ) representation of @xmath8 available , in which case the maximal localisation states have been calculated in @xcite .
let us now perform the analogous studies for the general case with @xmath27 , where position and momentum space representations are both ruled out . to this end
we use a hilbert space representation of @xmath8 on a generalised fock space .
the position and momentum operators can be represented as * x*= l(a^+ a ) * p*= ik(a^- a ) where the @xmath28 and @xmath29 obey generalised commutation relations a a^- q^2 a^a = 1 and act on the domain @xmath20 of physical states @xmath30 as : a 0 & = & 0 + a^ & = & +1 + a & = & -1 where @xmath31 $ ] denotes the partial geometric sum or ` @xmath32'- number = and where the @xmath33!)^{-1/2 } ( a^{\dagger})^n\vert 0\rangle , { { \mbox { } } } n=1, ... ,\infty$ ] are orthonormalised = _ n_1,n_2 and @xmath20 is analytic and dense in the hilbert space @xmath34 .
while @xmath18 and @xmath19 ordinarily are essentially self - adjoint , they are now merely symmetric , which is sufficient to insure that all expectation values are real .
the deficiency indices of @xmath18 and @xmath19 are @xmath35 , implying the existence of 1-parameter families of self - adjoint extensions .
while ordinarily self - adjoint extensions , e.g. for a particle in a box , need to and can be fixed , there is now the subtle effect of the self - adjoint extensions not being on common domains , which prevents the diagonalisation of @xmath18 or @xmath19 on physical states , as can also be understood through the uncertainty relations . for the full functional analytical details
see @xcite , where these structures have first been found .
we will come back to these functional analytical studies in sec.[inv ] where we will explicitly calculate the diagonalisations in @xmath17 .
they are of use for the calculation of inverses of @xmath18 and @xmath19 , which are not only needed to describe certain quantum mechanical potentials , but ultimately also to invert kinetic terms e.g. of the form @xmath36 to obtain propagators from the field theoretical path integral , see @xcite .
the absence of eigenvectors of @xmath18 or @xmath19 in all @xmath14-representations @xmath20 of the commutation relations physically implies the absence of absolute localisability in position or momentum i.e. there are no physical states that would have @xmath37 or @xmath38 .
more precisely , the uncertainty relation , holding in all @xmath20 , implies a ` minimal uncertainty gap ' : v d : x_v < x_0 v d : p_v < p_0 .
[ gap ] the state of maximal localisation in position @xmath39 with given position expectation @xmath40 and vanishing momentum expectation , is defined through [ maxlocdef ] = x , = 0 , ( x)__x = x_min . explicitly the minimal uncertainty in position then reads [ deltax ] ( x)__x^2 = l^2 ( 1 + ( q^2 - 1 ) ) with the corresponding ( now not infinite ) uncertainty in momentum : [ deltap ] ( p)__x^2 = k^2 ( 1 + ( q^2 - 1 ) ) .
we focus on maximal localisation in @xmath40 , the case of maximal localisation in @xmath41 is fully analogous .
as shown in @xcite a state of maximal localisation is determined by the equation ( ( * x*-*x * ) + i ( * p*-*p * ) ) _ x = 0 where @xmath42 . inserting eqs.[deltax],[deltap ]
we obtain = so that the condition reads : [ maxloceq ] ( ( * x*-*x * ) + i * p * ) _ x = 0 . in order to explicitly calculate those states that realise the now maximally possible localisation
we expand the @xmath39 in the fock basis [ bfrep ] _ x : = _
n=0^q^-3n/2 c_n(x ) where the @xmath43 are real coefficients and [ normdefinition ] ( x ) : = _
n=0^ q^-3n c_n^2(x ) is a normalisation factor ( the inserted factors @xmath44 will be convenient later).the condition for maximal localisation eq.[maxloceq ] reads in the fock representation : [ maxloceqbf ] ( ( q^2 + 1)(a^+a- ) - ( q^2 - 1)(a^-a ) ) _ x = 0 . inserting the ansatz eq.[bfrep ] we are led to the recursion relation x c_n(x ) = c_n+1(x ) + c_n-1(x ) .
[ reca ] together with c_-1(x ) = 0 c_0(x ) = 1 the coefficients @xmath43 are therefore determined as polynomials of degree @xmath45 in @xmath40 .
the coefficients @xmath43 are related to the so - called continuous @xmath32-hermite polynomials .
an excellent review on these and other @xmath32-orthogonal polynomials is @xcite .
we use the notation of shifted @xmath32-factorials @xcite [ qfactorials ] ( a;q^2)_n : = _ k=0^n-1 ( 1-a q^2k ) which obey the identity [ qinverse ] ( a;q^2)_n = ( -a)^n q^n(n-1 ) ( a^-1;q^-2)_n .
furthermore we define for later convenience [ xjfunction ] ( x ) : = , x ( ) = where [ defomega ] : = ( q+q^-1 ) .
the continuous @xmath32-hermite polynomials @xmath46 are defined through h_-1(z|q^2)=0 , h_0(z|q^2)=1 and the recurrence relation , see ref.@xcite : 2 z h_n(z|q^2 ) = h_n+1(z|q^2 ) + ( 1-q^2n ) h_n-1(z|q^2 ) .
it is not difficult to check that this recursion relation can be brought into the form of the recursion relation eq.[reca ] for the coefficients @xmath43 , by expressing them in terms of the @xmath47 as : [ correspondence ] c_n(x ) = i^-n h_n(i x polynomials @xmath46 can be written as [ binominalexpression ] h_n(z|q^2 ) = _ k=0^n ( c n + k ) _
q^2 e^i(n-2k ) , z = with the @xmath32-binomial coefficients : ( c n + k ) _ q^2= . inserting eq.[binominalexpression ] into eq.[correspondence ] and replacing @xmath31!$ ] by !
= = yields c_n(x ) = i^-n _ k=0^n ( c n + k )
_ q^2 e^i(n-2k ) , i x=. because of @xmath48 we may also write @xmath49 and therefore obtain the following exact expression for the coefficients @xmath43 : [ cexplicit ] c_n(x ) = _ k=0^n ( c n + k ) _ q^2 ( -)^k q^(n-2k)(x ) .
[ a3 ] we derive further useful properties of the @xmath43 . * * classical limit + for @xmath50 the recursion relation eq.[reca ] reduces to c_n(x ) = c_n+1(x ) + c_n-1(x ) . by substituting @xmath51 and @xmath52
we obtain @xmath53 and 2zh_n(z ) = h_n+1(z ) + 2nh_n-1(z ) which is the defining recursion relation for classical hermite polynomials @xmath54 .
thus the classical limit of the polynomials @xmath43 is given by _
q 1 c_n(x ) = h_n ( ) . * representation by the formula of rodriguez * + as a short notation we write @xmath55 as @xmath56 . then , introducing the @xmath32-difference operator df(x _ ) = [ r1 ] the polynomials @xmath57 can be expressed as [ rodriguez ] c_n(x _ ) = q^^2 d^n q^-^2 where _
n = ( ) ^n .
eq.[rodriguez ] generalises the formula of rodriguez @xmath58 for classical hermite polynomials .
a proof for eq.[rodriguez ] is outlined in appendix a. + * @xmath32-difference equation * + the generalised formula of rodriguez eq.[rodriguez ] implies that [ differenceequation ] c_n(x_+1)-c_n(x_-1 ) = ( q^+q^- ) c_n-1(x _ ) which generalises the differentiation rule @xmath59 for classical hermite polynomials . in order to prove this equation , we rewrite its l.h.s . using eqs.[r1],[rodriguez ]
c_n(x_+1)-c_n(x_-1 ) = ( x_+1-x_-1 ) d c_n(x _ ) = ( x_+1-x_-1 ) d q^^2 d^n q^-^2 . carrying out the first differentiation on the r.h.s . of this formula ( c.f .
eq.[productrule ] ) , one obtains a linear combination of the polynomials @xmath60 and @xmath61 which in turn can be expressed through the recurrence relation eq.[reca ] in terms of @xmath62 .
it can also be shown by induction that eq.[differenceequation ] implies the following @xmath32-difference equation for the polynomials @xmath43 [ qdiff ] q^c_n(x_-1 ) + q^- c_n(x_+1 ) = q^-n(q^+q^-)c_n(x _ ) which corresponds to the differential equation @xmath63 for classical hermite polynomials .
+ * orthogonality * + the polynomials @xmath43 obey the orthogonality relation [ orthorel ] _ = -^ ( x_2++1-x_2 + -1 ) q^-(2+)^2 c_m(x_2 + ) c_n(x_2 + ) = n _ q^n _ m , n where n_= _ = -^ ( x_2++1-x_2 + -1 ) q^-(2+)^2 . the parameter @xmath64 can be choosen arbitrarily and fixes a family of positions occurring in the sum .
+ eq.[orthorel ] can be proved as follows .
the case @xmath65 is trivial . for @xmath66 and
@xmath67 one can show that the l.h.s . of eq.[orthorel ]
is equal to _ = -^ ( x_2++1-x_2 + -1)d^m q^-(2+)^2 = .
d^m-1 q^-j^2 |_j=-^j=+ = 0 . keeping @xmath68 fixed , a further induction for @xmath69 completes the proof . + * generating function * + a generating function of the polynomials @xmath43 is ( tq^-(x ) ; q^-2)_(-tq^(x);q^-2)_= _
in order to verify this expression , we use the @xmath32-difference equation eq.[qdiff ] and get & & q^j(tq^-+1 ; q^-2 ) _ ( -tq^-1;q^-2)_+ q^-j(tq^--1 ; q^-2 ) _ ( -tq^+1;q^-2 ) _ + & & = ( q^j+q^-j ) ( tq^--1 ; q^-2 ) _ ( -tq^-1;q^-2)_which in turn can be proved by inserting the definition of the @xmath32-factorials ( c.f .
eq.[qfactorials ] ) .
generally , all information on position and momentum is contained in the matrix elements of the position and momentum operators , and matrix elements can of course be calculated in arbitrary bases , such as also the fock basis .
ordinarily , the position and the momentum information content of a state @xmath70 of the particle is easily obtained by writing the state as a position or momentum space wave function @xmath71 or @xmath72 , which is to project onto position or momentum eigenstates , i.e. to project onto states of maximal localisation in @xmath40 or @xmath41 . in the new
setting we can now project arbitrary states @xmath70 onto the states which realise the maximally possible localisation in position ( or in momentum ) , which are given by eqs.[maxlocdef],[bfrep],[a3 ] .
we call the collection of these projections the quasi - position wavefunction @xmath73 of @xmath70 : ( x ) : = ^ml_xhere @xmath73 yields the probability amplitude for finding the particle in a state of maximal localisation around the position @xmath40 with vanishing momentum expectation . as is easily seen from eqs.[maxlocdef ] ,
[ maxloceq],[maxloceqbf ] the generalisation to arbitrary momentum expectations is straightforward .
the framework for quasi - momentum wave functions ( p ) : = ^ml_pis analogous with @xmath74 being the probability amplitude for finding the particle in a state of maximal localisation in its momentum , with the momentum expectation @xmath41 and vanishing position expectation ( again the definition may easily be generalised to include arbitrary position expectations ) .
aiming at the calculation of examples of quasi - wave functions , we need to complete our studies on the maximal localisation states by calculating their norm and scalar product . to this
end an important technical tool will be the christoffel darboux theorem , for the application of which we will need the limiting cases of the coefficients @xmath43 of the maximal localisation states . as we prove in appendix b , the polynomials @xmath43 , for all odd and for all even @xmath45 have the nontrivial property that their limit for @xmath77 exists . denoting again @xmath78
these limits are : [ cplus ] c^+(x _ ) & = & _ m ( -)^m c_2m(x _ ) = aq^^2/2_2 ( , ) + [ cminus ] c^-(x _ ) & = & _ m ( -)^m c_2m+1(x _ ) = aq^^2/2_1 ( , ) where @xmath79 are the jacobi- , or elliptic @xmath80-functions defined as [ theta1 ] _ 1(z , ) & : = & 2 ^1/4 _ n=0^(-)^n ^n(n+1 ) ( ( 2n+1)z ) + _ 2(z , ) & : = & 2 ^1/4 [ theta2 ] _
n=0^^n(n+1 ) ( ( 2n+1)z ) and where in eqs.[cplus],[cminus ] the constants @xmath81 and @xmath22 are defined as [ constlambda ] : = e^ and [ consta ] a^2 : = = . in order to evaluate the scalar product of two maximally localised states = _
n=0^ q^-3nc_n(x)c_n(x^ ) the @xmath32-difference equation eq.[qdiff ] can be used to rewrite this expression as [ quasiscalarproductdef ] = where we defined f _ , : = _
n=0^ q^-nc_n(x)c_n(x^ ) and where we abbreviated @xmath82 and @xmath83 .
we can compute @xmath84 by applying the christoffel - darboux @xcite theorem [ christoffel ] _
n=0^m q^-nc_n(x)c_n(x^ ) = which can be proved as follows . for @xmath67 (
the case @xmath85 is trivial ) we use the recursion relation ( c.f .
. [ reca ] ) c_m+1(x ) = ( xc_m(x ) - c_m-1(x ) ) in order to replace @xmath86 and @xmath87 on the r.h.s . of eq.[christoffel ] which then takes the form = q^-mc_m(x)c_m(x^ ) - so that eq.[christoffel ] follows by induction .
since the polynomials @xmath88 have well defined limits as @xmath68 goes to infinity , the christoffel - darboux theorem implies that the expression @xmath89 is given by [ normalscalarproduct ] f _ , = _
n=0^ q^-nc_n(x)c_n(x^ ) = . inserting eqs.[cplus],[cminus ] yields f _ , = _
1((-),^2 ) _
4((+),^2 ) [ a60 ] with the definition of @xmath90 being _
4(z,^2 ) : = 1 + 2_n=1^ ( -)^n^2n^2(2nz ) . in the limit @xmath91 eq.[a60 ]
reduces to [ normalnorm ] f _ , = _
n=0^ q^-nc_n^2(x ) = _ 4(,^2 ) .
inserting eq.[normalscalarproduct ] into eq.[quasiscalarproductdef ] we eventually obtain an exact expression for the scalar product of two quasi - position states : [ quasiscalarproduct ] = _ 1((-),^2 ) _
4((+),^2 ) where g^s _ , = ( q^12(-)+s - q^-12(-)-s ) ( q^12(+)+s+q^-12(+)-s ) . note that the poles in the denominator of eq.[quasiscalarproduct ] are cancelled by the zeros of the @xmath92-function .
the limit @xmath91 yields the norm ( eq.[normdefinition ] ) [ quasinorm ] ( x ) = .
as an example we draw the graph of the quasi - position wave function @xmath73 for the state @xmath70 that describes maximal localisation around @xmath94 i.e. for @xmath95 , i.e. with ( x ) = 3truecm _ fig . 1 : quasi - position wave function @xmath73 for @xmath96 for @xmath97 drawn over @xmath98 _ 0.8truecm
the analytic form of the wave function is given in eq.[quasiscalarproduct ] .
for the width of the main peak note that the graph shows the overlap of pairs of localisation states , each with its finite position uncertainty .
we have thus generalised the treatment of @xcite where the corresponding graph was calculated and drawn for the special case without a minimal momentum uncertainty ( @xmath99 ) .
for potential applications of the formalism the parameters @xmath2 and @xmath3 in eq.[xp ] can be assumed small in which case useful simplifications hold . in the notation of eq.[xpq ] this is the case when @xmath50 .
then @xmath100 and the @xmath80-functions in eqs.[cplus],[cminus ] can be approximated by @xmath101 and @xmath102 .
this implies that [ capprox ] c^+(x _ ) & & ^+(x _ ) : = b q^^2/2 + c^-(x _ ) & & ^-(x _ ) : = b q^^2/2 where [ bdef ] b^4 = .
the relative error of this approximation is shown in the following table : @xmath103 & @xmath104 & @xmath105 & @xmath106 & @xmath107 using eq.[capprox ] we can give approximations of the scalar product for @xmath32 close to @xmath108 : [ approxscalarproduct ] ( - ) where [ ncnorm ] ( x ) ( x ) : = .
e.g. for @xmath109 the relative error of this approximation is less than @xmath110 .
it is interesting to consider also the limiting case where q 1 , k(q ) : = , l(q ) = .
[ limits ] in this limit the commutation relations eq.[xpq ] turn into the relations eq.[xp ] with @xmath3 finite but @xmath23 , which is the special case considered in @xcite .
there is then only a minimal uncertainty in position and no minimal uncertainty in momentum . as can be shown easily , the limit @xmath50 of the scalar product eqs.[approxscalarproduct]-[ncnorm ] is given by _
q 1 = . in the limit given by eqs.[limits ] ,
@xmath40 and @xmath111 are related linearly through @xmath112 .
we thus obtain the limiting expression for the scalar product : ^ml_x^ ^ml_x = ( - ( ) ^3)^-1 ( ) this result coincides with the expression found in @xcite , thus providing a nontrivial consistency check : we calculated the scalar product using @xmath32-analysis on a discrete @xmath32- fock space representation .
however , the calculation @xcite of this scalar product in the special case @xmath113 , which we here recover in the limit , had been performed with entirely different analytic methods in a continuous representation .
in this section we continue formal considerations of @xcite where it was proved that the operators @xmath18 and @xmath19 separately do have one - parameter families of self - adjoint extensions in @xmath17 . to be precise ,
@xmath18 on @xmath20 is symmetric , while its adjoint @xmath114 is closed but nonsymmetric .
@xmath115 is closed and symmetric and has deficiency indices ( 1,1 ) .
there are families of diagonalisations of @xmath18 in @xmath17 , though of course not in @xmath20 .
the same holds for @xmath19 .
the corresponding eigenvectors are unphysical states , separated from the physical domain by the minimal uncertainty gap , see eq.[gap ] . while in @xcite
the existence only of self - adjoint extensions had been proven , we can now explicitly solve the eigenvalue problem * x*.v_x = x v_x for the solution we make the ansatz v_x = n^-1(x ) _
n=0^ q^-n/2 d_n(x) yielding the recurrence relation d_n(x ) = d_n+1(x ) + d_n-1(x ) [ e80 ] with @xmath116 and @xmath117 .
in fact , eq.[e80 ] can be transformed into the recurrence relation eq.[reca ] , i.e. the @xmath118 can be transformed into our previously considered coefficients @xmath119 : d_n(x ) = c_n(2 x ( q+q^-1)^-1 ) . in the expansion of @xmath120 ,
the factor @xmath121 is different from the corresponding factor @xmath44 in the expansion of the @xmath122 , implying that the scalar product and normalisation constant of the formal eigenvectors are different from those of the maximal localisation states which we had calculated earlier : n(x ) = = = where now x(j ) : = | = and where we abbreviated again @xmath83.we draw the graph of the scalar product over @xmath111 for @xmath123 : 5truecm _ fig .
2 : scalar product @xmath124 of formal eigenvectors drawn over @xmath98 _ 0.8 cm from the zero s of @xmath92 we read off that the @xmath120 are mutually othogonal for @xmath125 . using @xmath126 as a parameter in the range @xmath127 we identify for each value of @xmath126 a diagonalisation of @xmath18 .
thus , @xmath128 labels the self - adjoint extensions with the corresponding eigenvalues @xmath129 being x_n
= l = l ( n ) [ a61 ] compare also with the graph of the scalar product which had been calculated only numerically in @xcite .
having found the analytic form of the scalar product in terms of @xmath80 functions we were able to determine the one parameter familiy of diagonalisations of @xmath18 , of which we had so far only known its existence . as is not difficult to see
we recover for vanishing minimal uncertainty in momentum , i.e. for @xmath130 , the linear spectrum found in @xcite for that special case .
analogously to above we obtain the eigenvalues of @xmath19 in its self - adjoint extensions ( @xmath131 ) : p_n = k = k ( n ) [ a62 ] we stress that the parameters @xmath132 of eqs.[a61],[a62 ] label _ different extensions of the domain @xmath20 of @xmath18 and @xmath19 .
recall that the uncertainty relation implies that the formal @xmath18- or @xmath19- eigenvectors which we here calculated do not lie in any _ common extension of the domain @xmath20 of @xmath18 and @xmath19 .
they are not physical states and are separated from the physical domain by the uncertainty gap , see eq.[gap ]
. however , these families of diagonalisations of @xmath18 or @xmath19 in @xmath17- can still be of use , e.g. for the calculation of inverses of @xmath18 and @xmath19 , which would have been difficult to invert as nondiagonal operators in the fock basis @xmath133 and @xmath134 . _ _
in quantum field theory , interaction terms which on ordinary geometry would be ultraviolet regular but nonlocal , can in fact be regular and strictly local on a geometry with a minimal position uncertainty .
the reason is that an interaction is to be considered strictly local if no nonlocality could be observed .
intuitivly this is the case if a small apparent nonlocality of the interaction term is unobservable due to a comparatively larger minimal uncertainty in the underlying space .
we already mentioned that , as has been shown in @xcite , quantum field theories can be naturally regularised when working on a generalised geometry with intrinsic minimal uncertainties . generally , in order to explicitly compare the size of nonlocality of an arbitrary interaction term with the size of the intrinsic uncertainty of the generalised geometry
it is crucial to have available the states of maximal localisation on this geometry .
similarly , maximal localisation in a momentum space with minimal uncertainty @xmath5 is of interest in the context of infrared regularisation . so far
, we have studied the properties of the maximal localisation states in one dimension only .
the generalised fourier transformations which map between the ( quasi- ) position and the ( quasi- ) momentum representations have only been studied in the special case @xmath113 . for
the general case techniques should be useful which have been developed for the fourier theory @xcite on quantum planes @xcite .
also , the unitary equivalence of _ all hilbert space representations , in the sense in which it holds for the ordinary commutation relations , has not yet been proven .
most interesting further physical insight into the nature of these generalised geometries can be expected from studies on maximal localisation in @xmath45 dimensions where @xmath0\ne 0 $ ] and @xmath1\ne 0 $ ] lead also to @xmath135 and @xmath136 .
work in this direction is in progress .
* acknowledgement * _
a.k . would like to thank prof .
t koornwinder for the kind invitation to give a seminar in amsterdam and a very useful discussion on @xmath32- hermite polynomials .
in the following we outline a proof for the equivalence of the formula of rodriguez eq.[rodriguez ] and the recurrence relation eq.[reca ] .
we use the notation @xmath137 which allows the differentiation of products to be written in the form : [ productrule ] d(fg ) = ( df)(ag ) + ( af)(dg ) .
we first prove the identity [ appeq1 ] d^n+1 q^-j^2 = - ( ) ^2 ( q^-2n+1 [ n]d^n-1 + xq^-n d^n ) q^-j^2 . the case @xmath66 can be verified easily by hand .
induction from @xmath45 to @xmath138 implies that [ appeq2 ] d^n+1 q^-j^2 = d(d^n q^-j^2 ) = -d ( ) ^2 ( q^-2n+3 [ n-1]d^n-2 + xq^-n+1 d^n-1 ) q^-j^2 comparing the r.h.s . of eqs.[appeq1 ] and
[ appeq2 ] and using eq.[productrule ] , one is led to the condition [ appeq3 ] ( ad^n+1 + 12(q - q^-1)xd^n - q^-nd^n-1 ) q^-j^2 = 0 which can be proved by a further induction where the identities @xmath139 turn out to be very useful .
once eq.[appeq1 ] is proved , one obtains the recursion relation eq.[reca ] by inserting the formula of rodriguez which completes the proof .
we prove the limits @xmath140 i.e. we prove eqs.[cplus],[consta ] : [ lemma1 ] c^+(x _ ) = _ m ( -)^m c_2m(x _ ) = q^^2/2_2 ( , ) let us first rewrite expression eq.[cexplicit ] by c_2 m ( x _ ) = _ k =- m^m ( 2 m m+k ) _ q^2 ( -)^m+k q^2k .
choosing an integer @xmath141 we split up this sum into two parts ( -)^m c_2m(x _ ) = s_m , r^(1 ) + s_m , r^(2 ) where s_m , r^(1 ) & = & _ k =- r^r ( 2 m m+k ) _
q^2 ( -)^k q^2k + s_m , r^(2 ) & = & _ k = r+1^m ( 2 m m+k ) _ q^2 ( -)^k ( q^2k+q^-2k ) now let @xmath68 go to infinity and keep @xmath142 fixed . since @xmath143 and eq.[qinverse ] we have the identity [ binomidentity ] ( 2 m m+k ) _ q^2 = q^2 ( m^2-k^2 ) ( 2 m m+k ) _ q^-2 .
because of _ m ( 2 m m+k ) _
q^-2 = the first part converges to s_r^(1 ) = _ m s_m , r^(1 ) = _ k =- r^r(-)^kq^2k-2k^2 .
the second part @xmath144 can be estimated as follows .
as can be seen from eq.[binomidentity ] the inequality ( 2 m m+k ) _
q^-2 = = implies that ( 2 m m+k ) _ q^2 . therefore |s_m , r^(2)| _ k = r+1^m ( 2 m m+k ) _
q^2k _ k = r+1^m q^2(k - k^2 ) so that [ estimation ] |s_r^(2)| = |_m s_m , r^(2)| _ k=1^ q^2(k - k^2 ) . since the sum on the r.h.s . is finite this expression tends to zero as @xmath142 goes to infinity .
thus we conclude that [ sumexpression ] c^+(x _ ) = _ r s_r^(1 ) = _ k=-^(-)^kq^2k-2k^2 .
the sum on the r.h.s . of this expression is essentially a jacobi @xmath145-function . in order to see this notice that its definition eq.[theta2 ] can also be written as _
2(z , e^- ) = _ k=0^ ( -)^k e^-(z - k)^2 . inserting @xmath146 and @xmath147
we can express the sum in eq.[sumexpression ] by _
k=-^ ( -)^k q^-2k^2 + 2k = q^^2/2 _ 2(z , ) which completes the proof for @xmath148 .
the proof for @xmath149 follows the same lines .
khuri , hep - ph/9405406 , preprint ru-94 - 5-b ( 1994 ) .
p. k. townsend , phys . rev . *
d15 * , 2795 ( 1976 ) .
jaeckel , s. reynaud , phys . lett . *
a185 * , 143 ( 1994 ) d. amati , m. cialfaloni , g. veneziano , phys . lett . *
b216 * , 41 ( 1989 ) .
balachandran , g. bimonte , e. ercolessi , g. landi , f. lizzi , g. sparano , p. teotonio - sobrinho , preprint ictp : ic/94/38 , dsf - t-2/94 , su-4240 - 567 ( 1994 ) .
k. konishi , g. paffuti , p. provero , phys . lett . *
b234 * , 276 ( 1990 ) .
m. maggiore , phys . lett . *
b319 * , 83 ( 1993 ) ; m.maggiore , phys.rev . * d49 * , 5182 ( 1994 ) . l. j. garay , int . j. mod
. phys . * a10 * 145 - 166 ( 1995 ) .
a. kempf , lett .
* 26 * , 1 ( 1992 ) .
a. kempf , j. math .
phys.*34 * , 969 ( 1993 ) .
a. kempf , proc .
xxii dgm conf . ,
cliff . alg .
( proc . suppl . )
( s1 ) 87 ( 1994 ) a. kempf , j. math . phys . * 35 * ( 9 ) , 4483 ( 1994 ) .
a. kempf , preprint damtp/94 - 33 , hep - th/9405067 ( 1994 ) .
a. kempf , czech .
( proc . suppl . ) * 44 * , 1041 ( 1994 ) a. kempf , g. mangano , r.b .
mann , phys.rev.*d52 1108 ( 1995 ) j. schwenck , j. wess , phys .
* b291 * , 273 ( 1992 ) .
a. hebecker , s. schreckenberg , j. schwenk , w. weich , j. wess , lett . math
* 31 * , 279 ( 1994 ) .
koornwinder , r.f .
swarttouw , trans .
ams * 333 * , 445 ( 1992 ) .
s. doplicher , k. fredenhagen , j. e. roberts , phys .
lett.*b331 * , 39 ( 1994 ) .
r. koekoek , r. f. swarttouw , report 94 - 05 of fac .
& inf . , techn univ .
delft ( 1994 ) g. szeg , _ orthogonal polynomials _ ,
23 , providence , ri ( 1939 ) a. kempf , s. majid , j. math . phys .
* 35 * ( 12 ) , 6802 ( 1994 ) l. d. faddeev , n. yu .
reshetikhin , l. a. takhtajan , alg .
anal . * 1 * , 178 ( 1989 ) . | small corrections to the uncertainty relations , with effects in the ultraviolet and/or infrared , have been discussed in the context of string theory and quantum gravity .
such corrections lead to small but finite minimal uncertainties in position and/or momentum measurements .
it has been shown that these effects could indeed provide natural cutoffs in quantum field theory .
the corresponding underlying quantum theoretical framework includes small ` noncommutative geometric ' corrections to the canonical commutation relations . in order to study the full implications on the concept of locality
it is crucial to find the physical states of then maximal localisation .
these states and their properties have been calculated for the case with minimal uncertainties in positions only . here
we extend this treatment , though still in one dimension , to the general situation with minimal uncertainties both in positions and in momenta .
8.5 in by -1.1truecm by -1.0truecm v#1#1 # 1#2#1#2 # 1#1 # 1#2#3#1 # 2 # 3 # 1#1 # 1_|^#1 # 1_^#1 # 1 ^ 2_|^#1 -18truecm 11truecm * damtp/95 - 50 * 18truecm |
in nuclear systems , an important feature is the independent single particle motion in a self - consistent mean field . as a result of spin - orbit splitting coming from the one - body spin - orbit potential ,
the @xmath7-coupling shell model has been able to explain the nature of various nuclei .
another important feature is the development of clusters .
usually , well - developed cluster structures appear in excited states close to their cluster decay thresholds , as predicted by the ikeda s threshold rule @xcite . in light nuclei
, it is known that cluster formation occurs in low - lying states and competes with the shell - model structure .
@xmath1c is a typical example of this competition @xcite . because the spin - orbit force usually tends to hinder the cluster formation ,
the ground state of @xmath1c is not the pure 3@xmath0 cluster state but is a mixture of 3@xmath0 cluster structure and @xmath7-coupling shell model structure of the @xmath2 subshell closure ; this is supported by the large level spacing between @xmath8 and @xmath9 states . on the other hand , in excited states of @xmath1c near the threshold energy , developed 3@xmath0 cluster structures have been discovered .
it is important to understand the magnitude of this competition and how much it affects the structures of the ground and excited states of @xmath1c . to clarify this problem ,
it is necessary to have a theoretical model that can describe cluster structures incorporating cluster breaking effects .
nevertheless , many studies have investigated the cluster structures of @xmath1c using 3@xmath0 cluster models without considering cluster breaking . in the following , we review studies of structures of @xmath1c on the basis of various theoretical approaches and introduce recent developments in experimental works . in 3@xmath0 cluster models ,
the ground and low - lying states have a compact triangle structure , whereas excited states near and above the 3@xmath0 cluster threshold energy have well - developed cluster structures @xcite .
for example , the @xmath10 state , which is well - known as the hoyle state , is considered to be an @xmath0 condensate state in which weakly interacting three @xmath0 clusters occupy an identical lowest orbit of a mean - field potential @xcite .
this interpretation was indicated by the so - called tohsaki - horiuchi - schuck - rpke ( thsr ) wave function .
the cluster developed structure in the @xmath4 state is also experimentally supported by a quite large monopole transition strength @xcite .
another interesting cluster structure was predicted in the @xmath5 state , which is considered to be a vibration mode of acute and obtuse triangle configurations @xcite . however
, 3@xmath0 cluster models can not describe the detailed properties of ground and excited states of @xmath1c because they ignore the components of @xmath0-cluster breaking caused by the spin - orbit force .
for example , the excitation energy of the @xmath9 state is underestimated . in the no - core shell model calculation , which is one of the _ ab initio _ calculations ,
many states of @xmath1c are reproduced well @xcite .
however , the model fails to reproduce some states observed near the cluster threshold such as @xmath4 and @xmath5 states .
this seems to be reasonable because these states are cluster developed states in which large proportion of nucleons are distributed over many harmonic oscillator orbits . to describe these cluster developed states ,
a model space is required that is significantly larger than those used in the usual no - core shell model calculation @xcite .
another _ ab initio _
calculation for @xmath1c has been performed using the chiral effective field theory @xcite . however , quantitative reproduction of structure properties , such as radii , was not satisfactory even though the calculation reproduced some experimental data such as excitation energies and @xmath11 transition strengths . for detailed discussion of cluster structures , further developments are required . the antisymmetrized molecular dynamics ( amd ) and the fermionic molecular dynamics ( fmd ) models have been successful in describing the nature of @xmath1c from the ground state to the higher excited states @xcite . in these models , the ground and low - lying states are a mixture of cluster and @xmath7-coupling shell model configurations .
because of this mixing , these models reproduce well the excitation energy of the @xmath9 state and @xmath11 transition strengths between the low - lying states . for states near the threshold , these models reproduce the cluster nature , but the detailed cluster structures obtained with these models are somewhat different from those of cluster model results . for example , the radius of the @xmath4 state can be somewhat smaller than that predicted by the cluster model , although both the models reproduce the monopole transition strength @xcite .
the @xmath5 state is considered to be a chain - like structure having an obtuse triangle configuration of three @xmath0 clusters instead of the vibrational state predicted by the cluster model .
the capability of reproducing the experimental data of @xmath1c is a significant advantage of amd and fmd .
recently , several interesting experimental data on @xmath1c for excited states near the threshold energy have been reported .
a new state , the @xmath6 state at 9.84 mev , was discovered @xcite . the @xmath12 transition strength from this @xmath6 state to the ground state
was measured to be 1.57@xmath13 e@xmath14@xmath15 @xcite , which many models fail to reproduce .
_ found that the broad @xmath3 state at 10 mev consists of two @xmath3 states @xcite .
theoretically , the existence of two @xmath3 states above the @xmath4 state was predicted by kurokawa and kat @xcite and also shown by ohtsubo _
_ @xcite using the orthogonal condition model and the complex scaling method ( ocm+csm ) .
a band structure including these new states has been an open problem . in particular , the assignment of the band - head state of the newly measured @xmath6 state is now controversial . in this paper , we investigate @xmath1c with a simple model to clarify @xmath0-cluster breaking effects on 3@xmath0 cluster structures . in this model , we superpose the brink - bloch ( bb ) cluster model wave functions and the @xmath2 subshell closure wave function to incorporate the mixing of cluster breaking components in cluster structures .
this model is regarded as an extension of the generator coordinate method of a cluster model .
the difference from traditional cluster models is the mixing of the @xmath2 subshell closure wave function , which is the lowest configuration of the @xmath7-coupling shell model .
the results obtained with this model are consistent with the amd and fmd results even though it is much simpler than those models . in this model space , we changed the strength of the spin - orbit force by hand to control the mixing of the @xmath0-cluster breaking component , i.e. , the @xmath2 subshell closure wave function .
we found that the mixing of the @xmath0-cluster breaking component significantly changes the cluster configurations and band structure of excited states .
this paper is organized as follows . in sec .
[ formulation ] , we explain the wave function and hamiltonian used in the present study . in sec .
[ results ] , we show calculated results . in sec
. [ discussion ] , we discuss differences in cluster configurations in @xmath3 states and band structure between calculations with and without @xmath0-cluster breaking .
we compare our results with other theoretical calculations .
finally , we summarize the findings of the present study in sec .
[ summary ] .
in this work , we investigate @xmath1c using a simple model that combines the bb cluster model with the lowest wave function of the @xmath7-coupling shell model . we superpose many 3@xmath0 bb wave functions and the single @xmath2 subshell closure wave function .
the bb wave function @xcite for three @xmath0 clusters is written as @xmath16,\ ] ] @xmath17 @xmath18 , \label{eq : brink_single}\ ] ] where @xmath19 is a real vector that indicates the position of the @xmath20th @xmath0 cluster .
for the width parameter , we use the value @xmath21 @xmath22 same as that used in the previous study of @xmath1c using amd @xcite .
this parameter was originally from a variational calculation for the ground state of @xmath23be in ref .
the configuration of the bb wave function is characterized by the distances between @xmath0 clusters , @xmath24 and @xmath25 , and the bending angle , @xmath26 , as shown in fig .
[ fig : configuration ] . because of symmetry , we can limit distances to @xmath27 without loss of generality . to describe relative motions of clusters sufficiently ,
we generate 252 bb wave functions by taking the distance @xmath28 fm and the bending angle @xmath29 with @xmath30 .
to describe the @xmath2 subshell closure wave function , we adopted the method proposed in refs .
@xcite . in this method ,
the single particle wave function is described by a gaussian wave packet in the same manner as in the bb wave function eq . .
however , the center of the wave packet for the @xmath31th nucleon belonging to the @xmath20th @xmath0 cluster is replaced by a complex parameter @xmath32 , @xmath33 where @xmath34 is a unit vector for the intrinsic spin direction of the @xmath31th nucleon . in the limit of @xmath35 ,
this wave function describes the @xmath2 subshell closure wave function . for the width parameter of the @xmath2 subshell closure wave function
, we use the same value @xmath21 @xmath22 of the bb wave function to exactly separate the center - of - mass motion from the total wave function . as the final wave function , we superpose the 252 3@xmath0 bb wave functions and the single @xmath2 subshell closure wave function with the parity and angular momentum projection , @xmath36 where @xmath37 is the parity and angular momentum projection operator .
the coefficients are determined by the diagonalization of the norm and hamiltonian matrices .
the superposition of bb wave functions without the @xmath2 subshell closure wave function corresponds to the traditional 3@xmath0 cluster gcm calculation without @xmath0-cluster breaking . in the present model , @xmath0-cluster breaking caused by the spin - orbit force
is incorporated by adding the @xmath2 subshell closure wave function to the 3@xmath0 cluster model space .
note that the final wave functions for the non - zero angular momentum states are the same as those obtained by the 3@xmath0 cluster gcm calculation because the single @xmath2 subshell closure wave function is a @xmath38 eigenstate ; therefore , @xmath39 vanishes for @xmath40 .
in other words , the @xmath0-cluster breaking caused by the spin - orbit force affects only the @xmath3 states through the mixing of the @xmath2 subshell closure wave function . in the present work ,
we omit cluster breaking components for the @xmath40 states because they give only minor effect compared with the significant cluster breaking effects in the @xmath38 states .
indeed , the structure of the cluster developed states does not change qualitatively even if we adopt the cluster breaking components for the @xmath40 states . to simplify the discussion ,
we adopt only the @xmath2 subshell closure wave function , which gives the major contribution of the @xmath0-cluster breaking effect .
the hamiltonian consists of the kinetic energy @xmath41 , effective central force @xmath42 , effective spin - orbit force @xmath43 , and coulomb force @xmath44 , @xmath45 where the center - of - mass kinetic energy @xmath46 is subtracted . for the effective central force
, we used the volkov no . 2 force @xcite with majorana parameter @xmath47 .
for the effective spin - orbit force , we used the spin - orbit term of the g3rs interaction @xcite , @xmath48 \hat{p}(^{3 } \text{o } ) \hat{\bm{l } } \cdot \hat{\bm{s}},\ ] ] where @xmath49 is a projection operator onto the triplet odd state .
we chose @xmath50 , where @xmath51 is the strength of the spin - orbit force .
we used the same interaction with @xmath52 mev as that used in a previous study of @xmath1c using amd @xcite .
we also investigate the @xmath51 dependence of energy levels and the degree of @xmath0-cluster breaking by changing the strength @xmath51 from 0 mev to 3200 mev .
in fig . [ fig : levels_0+_2 + ] , we show the energy levels of @xmath3 and @xmath53 states calculated with the present model using @xmath52 mev , which is the same strength used in our previous study of @xmath1c using amd @xcite , and those obtained with the @xmath54-cluster gcm calculation .
the former and the latter correspond to calculations with and without @xmath0-cluster breaking , i.e. , the mixing of the @xmath2 subshell closure wave function caused by the spin - orbit force , which we call `` 3@xmath0+@xmath2 '' and `` 3@xmath0 '' , respectively .
experimental data are taken from refs .
energies are measured from the @xmath54 threshold energy .
the theoretical and experimental threshold energies are @xmath55 mev and @xmath56 mev , respectively . in the `` 3@xmath0 '' case ,
the spin - orbit force provides zero energy contribution because the expectation value of the spin - orbit force vanishes in the pure @xmath54 cluster model space . in the `` 3@xmath0+@xmath2 '' result ,
the @xmath53 states are completely the same as those of the `` 3@xmath0 '' result , but the @xmath3 energies are decreased by the spin - orbit force because of the mixing of the @xmath2 subshell closure wave function . the results for `` 3@xmath0+@xmath2 '' agree with experimental results very well except for the absence of a @xmath3 state observed around 10 mev .
in particular , the level spacing between @xmath8 and @xmath9 states of `` 3@xmath0+@xmath2 '' agrees well with the experimental one , which is largely underestimated in `` 3@xmath0 '' .
this large spacing comes from the energy gain of the spin - orbit force in the @xmath8 state with the mixing of the @xmath2 subshell closure wave function .
any microscopic @xmath0 cluster model fails to reproduce this large level spacing , and therefore , this is an evidence for @xmath0-cluster breaking . as shown later ,
the `` 3@xmath0+@xmath2 '' result is qualitatively consistent with the amd and fmd results for @xmath1c .
this implies that the major difference between the amd and fmd results and 3@xmath0 cluster model calculations is the @xmath0-cluster breaking effect , which can be qualitatively simulated by the mixing of only one configuration of the @xmath2 subshell closure to the @xmath54 cluster model space in the present model . in sec .
[ discussion ] , we discuss in detail the structure differences between @xmath3 states with and without @xmath0-cluster breaking . as mentioned above , @xmath0-cluster breaking caused by the spin - orbit force gives a significant effect in @xmath1c . to clarify the mechanism of the mixing of the @xmath0-cluster breaking component
, we consider the dependence of @xmath3 states on the strength of the spin - orbit force by changing the strength parameter @xmath51 from 0 mev to 3200 mev in the present model .
the calculation using @xmath57 mev is equivalent to the `` 3@xmath0 '' calculation . as the strength of the spin - orbit force increases , @xmath3 energy levels change due to @xmath0-cluster breaking , i.e. , the @xmath2 subshell closure wave function . in fig .
[ fig : level_ls_dependence ] , we show the change in energy levels of @xmath3 states against the strength of the spin - orbit force @xmath51 . the energy of the @xmath2 subshell closure wave function is also shown for reference . for @xmath58 mev ,
the energies are almost independent of the strength of the spin - orbit force . in this region , the energy levels and wave functions themselves are almost equal to those obtained from the pure cluster model calculation without @xmath0-cluster breaking .
when @xmath59 mev , the @xmath3 levels except for the lowest one do not depend on the strength of the spin - orbit force ; however , the energy of the lowest @xmath3 state linearly decreases and shows almost the same @xmath51 dependence as that of the @xmath2 subshell closure state .
these imply that structures of the states do not change when the strength of the spin - orbit force is extremely weak or strong .
however , around @xmath52 mev , energy levels depend on the strength @xmath51 because of level crossings of the cluster states and the @xmath2 subshell closure wave function . in this transient region , a change in the wave function occurs for each state .
.percentages of squared overlaps between @xmath3 gcm wave functions and the @xmath2 subshell closure wave function for @xmath60 mev . [ cols="^,^,^,^,^,^,^ " , ] in table [ table : comparison_e2 ] , we compare @xmath11 transition strengths from the present results with those from several theoretical models .
the labels in the table are the same as those in fig .
[ fig : comparison_levels ] .
the microscopic @xmath0-cluster models without @xmath0-cluster breaking tend to overestimate the @xmath61 transition strengths , whereas models with @xmath0-cluster breaking reproduce it properly . for the reproduction of the @xmath62 transition strength , the present 3@xmath0+@xmath2 agrees with the experimental data better than the other models .
the @xmath11 transition strengths from the @xmath6 state are important for determining the band assignment for this state , as discussed before .
although the @xmath6 state has strong @xmath11 transitions to the @xmath4 and @xmath5 states , the major transition differs between calculations with and without @xmath0-cluster breaking .
the @xmath63 transition strength is significantly larger than the @xmath64 one in the @xmath65 and amd calculations with @xmath0-cluster breaking .
however , the relation of strengths is opposite and the @xmath64 transition is dominant in the 3@xmath0 and thsr calculations without @xmath0-cluster breaking .
this implies that the band - head of the @xmath6 state is changed from the @xmath4 state to the @xmath5 state by @xmath0-cluster breaking .
this difference between calculations with and without @xmath0-cluster breaking comes from the difference of cluster configurations in the @xmath4 and @xmath5 states due to @xmath0-cluster breaking . as explained above , with @xmath0-cluster breaking , the cluster development of the @xmath4 state
is reduced to decrease the @xmath11 transition from the @xmath6 state . moreover ,
the structure of the @xmath5 state changes from the vibration mode of acute and obtuse triangle configurations of three @xmath0 clusters to a chain - like open triangle structure .
therefore , the phase with the @xmath6 becomes more coherent in the results with @xmath0-cluster breaking .
we stress that @xmath0-cluster breaking caused by the spin - orbit force has significant effects on transition strengths and changes band structure .
this implies that , to investigate structures of excited cluster states in @xmath1c , it is important to consider @xmath0-cluster breaking , in particular , mixing of the @xmath2 subshell closure configuration in @xmath3 states . in the results with @xmath0-cluster breaking , we assigned the @xmath5 and @xmath6 states as the band members because of the major @xmath11 transition as discussed above . since the results with @xmath0-cluster breaking reproduce the experimental data such as the large level spacing between @xmath8 and @xmath9 states , rms - radii , and @xmath11 transition strengths between low - lying states well , we consider the @xmath5 state is likely to be the band - head of the @xmath6 , although there is no experimental data for the @xmath11 transitions from the @xmath6 state to the @xmath4 state and the @xmath5 state . strictly speaking , the structure of the @xmath6 state is different from that of the @xmath5 state because of 16% of the @xmath0-cluster breaking component in the @xmath5 state ( see table [ table : squared_overlap_p3/2 ] ) .
it means that they are not an ideal rotational band constructed from an intrinsic state .
however , when we project the structures on the 3@xmath0 cluster configuration space , we found that both of @xmath5 and @xmath6 states have maximum amplitude around obtuse triangle region . considering these points ,
the stronger @xmath11 transition strength and maximum amplitude around the chain - like open triangle structure , this band can be considered a linear chain - like band . in this band ,
the energy position of the @xmath5 state is slightly higher than that of the @xmath6 state in the present calculation with @xmath0-cluster breaking .
the order of the @xmath3 state and the @xmath53 state is reverse to an ordinary rotational band .
this may come from the mixing of the intrinsic structure of the @xmath4 state in the @xmath6 state , though the dominant component of the @xmath6 state is the intrinsic structure of the @xmath5 state .
this can be understood from that the @xmath11 transition from the @xmath6 state is not weak to the @xmath4 state though the strength is relatively smaller than that to the @xmath5 state .
to clarify @xmath0-cluster breaking effects on 3@xmath0 cluster structures in @xmath1c , we investigated @xmath1c with a hybrid model composed of the bb cluster model and the @xmath2 subshell closure wave function .
we superposed 252 bb wave functions to sufficiently describe the relative motions of clusters .
the @xmath0-cluster breaking caused by the spin - orbit force affects @xmath3 states through the mixing of the @xmath2 subshell closure wave function .
the present result reproduces well the experimental results with the significant mixing of @xmath0-cluster breaking .
we have found that @xmath0-cluster breaking significantly changes the cluster structures of @xmath3 states through orthogonality to lower states . as a result of the structure changes of @xmath3 states , the band assignment for the @xmath6 state is changed . to investigate @xmath0-cluster breaking effects , we compared 3@xmath0 configurations contained in @xmath3 states obtained from calculations with and without @xmath0-cluster breaking . in calculations with and without @xmath0-cluster breaking ,
the @xmath8 state has a compact @xmath54 cluster structure as the dominant component and also contains components of developed cluster configurations .
however , in the calculation with @xmath0-cluster breaking , 3@xmath0 cluster components becomes relatively small because of significant mixing of the @xmath2 subshell closure component . in both the calculations ,
the @xmath4 state is described by the superposition of various triangle configurations of three @xmath0 clusters . however , in the calculation with @xmath0-cluster breaking , the spatial development of the cluster structure and the gas - like feature of the @xmath4 state are slightly reduced .
orthogonality to the @xmath8 state can not hinder the @xmath4 state including more compact @xmath0 cluster components because the 3@xmath0 cluster component in the @xmath8 state decreases with @xmath0-cluster breaking .
because of @xmath0-cluster breaking , the structure of the @xmath5 state changes from the vibration mode of acute and obtuse triangle configurations to the chain - like open triangle structure .
the components of acute triangle configurations increases in the @xmath4 state ; therefore , the @xmath5 state loses components of acute triangle configurations because of orthogonality to the @xmath4 state . because of structure changes of @xmath3 states by @xmath0-cluster breaking , transition strengths change significantly .
the band assignment for the @xmath6 state differs between calculations with and without @xmath0-cluster breaking .
namely , in the model calculation without @xmath0-cluster breaking , the @xmath4 state is assigned to be the band - head of the @xmath6 state . however , when we incorporate @xmath0-cluster breaking caused by the spin - orbit force , the @xmath5 state is regarded as the band - head of the @xmath6 state .
the @xmath5 state is likely to be the band - head of the @xmath6 state instead of the @xmath4 state because the present calculation reproduces experimental data of the low - lying states well with @xmath0-cluster breaking .
considering the stronger @xmath11 transition strength and maximum amplitude around the chain - like open triangle structure , this band can be considered a linear chain - like band .
unfortunately , there is no experimental data for the @xmath11 transitions from the @xmath6 state to excited @xmath3 states .
moreover , experimental information about two @xmath66 states around 10 mev is not enough to assign the theoretical @xmath5 state to either of the experimental @xmath3 states .
further experimental data related to @xmath3 and @xmath53 states in this energy region are required to clarify the band assignment of these states .
we stress that @xmath0-cluster breaking caused by the spin - orbit force gives significant effect on cluster structures , transition strengths , and band structure . to investigate structures of excited cluster states in @xmath1c , it is important to consider @xmath0-cluster breaking , in particular , mixing of the @xmath2 subshell closure configuration in @xmath3 states .
the computational calculations of this work were performed by the supercomputers at yitp .
this work was supported by jsps kakenhi grant nos .
25887049 and 26400270 . 99 k. ikeda , n. takigawa , and h. horiuchi , prog .
. suppl . * e68 * , 464 ( 1968 ) .
n. itagaki , s. aoyama , s. okabe , and k. ikeda , phys .
c * 70 * , 054307 ( 2004 ) .
y. kanada - enyo , phys .
lett . * 81 * , 5291 ( 1998 ) .
t. neff and h. feldmeier , nucl .
a738 * , 357 ( 2004 ) .
y. kanada - enyo , prog .
theor . phys . *
117 * , 655 ( 2007 ) .
m. chernykh , h. feldmeier , t. neff , p. von neumann - cosel , and a. richter , phys .
lett . * 98 * , 032501 ( 2007 ) .
y. kanada - enyo , m. kimura , and a. ono , prog .
* 2012 * , 01a202 ( 2012 ) .
y. fukuoka , s. shinohara , y. funaki , t. nakatsukasa , and k. yabana , phys .
c * 88 * , 014321 ( 2013 ) .
e. uegaki , s. okabe , y. abe , and h. tanaka , prog .
. phys . * 57 * , 1262 ( 1977 ) .
y. fukushima and m. kamimura , _ in proceedings of the international conference on nuclear structure , tokyo , 1977 _ , edited by t. marumori [ j. phys .
* 44 * , 225 ( 1978 ) ] .
a. tohsaki , h. horiuchi , p. schuck , and g. rpke , phys .
lett . * 87 * , 192501 ( 2001 ) .
y. funaki , a. tohsaki , h. horiuchi , p. schuck , and g. rpke , phys .
c * 67 * , 051306(r ) ( 2003 ) . y. funaki , a. tohsaki , h. horiuchi , p. schuck , and g. rpke , eur .
j. a * 24 * , 321 ( 2005 ) .
t. suhara , y. funaki , b. zhou , h. horiuchi , and a. tohsaki , phys .
* 112 * , 062501 ( 2014 ) .
y. funaki , arxiv:1408.5855 [ nucl - th ] and private communication .
t. yamada , y. funaki , h. horiuchi , k. ikeda , and a. tohsaki , prog .
. phys . * 120 * , 1139 ( 2008 ) .
p. navrtil and w. e. ormand , phys .
c * 68 * , 034305 ( 2003 ) .
p. navrtil , s. quaglioni , i. stetcu , and b. r. barrett , j. phys .
g : nucl . part .
phys . * 36 * 083101 ( 2009 ) .
t. neff and h. feldmeier , j. phys .
ser .. * 111 * 012007 ( 2008 ) .
e. epelbaum , h. krebs , t. a. lhde , d. lee , u. meiner , phys .
lett . * 109 * , 252501 ( 2012 ) .
t. suhara and y. kanada - enyo , prog .
* 123 * , 303 ( 2010 ) .
m. freer _ et al _ , phys .
c * 80 * , 041303(r ) ( 2009 ) . m. itoh _
et al _ , .
c * 84 * , 054308 ( 2011 ) .
w. r. zimmerman _
et al _ , phys .
rev . lett . * 110 * , 152502 ( 2013 ) .
w. r. zimmerman , `` direct observation of the second 2@xmath67 state in @xmath1c '' ( 2013 ) .
doctoral dissertations .
paper 230 .
et al _ , j. phys .
. ser . * 436 * , 012006 ( 2013 ) . c. kurokawa and k. kat , phys .
c * 71 * , 021301 ( 2005 ) . c. kurokawa and k. kat , nucl
a792 * , 87 ( 2007 ) . s. ohtsubo , y. fukushima , m. kamimura , and e. hiyama , prog .
exp . phys . * 2013 * , 073d02 ( 2013 ) .
d. m. brink , _ proc .
school of physics enrico ferm , course 36 _ , varenna , ed . c. bloch ( academic press , new york , 1966 ) .
s. okabe and y. abe , prog .
. phys . * 61 * , 1049 ( 1979 ) .
n. itagaki , h. masui , m. ito , and s. aoyama , phys .
c * 71 * 064307 ( 2005 ) .
t. suhara , n. itagaki , j. cseh , and m. poszajczak , phys .
c * 87 * , 054334 ( 2013 ) .
a. volkov , nucl .
* 74 * , 33 ( 1965 ) . n. yamaguchi , t. kasahara , s. nagata , and y. akaishi , prog .
1018 ( 1979 ) . f. ajzenberg - selove and j. h. kelley , nucl .
a506 * , 1 ( 1990 ) .
a. ozawa , t. suzuki , and i. tanihata , nucl .
a693 * , 32 ( 2001 ) .
danilov , t.l .
belyaeva , a.s .
demyanova , s.a .
goncharov , and a.a .
ogloblin , phys .
c * 80 * , 054603 ( 2009 ) . | to clarify the effects of @xmath0-cluster breaking on 3@xmath0 cluster structures in @xmath1c , we investigate @xmath1c using a hybrid model that combines the brink - bloch cluster model with the @xmath2 subshell closure wave function .
we have found that @xmath0-cluster breaking caused by spin - orbit force significantly changes cluster structures of excited @xmath3 states through orthogonality to lower states .
spatially developed cluster components of the @xmath4 state are reduced .
the @xmath5 state changes from a vibration mode in the bending motion of three @xmath0 clusters to a chain - like 3@xmath0 structure having an open triangle configuration . as a result of these structure changes of @xmath3 states , the band assignment for the @xmath6 state is changed by the @xmath0-cluster breaking effect .
namely , in model calculations without the @xmath0-cluster breaking effect , the @xmath4 state is assigned to be the band - head of the @xmath6 state . however , when we incorporate @xmath0-cluster breaking caused by the spin - orbit force , the @xmath5 state is regarded as the band - head of the @xmath6 state . |
prospects for the direct detection of dark matter depend crucially on the phase space distribution of dark matter near the sun @xcite .
unfortunately , detailed predictions for the dark matter distribution around the earth are extremely difficult to construct from first principles ; they require an understanding of spatial clumping on the scale of the solar system @xcite , and are almost certainly affected by poorly - understood baryonic processes like the formation of the galactic disk .
recently , @xcite emphasized that the process of cosmological disk galaxy formation can significantly alter the dark matter distribution compared to canonical predictions that rely on simulations of the standard halo model . during the process of hierarchical structure formation
, merging satellite galaxies can get dragged into the plane of their host disk and deposit their dark matter in a structure , dubbed the _ dark disk _
@xcite , that is co - rotating with the milky way stellar disk and morphologically resembles a thick disk . if the @xmath0cdm cosmology represents the correct model of structure formation in the universe , it is certain that dark disks are virtually ubiquitous in disk galaxies .
however , this contribution relative to the smoother halo component will depend sensitively on the formation process of each galaxy individually . in this work ,
we provide the first focused attempt at constraining the dark disk contribution in the galaxy , improving upon the initial discussion of the milky way dark disk in @xcite by comparing our simulations to the detailed kinematic and morphological properties of the galactic thick disk .
if a substantial dark disk exists within the milky way , then it has dramatic implications for the direct and indirect detection of dark matter , should that species consist of weakly interacting massive particles ( wimps ) , which are pervasive in models of new physics at the weak scale .
a co - rotating disk of dark matter will leave temporal modulation signals in terrestrial nuclear - recoil experiments .
it will also increase wimp capture in the sun and earth and enhance the resultant flux of neutrinos from wimp self - annihilation ; both of these results have been discussed in detail by @xcite .
we show in this work that a disk - like structure could be discernible in the self - annihilation signal ( depending on the total flux ) from the center of the galaxy .
since the scale height of the dark matter disk will be larger than the thin disk scale height and smaller than any background halo flattening , such a morphological feature could provide an important handle on this indirect detection signal .
the fraction of dark matter locked up in a co - rotating component is expected to depend sensitively on merger history @xcite , and this creates an important link between dark matter detection experiments and efforts in galactic astronomy to constrain the accretion history of the milky way .
evidence is mounting that the merger history of the galaxy is unusually quiescent compared to _ typical _
@xmath0cdm expectations @xcite .
high - resolution simulations have shown that 1:10 mass - ratio accretion events , despite being fairly common in the @xmath0cdm paradigm @xcite , are irreconcilable with the cold and thin galactic disk ( * ? ? ? * hereafter pkb09 ) .
this conclusion is consistent with the results of @xcite and @xcite , who used simulations to show that 1:10 mergers generate heated systems that are grossly consistent galactic _
thick disks _ , but that are clearly incommensurable with the dominant thin disk of the galaxy .
more recently , the simulations of @xcite showed that while these common mergers may be consistent with the broad population of galaxies , they produce remnants that are clearly thicker than the milky way within @xmath4 pc of the disk plane .
empirically , the galaxy appears to be deficient in stellar mass and angular momentum when balanced against a sample of local spiral galaxies @xcite , further motivating the need for a focused program that constrains the milky way s dark disk specifically , rather than relying on general cosmological expectations .
= 3.4 in we extend the methodology of pkb09 and use high - resolution simulations to study dark disk production .
our main observational constraint comes from observed kinematic and spatial properties of the old ( @xmath5 gyr ) thick disk of the milky way , which comprises at most @xmath6 of the total mass of the galaxy ( _ e.g. _ * ? ? ? * ) .
we initialize stellar disks that are consistent with the observed properties of the milky way today and consider the impact of fairly massive , cosmologically common accretion events .
this approach is conservative : by initializing a disk that is as massive as the milky way disk today , we are exploring a case that is both more resistant to heating and more efficient in capturing dark matter from the accreted satellite than any realistic progenitor of the milky way s thick disk . more generally , by focusing on the oldest , thickest , and hottest stellar component of the galactic disk , we are able to provide a most conservative constraint on the milky way s merger history .
the heating of the initial disk and the creation of the dark disk are intimately connected , since heating requires the transfer of kinetic energy from the incoming satellite to the disk .
for a fixed set of orbital parameters , more massive accretion events produce more stellar heating , and they also deposit more dark matter into a disk - like configuration .
the transfer of kinetic energy depends on how close the satellite gets to the disk and its orbit about the disk ; a closer and slower orbit results in larger heating and concomitantly larger stripping of dark matter from the satellite by the tides of the milky way .
thus , by comparing the heated stellar disk to the observed milky way thick stellar disk , we are able to bound the amount of coherently rotating dark matter that is deposited during the disruption of a large satellite galaxy . as alluded to above , the milky way has both a thin and thick disk @xcite , with the latter component being significantly hotter and thicker than the former . by comparing the stellar remnants in our simulations with those of the galactic thick disk implicitly , we allow for the possibility that the thin disk of the milky way is regrown later through fresh accretion . while the regrowth of a thin disk could potentially act to make the heated disk thinner , it can not make the heated disk colder .
thus our constraints based on the resultant velocity ellipsoid are the most robust .
the majority of the accretion events we consider have mass ratios @xmath7:@xmath8:@xmath9 .
this type of merger represents the dominant mode by which dark matter halos grow in @xmath0cdm models @xcite .
they are also the most relevant for the formation of co - rotating dark disks , because more massive mergers are both quite rare and catastrophically destructive to primary disks , while smaller mergers are less heavily affected by dynamical friction and leave behind much less dark matter ( we explore such a case below ) .
thus the 1:10 events we consider here are both cosmologically common and capable of producing a co - rotating dark disk without completely destroying the primary galaxy .
note that while a number of less massive mergers may incrementally build up a thick disk ( see * ? ? ?
* ) , they will not produce a coherently rotating dark disk unless the angular momenta contributed by the individual accretion events also coincide constructively .
this tendency toward a series of prograde accretions is not found in @xmath0cdm halos , at least for the few merger histories that have been explored in detail @xcite .
= 6.5 in in addition to collisionless experiments , we perform a hydrodynamical simulation in which the primary galaxy also hosts a gaseous disk , ( a much more thorough treatment of these and similar results can be found in kazantzidis , purcell , & bullock 2009 , in preparation ; see also * ? ? ?
* ) to determine whether this component plays a substantial role in the formation of a dark disk .
we find that our results do not depend sensitively on the presence of gas as we have modeled it , though we caution that the treatment of ism gas physics in galaxy simulations remains uncertain .
based on our implementation , the presence of gas does not curtail the heating of a primary disk that occurs during dark disk creation .
in a suite of high - resolution collisionless simulations , pkb09 investigated the response of a galactic disk to the infall of cosmologically - common accretion events involving satellites with one - tenth the mass of the host halo . for our primary system
, we focus here on the galaxy 1 model from pkb09 , which is a good match to the milky way today , and should provide a higher probability of dark disk creation via satellite dragging than any presumably less - massive progenitor of the galaxy at higher redshift .
we now provide a brief overview of our simulations and refer the reader to pkb09 for a more complete discussion of the methods and results regarding the morphological and dynamical changes undergone by stellar disks in response to these accretion events .
our primary galaxy is constructed according to the fully self - consistent distribution functions prescribed by @xcite and is therefore an equilibrium solution to the coupled collisionless boltzmann and poisson equations .
the model s stellar disk is initialized with an exponential scale length @xmath10 kpc and a vertical distribution described by a sech@xmath11 function with scale height @xmath12 kpc , containing @xmath13 particles with a total disk mass @xmath14 .
the massive central bulge follows a srsic profile of effective radius @xmath15 kpc and index @xmath16 , and has a mass @xmath17 distributed among @xmath18 particles .
these stellar components are embedded in a dark host halo composed of @xmath19 particles which follow the canonical nfw density profile of @xcite , with scale radius @xmath20 kpc and virial mass @xmath21 .
this particular set of parameters was chosen in order to minimize secular effects such as bar formation , as well as artificial heating induced by the interaction of disk particles with more massive halo particles .
the abscissa of figure [ fig : properties ] as well as the top row of table [ accretedtable ] denote the infall models we consider .
every simulation tracks the infall of a satellite galaxy modeled within a subhalo of virial mass @xmath22 at @xmath23 .
the satellites are each composed of @xmath24 particles in a density structure well fitted by an nfw profile with concentration @xmath25 , and contain a stellar mass @xmath26 distributed among @xmath27 particles in a spheroid with srsic index @xmath28 .
the stellar masses for the satellite galaxies are drawn from the model of @xcite , which describes the redshift - dependent connection between the observed spatial abundance of galaxies and the corresponding abundance of predicted dark matter halos .
fiducially , we investigate five orbital infall inclinations , including four prograde orbits with @xmath29p , and @xmath30 and one retrograde @xmath31r orbit .
the satellites are initialized 120 kpc from the host halo center , and we set our initial orbital vectors according to the distribution of substructure accretions drawn from cosmological simulations @xcite , with radial and tangential velocity components equal to @xmath32 km s@xmath33 and @xmath34 km s@xmath33 respectively .
additionally , in a model we designate @xmath35s , we investigate a case where the same satellite galaxy is traveling more slowly than in the fiducial @xmath35 case , with @xmath36 km s@xmath33 and @xmath37 km s@xmath38 ; these orbital parameters correspond roughly to the lower @xmath39 limit of @xcite . to test the effect of a full treatment of hydrodynamics on the morphological and dynamical changes induced by an accretion event , we repeat the simulation involving a massive satellite galaxy infalling along a prograde orbit with an inclination of @xmath40 ( hence the designation @xmath40 g ) , having converted a modest fraction @xmath41 of the initial primary galaxy s star particles into a gaseous component with the same density structure as the stellar disk .
our hydrodynamical prescription includes atomic cooling for a primordial mixture of hydrogen and helium , and the star formation algorithm is based on the work of @xcite , in which gas particles in cold and dense regions form star particles at a rate proportional to the local dynamical time ; star formation occurs when the gas density exceeds 0.1 @xmath42 and the gas temperature drops below @xmath43 k. supernova feedback is implemented according to the blast - wave model described in @xcite , in which the energy deposited by a type - ii supernova into the surrounding gas is @xmath44 ergs .
this parameter set produces realistic galaxies in cosmological simulations @xcite .
again we refer the reader to kazantzidis , purcell & bullock ( 2009 , in preparation ) for more details regarding the suite of experiments to which this current test case belongs . as a test of the fractional contribution made by a much smaller satellite galaxy
, we simulate an additional accretion event with identical orbital parameters to the fiducial infall , involving a subhalo of mass @xmath45 , i.e. with mass ratio 1:25 .
for this case , we choose a planar subhalo orbit ( @xmath46 ) in order to maximize the likelihood of dark disk formation .
however , this experiment resulted in a trivial increase ( @xmath47 ) in the fraction of dark matter in the solar neighborhood , and is thus not listed in table [ accretedtable ] .
we summarize in figure [ fig : properties ] the observable properties of our initial and remnant stellar disks , with each model indicated by name along the horizontal axis .
the upper panel shows the disk scale height @xmath48 , derived from fitting the minor axis surface density profile at a defined solar radius ( @xmath49 kpc ) , using the two - component form @xmath50 this decomposition provides a minimum estimate of the resultant disk scale height ( @xmath48 ) by allowing for a secondary , much thicker component ( @xmath51 ) , which dominates at large height and low surface brightness .
generally , we find significant variance in these values with orbital inclination angle ; the disk scale height @xmath52 kpc , while the faint component typically has @xmath53 kpc .
note that the initial thin disk of the primary galaxy has been destroyed : our minimal scale heights are comparable to or larger than the galactic thick disk scale height , as demonstrated in the comparison of figure [ fig : properties ] , where observed scale heights for the milky way thin and thick disks are shown as shaded horizontal bands .
these values are drawn from @xcite , in which an exponential disk scale height is presented ; we convert this value here to a sech@xmath11 scale height such that the two vertical surface density profiles decrease by roughly the same amount within @xmath54 scale height of the disk plane .
the total three - dimensional velocity dispersion @xmath55 of each disk at the solar neighborhood is presented in the bottom panel of figure [ fig : properties ] . the significant increase in this quantity from the initial model to
the resultant disk betrays the extraordinary heating undergone by the primary disk during the merger .
the shaded bands again reflect observed velocity dispersions for the milky way thin and thick disks , as drawn from @xcite and @xcite , respectively .
= 3.7 in = 3.4 in if we associate our remnant disks with the milky way thick disk , most of our resultant systems are thicker and hotter than the galaxy .
this comparison is particularly constraining for two reasons .
first , the age distribution of stars in the _ thin disk _ includes stars that are 8 - 10 gyr old , suggesting than any merger of this kind must have happened well before @xmath56 ( when the primary disk was even smaller and less able to cause significant dragging into the disk plane ) .
moreover , while regrowth of a new disk could potentially reduce the scale height of the thickened disk , it would not reduce the velocity dispersion of the stars .
only the prograde and retrograde @xmath40 orbits and the @xmath57 degree orbits result in a system that is marginally viable , both morphologically and dynamically .
each of the merger simulations provides an accreted dark component and an accreted stellar component , but the nature of these components depends sensitively on the interaction , as illustrated in figure [ fig : densitymaps ] for three of our simulations . for reference , the upper left panel presents a central slice of the dark matter density in the primary galaxy s host halo prior to the accretion , the middle left panel shows a slice in the stellar distribution in the primary galaxy , and the lower panels represent the heated stars belonging to the original disk only .
the three columns on the right for the upper two rows display the resultant distributions of accreted dark matter and accreted stars ; as might be expected , the lower - latitude event produces the most disk - like accreted dark matter morphology , with a qualitatively similar accreted stellar distribution that is somewhat more rotationally supported owing to the fact that the stars are more tightly bound in the satellites than the dark matter .
as noted in table [ accretedtable ] , the low - latitude simulations also produce the largest fractional density @xmath58 of accreted dark matter compared to background halo dark matter in the solar neighborhood kpc and bounded vertically by @xmath59 kpc ; we also note that variance in these boundaries does not lead to significantly altered results . ] .
high - latitude events create less discernible dark disks at the solar position and deposit proportionally less dark matter there ; we also note that our test model involving the planar infall of a satellite galaxy with mass ratio @xmath7:@xmath8:@xmath60 evolves over a much longer timescale than that of the fiducial models , and also results in a negligible contribution to the local dark matter quotient , indicating that even a multitude of small subhalos with similar orbits may not be able to form a significant dark disk .
for the models with mass ratio @xmath7:@xmath8:@xmath9 , table [ accretedtable ] provides dark disk and accreted stellar disk scale heights at the solar neighborhood , as computed by fitting a sech@xmath11 profile .
an additional characterization of interest is the velocity distribution of the dark disk component .
it is useful to parametrize the dark matter in the solar neighborhood of our simulations with a double gaussian distribution that represents the original host and accreted subhalo material : @xmath61 where in each term @xmath62 the sum involves each velocity component , @xmath63 .
we work in a frame that co - rotates with the resultant stellar disk such that the host halo produces a net dark matter wind in the @xmath64 direction that is close to the primary galaxy rotation speed @xmath65 km / s and @xmath66 .
accreted matter can co - rotate or anti - rotate with the stellar disk and we quantify this rotation in accreted material by the lag speed @xmath67 , such that a zero lag corresponds to precise co - rotation , while a positive lag trails the stellar disk and a negative lag means that the dark disk is rotating faster than the stellar disk .
the resultant velocity distribution parameters for each accreted component in our simulations are presented in table [ accretedtable ] . in figure
[ fig : velocity ] , we show the collapsed distributions in rotational velocity @xmath68 for the background host halo as well as accreted dark matter , along with our double - gaussian fit to the total amount of dark matter in the solar neighborhood , for the same three resultant galaxies that were illustrated in figure [ fig : densitymaps ] : @xmath69 , @xmath40p , and @xmath40 g .
the center panel also presents the velocity distribution for the @xmath40r retrograde merger .
the filled histograms in each panel show the velocity distributions for accreted stars , normalized relative to the total stellar mass at the solar location .
two points are immediately apparent ; first , that lower - latitude accretion events produce smaller lag velocities in both accreted species , and secondly that the lag speed of accreted stars grossly mimics that of the accreted dark matter for accretion events at all latitudes . the kinematic relationship between the accreted dark and stellar disks ( a phenomenon also reported by * ? ? ?
* ; * ? ? ?
* ) is important because it may provide an avenue to constrain the dark disk velocity distribution via observational studies that can isolate an accreted stellar component . in figure
[ fig : correlation ] , we show the correlations between accreted dark matter and accreted stars in each velocity dispersion component , as well as the rotational lag speed , which correlates very well between the two accreted species . we also note that the velocity dispersion of accreted stars generally provides a lower limit on the velocity dispersion of the dark disk in each directional component ; the correlations in figure [ fig : correlation ] indicate that dark matter dispersions are consistently at least a factor of @xmath70 larger than their stellar counterparts .
though not explicitly shown here , we also find a connection between the velocity - ellipsoid axis ratio @xmath71 of the accreted dark matter and that of the stars formerly belonging to the satellite galaxy , in that the two collisionless events involving the high inclination angle of @xmath72 both have vertical to radial velocity dispersion ratios of approximately unity for all accreted material , while the velocity ellipsoids resulting from the low - latitude events are significantly more oblate , with both stellar and dark axis ratios roughly equal to @xmath73 : a similar value to that obtained observationally for the galactic solar neighborhood @xcite .
it should be noted here that the polar infall results in a final velocity - ellipsoid axis ratio coincident with that of the low - latitude events , due to a very large dispersion in the radial velocity of both accreted species in this case .
we use two simple metrics to characterize our main results in figure [ fig : darkproperties ] ; as in figure [ fig : properties ] , each model is indicated along the horizontal axis .
the upper panel presents the accreted dark matter fraction @xmath58 in the solar neighborhood ; the lower panel presents the fraction of local dark matter that is roughly co - rotating with the stellar disk , which is potentially the most salient characteristic of an accreted dark disk . specifically , we define the latter quantity to be the fraction of accreted dark particles with rotational speeds within 50 km / s of the stellar disk s velocity , _
i.e. _ @xmath74 km / s .
note that by this definition , even the initial spherical halo has a non - zero co - rotating dark matter fraction ; for reference , the dotted line in the lower panel of figure [ fig : darkproperties ] shows the fraction of dark matter particles initially from the background halo that are co - rotating with the stellar disk in the final simulation snapshot .
as might be expected , the most significant dark disks form during the planar accretion events @xmath35 and @xmath35s , which respectively produce @xmath75 and @xmath76 accreted dark matter fractions in the solar vicinity and factors of @xmath77 and @xmath78 enhancements in the co - rotating dark matter fraction .
however , these planar events also produce stellar disks that are much hotter than the thick disk of the milky way ( as shown in figure [ fig : properties ] ) , indicating that these cases are not appropriate baseline models for the galaxy .
though not evident in figure [ fig : properties ] , the slower orbit actually produces more radial heating than the fast orbit ( see table [ accretedtable ] ) and is therefore more discrepant with the milky way thick disk velocity structure than the standard orbit . the three shaded bands in figure [ fig : darkproperties ] highlight the three model cases that produce stellar disks marginally consistent with the thick disk of the milky way , when considering morphology as well as dynamical temperature .
the most optimistically viable model for a dark disk is the 60-degree prograde orbit , which results in a @xmath79 accreted dark matter contribution and a @xmath80 co - rotating dark matter fraction ; the latter quantity represents a @xmath81 increase compared to the co - rotating dark matter fraction in the smooth halo case . as noted in table [ accretedtable ] , the dark disk scale height in this case is @xmath82 kpc and lags the lsr with @xmath83 km / s .
= 6.5 in although the thick disk we produce is obviously more massive than the galactic thick disk , the similarity in morphology and kinematics motivate us to move beyond limits , and speculate on the properties of the actual dark disk in the milky way . given the range of parameters probed here and in @xcite , and operating under the assumption that the milky way s thick disk was formed in an event similar to the type we simulate here
, we expect that the real galactic dark disk contributes roughly @xmath84 to the local dark matter density . as we discuss below in [ sec : summary ] , such a dark disk does not significantly enhance the possibility of direct detection of dark matter ; however , it will be important to include in any detailed analysis aimed at interpreting an observed wimp detection signal .
moreover , such a dark disk may provide a detectable morphological signature towards the galactic center in experiments designed to indirectly detect dark matter via its self - annihilation products .
in a recent effort to quantify disk heating in a @xmath0cdm context , @xcite used simulations similar to those presented here in order to show that massive satellite accretion events involving a galactic - type thin disk not only produce thick stellar systems but also result in the deposition of accreted dark matter into a disk - like component that co - rotates to some degree with the primary galaxy . specifically , they analyzed dark disk formation during the infall of various subhalo types ( of particular interest here are their lmc and llmc models , representing satellite galaxies with total mass @xmath85 and @xmath86 respectively ) , finding that for low - inclination mergers , the resulting density ratio of accreted dark matter to that of the background host halo in the solar neighborhood was @xmath87 for the smaller subhalo and @xmath88 for the more massive satellite that corresponds closely to our fiducial infalling system .
in the follow - up analysis of cosmological simulations of milky - way - type disk galaxies @xcite , the authors found similar results for systems with a wide range in accretion history activity , showing that several massive mergers at late times can result in a dark disk with local density similar to that of the host halo itself .
although our simulated phenomena are qualitatively similar to those presented by @xcite , there are two key points of distinction to be drawn .
the two methods are only significantly discrepant in the choice of initial radii and velocities of the infalling satellites .
our work adopts satellite galaxy velocity vectors drawn from subhalo infall distributions found in cosmological simulations , and thus we set the initial position of the subhalo far ( @xmath89 kpc ) from the host halo s center in order to minimize the sudden change in potential felt by the primary galactic disk and also to imitate the cosmological velocity conditions , which are measured at the host s virial radius .
in contrast , the infalling satellites of @xcite begin much closer ( @xmath90 kpc ) to the primary disk s center and involve subhalos traveling at speeds slower than our fiducial initial velocity by roughly fifty percent ; slightly faster than our @xmath91s case .
the authors motivate the choice of a slow orbit at a small initial galactocentric apocenter by suggesting it replicates the orbital parameters of satellites found in a loose group environment that is accreted onto a galaxy - sized host .
however , it still remains to be shown that such a velocity vector at @xmath90 kpc can arise naturally for a massive satellite within a cosmological setting .
moreover , any process that would act to bring a subhalo to such a low energy state in the disk plane would almost certainly heat the primary disk to a degree unallowable by observations . the difference in initial conditions
may be largely responsible for the small systematic differences between our simulated dark disk results and those of @xcite . while our maximal local density ratio is @xmath92 , they find @xmath88 in their experiment labeled llmc-@xmath93 , an accretion event involving a subhalo of similar mass .
the methodological variance may also explain why our disks are slightly dynamically hotter .
our @xmath35 case has @xmath94 km / s and our @xmath35s case has @xmath95 km / s , while their llmc-@xmath93 model has @xmath96 km / s .
it is important to emphasize , however , that both sets of simulations find resultant thick disks that are hotter than the observed thick disk of the milky way , which has a total dispersion of just @xmath97 km / s .
therefore , in a broad sense , we agree that these disks are not ideal analogues to the milky way .
we note that a more substantial dark disk component could emerge following multiple similar accretion events , as investigated in cosmological simulations of galaxy - sized disks by @xcite , but such a series of late - time mergers can not be reconciled with the relatively thin disk of the milky way . unfortunately for experiments aimed at detecting local dark matter ,
the quiescent galactic accretion history favored by our result indicate that the relevant observable quantities are probably far less affected by a dark disk than we might have hoped .
in our experiments , we have adopted assumptions throughout that maximize the likelihood for producing a dark disk remnant and for preserving a thin , cold disk by initializing a primary disk that is as massive as the milky way disk today .
this approach is quite conservative because we compare the heated stellar remnant to the old _ thick disk _ of the milky way , which is at least five times less massive than our primary system ( _ e.g. _ * ? ? ?
* and references therein ) .
we have argued that the dark disk formed in our @xmath31 prograde merger case provides a fair limit on the properties of an underlying accreted dark component of our galaxy , with a local contribution @xmath98 that is at the low end of dark disk contributions estimated by @xcite for milky way type galaxies in @xmath0cdm .
based on this limit , we expect that nuclear recoil experiments designed to directly detect local dark matter would only receive small boosts to the event rate , preferentially at lower energies .
dark disk fractions of @xmath99 were explored by @xcite , who found an order unity enhancement in the differential event rate at kev recoil energies for a 100 gev dark matter particle . for a tev mass dark matter particle , such an enhancement would be present at higher energies ( accessible to cdms - ii ) .
while an order unity enhancement will not significantly impact the event rate , @xcite showed that the phase of the annual modulation signal is sensitive to both the dark disk and the mass of the dark matter particle .
the phase of the annual modulation signal produced by the dark disk depends on the motion of the sun relative to the dark disk .
the phase of the modulation of the total event rate in a given energy window depends on both the mass of dark matter particle and the dark disk fraction .
if the dark disk fraction is bracketed from the lower end , this will result in a prediction for the phase given a dark matter particle mass .
this may be testable at future detectors if a positive signal is seen .
in addition , these results also imply that directional detectors should see the wimp wind direction change as a function of recoil energy , the details of which would depend on the energy window , the mass of the wimp , the dark disk fraction and the motion of the sun with respect to the dark disk .
an increase in the local number density of particles that co - rotate with the sun will also enhance wimp capture in the sun and the earth .
these captured particles self - annihilate into standard model particles , including neutrino pairs . thus a larger co - rotating fraction of dark matter will result in a larger neutrino flux potentially observable with experiments like super - kamiokande and icecube . in detail ,
the capture rate depends sensitively on @xmath58 , the dark disk lag speed @xmath100 , and the velocity dispersion of the dark disk .
the timescale for capture of wimps at earth is small compared to the age of the planet and hence the wimp density in the earth has not yet reached equilibrium . in the sun , as shown by @xcite , the situation is the opposite for most regions of parameter space and the number of wimps would have reached an equilibrium value .
thus the annihilation rate of wimps inside the earth is proportional to the capture rate squared , while in the sun it is equal to half the capture rate ( given the equilibrium condition there ) .
@xcite show that this leads to two to three orders of magnitude enhancement in the flux from the earth and an order of magnitude enhancement in the flux from the sun for @xmath101 and an assumed isotropic gaussian velocity distribution for the dark disk with lag @xmath102 km / s and 1-d dispersion @xmath103 .
they also consider dark disk parameters @xmath104 and @xmath105 km / s and state that this does not lead to a large boost in the signal , owing primarily to the large velocity dispersion . by comparison , our simulated prograde accretion event with @xmath72 results in a local density ratio of @xmath106 , velocity dispersion @xmath107 km / s , and similar lag speed @xmath108 km / s .
thus we do not expect large boosts to the neutrino flux from the earth or the sun , and our results imply that the likelihood of wimp detection through this indirect channel is not significantly increased due to an accreted dark disk .
although the prospects for direct detection and indirect detection via wimp capture are not significantly altered by the accreted dark disk , we note that the accreted dark matter distribution is relatively more prominent toward the center of the galaxy , where most of the satellite galaxy s material settles .
the center of the galaxy is a prime target for experiments aimed at detecting dark matter indirectly via high energy annihilation products @xcite , as the flux scales proportionally to the square of the wimp density . in figure
[ fig : annihilation ] , we view the halo s center from the vantage point of the solar neighborhood , after the prograde @xmath40 satellite infall , and estimate the annihilation signal at earth by integrating @xmath109 along lines of sight .
we notice immediately that while the host s dark matter has retained approximate sphericity ( left ) , the combined signal ( right ) which includes the accreted subhalo material ( middle ) has a significantly more oblate contour .
approximately @xmath110 of the total @xmath109 contribution at galactic longitudes between @xmath111 and @xmath93 comes from accreted dark matter . as in the case of direct detection
, the dark disk does not significantly change the likelihood of detection ; however , if there is a positive detection , then our results provide motivation to search for a disk - like component with scale height and length that are different from that of the milky way thin disk .
our focus in this work has been on the scenario of @xcite , who showed that dark disks could be created via satellite accretion events .
we have argued that under this scenario , the relative importance of the accreted dark disk can be constrained via detailed comparison to the properties of thick disk stars in the galaxy and that this constraint limits the dark disk s density contribution to @xmath112 in the solar neighborhood .
however , it is interesting to consider the theoretical possibility that a dark disk with density ratio @xmath113 could form in response to some other process that did not heat the disk .
for example , one could imagine that large infalling gas clouds could transfer angular momentum to the dark matter , creating a disk - like dark component without disturbing the primary disk significantly
. such a process may not be easy to arrange , given the large amount of angular momentum transfer required to produce a co - rotating dark disk as well as the fact that observed disk galaxies seem to require something close to angular momentum conservation .
nonetheless , it is an instructive exercise to determine the effect of an @xmath113 dark disk on indirect detection signals . to investigate this
, we have scaled the local density ratio for the @xmath40 prograde case such that @xmath114 , and the resultant annihilation signal s sky map is shown in the lower panels of figure [ fig : annihilation ] .
the disk - like nature of the signal is clearly apparent and could be easily distinguished if any dark matter self - annihilation is observed toward the center of the galaxy .
therefore , if such a dark disk could somehow form without involving satellite accretion events , it would be discernible through indirect detection experiments .
coupled with the boost in the direct detection signal that is expected for such an @xmath113 dark disk @xcite , the jointly observable consequences of such a component would be significant . in a similar vein ,
it is worth noting that the quantity of interest for direct detection experiments is the fraction of slow - moving dark matter particles . from figure
[ fig : darkproperties ] , we see that the @xmath115 case shows an order unity enhancement in slow - moving dark matter particles , and figure [ fig : properties ] shows that this run produces a disk that is only marginally hotter than the thick disk of the milky way . however , the scale height of the disk is considerably larger .
one may , however , hypothesize that a later stage of disk formation could change the disk scale height substantially without heating the stars further . therefore
, we urge the reader to keep in mind that @xmath113 enhancement in slow - moving particles is possible in the accreted dark disk scenario of @xcite , but that the expected velocity dispersion in the @xmath64 direction is still large in our @xmath115 case , it is 83 km / s ( see table [ accretedtable ] ) .
in this work we have argued that within the context of the accreted dark disk scenario of @xcite , it is likely that the dark disk of the milky way contributes approximately @xmath116 to the local dark matter density .
if so , then its presence may be important to include in any attempt to interpret a detection signal of wimp dark matter in the solar neighborhood and for the indirect detection signal from the galactic center . given this , it will be important to constrain the dark disk to a higher degree of accuracy .
both @xcite and we have demonstrated that there is a significant dynamical relationship between dark matter accreted during the infall of a massive subhalo and the stellar mass contributed by that satellite galaxy to the primary system ( table [ accretedtable ] and figure [ fig : correlation ] )
. therefore one can anticipate constraining the properties of any milky way dark disk by isolating a dominant subpopulation of accreted galactic disk stars , perhaps by some combination of chemical and dynamical tags .
for example , if we were to consider the unlikely scenario where the entire thick disk of the milky way was deposited by a single accretion event ( _ i.e. _ it is devoid of any pre - existing disk stars ) , then we may consider the dynamical properties of the galaxy s thick disk as a means to constrain the dark disk : @xmath117 km / s , @xmath118 km / s @xcite , and @xmath119 kpc @xcite . under this interpretation , our simulation results would imply that the lag speed of the mw dark disk is @xmath120 km / s and that it must be quite hot , with 1.25 @xmath121 km / s , and with a very thick @xmath122 kpc .
these numbers are almost certainly lower limits ; as we have shown in table [ accretedtable ] , accreted stellar distributions tend to be both hotter and thicker than the primary disks they heat .
if there is any pre - existing disk population associated with the thick disk of the milky way , the accreted portion is almost certainly hotter than the composite thick disk values used in our analysis .
although theoretical predictions based on galaxy formation models as they relate to dark matter detection are still in the nascent stage , it is clear that the standard halo model is insufficiently equipped to allow precise predictions for measurable quantities of interest ; scattering rates and flux scalings depend crucially on the presence of locally coherent substructural flows .
observations have thus far been unable to accurately estimate the degree to which the galactic thick disk s population is composed of stars stripped from a satellite galaxy during tidal disruption , as opposed to material heated to ejection from the thin stellar disk ( as investigated by * ? ? ?
* ) . in principle
, however , a combination of high - precision kinematic measurements and detailed chemical composition data should be able to define a phase - space in which accreted and original stars occupy disparate regions .
placing such dynamical constraints on the population of stars obtained by the milky way during a significant accretion event may enable us to identify the kinematics of the dark matter deposited during that same event , thereby placing more stringent limits on the availability of a dark galactic disk for the purpose of detection experiments .
we thank justin read , laura baudis , tobias bruch , and george lake for discussions that improved the accuracy and clarity of our presentation .
stelios kazantzidis assisted in the construction of of the initial conditions used in this work , and we thank him for allowing us to present the results of our collaborative disk - satellite experiments here , including the hydrodynamical experiment described above .
we also thank larry widrow and john dubinski for kindly making available the software used to set up the initial galaxy models .
cwp and jsb are supported by national science foundation ( nsf ) grants ast-0607377 and ast-0507816 , and the center for cosmology at uc irvine .
mk is supported by nsf grant ast-0607746 and nasa grant nnx09ad09 g .
the numerical simulations were completed primarily on the ia-64 cluster at the san diego supercomputing center , with ancillary experiments performed on the greenplanet cluster at uc irvine . | massive satellite accretions onto early galactic disks can lead to the deposition of dark matter in disk - like configurations that co - rotate with the galaxy .
this phenomenon has potentially dramatic consequences for dark matter detection experiments .
we utilize focused , high - resolution simulations of accretion events onto disks designed to be galaxy analogues , and compare the resultant disks to the morphological and kinematic properties of the milky way s thick disk in order to bracket the range of co - rotating accreted dark matter .
we find that the milky way s merger history must have been unusually quiescent compared to median @xmath0cdm expectations and therefore its dark disk must be relatively small : the fraction of accreted dark disk material near the sun is about @xmath1 of the host halo density or smaller and the co - rotating dark matter fraction near the sun , defined as particles moving with a rotational velocity lag less than 50 km / s , is enhanced by about @xmath2 or less compared to a standard halo model .
such a dark disk could contribute dominantly to the low energy ( of order kev for a dark matter particle with mass 100 gev ) nuclear recoil event rate of direct dectection experiments , but it will not change the likelihood of detection significantly .
these dark disks provide testable predictions of weakly - interacting massive particle dark matter models and should be considered in detailed comparisons to experimental data .
our findings suggest that the dark disk of the milky way may provide a detectable signal for indirect detection experiments , contributing up to about @xmath3 of the dark matter self - annihilation signal in the direction of the center of the galaxy , lending the signal a noticeably oblate morphology . |
as the youngest known isolated rotation powered neutron star , and having the highest known spin - down flux density , the crab pulsar is not surprisingly the most efficient source of high energy magnetospheric emission .
all theoretical models attempting to elucidate the processes of nonthermal emission must satisfy the empirical constraints provided by the extensive observational datasets available for this object from @xmath6-rays to the infrared .
contemporary theoretical frameworks in existence to explain the nonthermal emission from this and other isolated rotation - powered neutron stars may be broadly divided into two schools .
the first places the source of emission close to the polar cap region , e.g. sturrock ( 1971 ) , and the second that place the source at a considerable distance above the neutron star surface , in the outer magnetosphere ( e.g. cheng , ho & ruderman 1986 ) . in recent times , the development of these models has been to some extent restricted to the problem of @xmath6-ray emission .
these efforts have been predominantly analytical in nature , as in this energy regime primary emission is being sampled , which avoids complications associated with the lower energy sources of emission . in these latter cases ,
the emission is secondary in nature and a result of interactions that require numerical representation .
thus the polar cap school ( e.g. daugherty & harding 1982 ) concluded with correct flux estimates but poor light curves , and the outer gap school ( cheng , ho & ruderman 1986 ) with reasonable light curves and approximate flux distributions but requiring extraordinary lines - of - sights and magnetospheric physics to justify the emission .
the group at stanford university lead by r. romani is generally credited by being the first to rigourously put the existing models to the test numerically .
their conclusions in many ways matched the intuitive premonitions regarding the two model frameworks , with perhaps the ` evolved ' outer gap model of cheng , ho & ruderman ( 1986 ) being the more likely candidate ( chiang & romani 1992 ; romani & yadagiroglu 1995 ) . recently
romani ( 1998 ) has indicated that it is only by both the development of more advanced numerical simulations that incorporate the full emission physics ( from @xmath6-rays to the ir ) in a truly self - consistent manner and the accurate characterisation of pulsar light curves in the lower energy regimes that one can hope to disentangle the problem of nonthermal pulsar emission .
optically , spectroscopic , integrated and high speed single - pixel photometry have produced some critically important results in this regard .
however , true phase - resolved data acquisition with accurate isolation of the pulsar s emission from that of the nebula requires high time resolved ( @xmath7 1 @xmath1s ) 2-d photometry to confirm unambiguously earlier conclusions and to determine the true nature of the magnetosphere s emission in the range @xmath7 1 - 10 ev . in a previous paper ( golden et al .
2000 ) , we presented an analysis of such observations of the crab pulsar in which we isolated and theoretically speculated upon the unpulsed component of the pulsar s optical emission .
these observations were made in january 1996 with the triffid 2-d high speed photometer in three colourbands ( @xmath0 ) using the 6 m bta of the special astrophysical observatory in the russian caucasus .
here we present a thorough photometric and temporal analysis of the same data .
following a review of the pulsar s relevant empirical and theoretical characteristics , we briefly outline the observational and data reduction methodology - the reader is encouraged to consult golden et al .
( 2000 ) for a more rigorous treatment of this process .
we then detail the results of the photometric and temporal analysis , and conclude with a discussion on the implications of these results for current thinking on this pulsar , with wider ramifications for pulsar emission theory in general .
the crab pulsar is identified as the south - west member of the double stars located in the central region of the crab nebula ( cocke et al .
radio observations , initially by staelin @xmath8 reifenstein ( 1968 ) , and regularly ever since , have provided perhaps one of the most detailed pulsar rotational datasets to date .
the pulsar has a @xmath9 @xmath7 33 ms and @xmath10 @xmath7 4.211@xmath11 s@xmath12 suggesting a canonical age of @xmath13 @xmath7 1260 years and surface magnetic field @xmath14 @xmath7 3.7 @xmath15 gauss .
historically , high speed optical photometry has been generally dominated by the use of single pixel detectors , with variable time resolution .
following the detection of cocke et al .
( 1969 ) , numerous targetted observations followed ( kristian et al .
1970 ; wampler et al . 1969
; warner et al . 1969 ; becklin et al . , 1973 ; muncaster and cocke , 1972 ; cocke and ferguson 1974 ; groth 1975a , b ) with the pulsar generally observed in the @xmath16 wavebands at typical accurate time resolutions of order @xmath7 ms , and the datasets ` added ' together in a least - squares fashion - there being no common stabilised ` clock ' from which to time the observations , nor a definitive radio ephemeris with which to accurately obtain _ in modulo _ lightcurves at that time .
the development of more `` temporally consistent '' instrumentation in conjunction with other advances in technology allowed for more detailed observations to be taken .
peterson et al . ( 1978 ) applied for the time rather novel techniques in the image processing of data obtained via the use of a 6.2 ms time resolved 2-dimensional image photon counting system camera .
significant emission from the unpulsed presumed ` off ' component was noted in the @xmath17 and @xmath18 bands , with peterson et al . estimating that it was some 2% of the total pulsed emission . at an effective time resolution of 20 @xmath1s ,
smith et al .
( 1978 ) were able to make unprecedented observations via a s20 photocathode of the crab s main pulse region , indicating that the peak had a plateau of emission @xmath5 100 @xmath1s in duration .
photon binning at 1 khz allowed for phase resolved tests to be made of the counting statistics incident on the detector .
they yielded statistical distributions that did not deviate significantly from random fluctuations expected , making an allowance for variations in sky transparancy .
subsequent high speed observations incorporating an appropriate polarisation set - up yielded phase resolved polarimetry ( jones et al .
1981 , smith et al . 1988 ) . under analysis and corrected for interstellar polarisation , these datasets show evidence for sharp swings in polarisation angle around both principle peaks , in addition to some form of evolution of the polarisation in the bridge area . remarkably , the results indicated substantial polarisation associated with what was always assumed to be ` off ' or at the very least , the pulsar at a minimum , as noted earlier by peterson et al .
( 1978 ) .
the high speed photometer ( hsp ) on board the hubble space telescope ( @xmath19 ) was used to observe the crab pulsar in october 1991 and again in january 1992 , as reported by percival et al .
the passbands used were f400lp and f160lp , which crudely approximate to standard @xmath20 and @xmath21 . although the on - board clock is accurate to 10.74 @xmath1s , accuracy to utc was typically @xmath7 @xmath3 1.05 ms per observing run .
each resulting light curve had an effective temporal resolution of 21.5 @xmath1s .
the pertinent results of percival et al .
( 1993 ) are as follows .
they estimated that the main pulse in the visible leads the radio by -1.47 @xmath3 1.05 ms , and for the uv , -0.91 @xmath3 1.05 ms , with both @xmath21 and @xmath20 main peaks offset by 0.56 @xmath3 0.92 ms . via autocorrelation and least - squares analysis of the light curves , the full width half maximum ( fwhm ) , peak separation and peak ratios
were determined . despite the summing and normalising of the individual light curves which might be expected to introduce an element of bias in any analysis , percival et al .
( 1993 ) obtain consistent and reasonable estimates for these parameters .
the implication is that there is a trend whereby the pulse peaks ` tighten ' and come together as a function of energy .
resolution of the main peak cusp was attempted via a least - squares polynomial fit , and was found to be @xmath5 150 @xmath1s in extent . despite a number of caveats outlined previously percival et al .
( 1993 ) can claim to have shown to a reasonable degree that 1 .
the peaks fwhm and separation contract as a function of energy .
the @xmath21 and @xmath20 main peaks appear to be coincident to within @xmath7 1 ms .
the @xmath21 and @xmath20 main peaks show signs of leading the radio main peak .
4 . a plateau of emission is not definitively resolvable to within 150 @xmath1s .
5 . the photon statistics would appear to be poissonian throughout the light curve .
later observations , using the @xmath22 in polarisation mode , were made of the pulsar ( smith et al .
qualitatively the results were similar to earlier optical observations .
the conlusions reached were no different to the original 1988 paper , other than that any model of the emission regions must be essentially wavelength independent . following on from @xmath23 band observations of ransom et al .
( 1994 ) , eikenberry et al .
( 1996 ) made observations in the @xmath24 , @xmath25 and @xmath26 bands using a single - pixel solid state photomultiplier , set - up to record using 20 @xmath1s time bins .
light curves were then obtained in the usual manner .
the analysis proceeded along three avenues .
the first was determination of the half width half maximum ( hwhm ) , essentially a truncated fwhm , with each side of the fwhm bisected by the maximum of the peak in question , yielding a ` leading ' and ` trailing ' hwhm component .
the second was the determination of phase resolved spectra , and the third an in - depth examination of the colour variation around the main peak .
monte carlo simulations were used to build up a statistical picture of the optimum fits to the light curve data .
this analysis showed clear evidence that the peaks ( morphologically and in terms of flux ) change as a function of energy throughout the light curve , and also that time limit over which such rapid spectral changes occured was @xmath5 180 @xmath1s .
assuming emission occurs in the vicinity of the light cylinder , then one can estimate the typical coherence dimensions for the region , being less than or equal to 1.8 @xmath27 s @xmath28 @xmath29 = 54 km - indicating that the emission region is consistent with a localised region of the magnetosphere . in a subsequent paper ( eikenberry & fazio 1997 )
these techniques were applied to a range of light curves , from @xmath6-rays ( ulmer et al .
1995 ) , x - rays ( reconstructed from the @xmath30 archives ) , @xmath31 ( percival et al . 1993 ) and @xmath32 ( from previous work ) .
the overall energy range spanned 0.47 ev to 10 mev .
the principle intention was to further test the hypothesis that the actual peaks themselves altered morphologically with energy .
the actual mechanics of how the datasets were analyzed was similar to the earlier analysis of eikenberry et al .
( 1996 ) - it will suffice to note their more pertinent conclusions ; 1
. the ratios of the flux of the bridge and peak 2 to peak 1 show an overall trend to increase with energy , but reverses at low energies ( @xmath33 1ev ) .
the peak - to - peak separation shows a ` nonlinear ' trend to decrease with energy over the range , which is consistent with energy stratification of the emission region .
the peaks fwhm show neither a definite trend nor pattern with energy , yet display significant variability over the range .
4 . there is a strong energy dependency as regards the evolution of the hwhm for leading and trailing edges of both peaks , with some evidence of a continous functionality over the range - except at the @xmath23 region .
the hwhm behaviour explains the rather puzzling fwhm behaviour .
peak 2 appears to change profoundly as regards morphology between the @xmath23-optical and x - ray-@xmath6-ray range , showing a fast rise and slow fall for the former , and the converse for the latter .
this is not explicable by theory .
many of the pulse shape parameters show maxima or minima at the energies of 0.5 - 1ev in the @xmath23 regime , suggesting some interesting phenomena occuring in this waveband .
eikenberry & fazio ( 1997 ) concluded that the many unusual energy dependent morphological phenomena that are observed under this analysis were incompatible with current model frameworks - in particular the single - pole emitting outer - gap ( romani & yadigaroglu 1995 ) . in this model
, the emission originates from a topologically extended region that effectively maps out a hemisphere about the open field region .
consequently , with the emission profile a result of viewing geometry and crossing caustics to the line of sight , one would expect either no evident effect , or a common effect that should manifest itself in a smooth and continuous manner at all such energies , as the effect would be essentially the averaging of many seperate spatially emission sources .
these expectations contrast with the empirically determined results .
free precession of an isolated pulsar , most likely due to some moment of inertia inequality , possibly as a result of irregular distributions of neutron star matter on its surface , or perhaps as a result of internal interactions and effects , would manifest itself as an appropriate modulation in the observed temporal signal .
previous reports by groups examining the variations in the ratios of peak 1 to peak 2 in @xmath6-rays have generally indicated periods of the order of years ( kanbach et al .
1990 ; nolan et al . 1993 ; ulmer et al .
previous reports of something akin to a precessional like signal from the crab pulsar in the radio have tended to suggest a precessional period of order months / years ( e.g. jones 1988 ) . at higher frequencies , evidence of precession could provide both direct constraints to the condensed matter equation of state , and imply the emission of gravitational radiation .
it is well known that a precessing gravitationally compact body rotating at high speeds would be expected to radiate gravity waves , at a frequency twice its rotational frequency ( e.g. misner , thorne & wheeler , 1971 ) .
various other emission modes are theoretically possible .
planned ground based gravitational detectors are being designed that could have the capability to detect displacement amplitudes of order @xmath34 @xmath7 @xmath35 .
however do this , these detectors must be ` tuned ' to a particular tight frequency band for up to 1 year to build up sufficient @xmath36 .
clearly , one needs a guaranteed signal to lock on to - and pulsars may very well provide such a source .
hence the great interest in the reported suggestions of a precession signal with a period of order 60 seconds from the crab by cadez and galicic ( 1996 ) .
this was based on a time - series analysis of the high speed photometer data of percival et al .
( 1993 ) , and their own ground based observational ccd dataset . despite nearly three decades of radio observations , no such modulation has been reported in the literature .
furthermore , an analysis of high speed optical data by jones et al .
( 1980 ) with early photometers indicated a lack of optical source variations at the 1 % level over a ten year period .
it is clearly important to try and test this with independent data , as a confirmation would have important possible implications for other fields , particularly that of gravitational wave detection .
the details of these triffid high speed observations of the crab pulsar , and the subsequent preliminary analysis have been published elsewhere ( golden et al . 2000 ) .
we shall thus limit our description accordingly .
the observations took place over 5 nights in january 1996 using the special astrophysical observatory s 6 m telescope in the caucasus .
triffid high speed camera used incorporated a 2-dimensional photon counting mama detector ( @xcite ; @xcite ) with standard @xmath17 , @xmath18 and @xmath20 filters .
timing was accurate to within 400 ns per 10 second period using a gps receiver .
there were 6 @xmath17 band exposures totalling 5546 seconds , 5 @xmath18 band exposures totalling 5750 seconds , and 9 @xmath20 band exposures totalling 6759 seconds .
the photon [ x , y , t ] data stream per dataset was corrected via a weiner filter modified shift - and - add algorithm for telescope wobble and gear drift ( redfern et al .
1993 ) . processing the datasets in this way and incorporating the derived flat fields yielded full field images of the inner crab nebula from which the pulsar and its stellar companions were registered .
figure [ 96zs2 ] shows the integrated 96zs2.0.0 dataset following this technique .
the average seeing for this dataset was @xmath7 1.3@xmath37 .
the crab pulsar is evident as the slightly dimmer upper component of the central double star . as detailed in golden et al .
( 2000 ) , phase resolved images were obtained by firstly extracting all photons with a fixed pixel radius of the crab pulsar , barycentering them , and then geometrically aligning each dataset in [ x , y ] per colour band .
this yielded a sequence of time - series per colour band that , when read by a modified epoch - folding algorithm , generated both standard light curves and phase - resolved 2-dimensional image within a certain specified phase range .
the latter used the appropriate data taken from the jodrell bank crab ephemeris ( lyne & pritchard , 1996 ) .
it is absolutely crucial in this analysis to ascertain the temporal quality of our datasets , and in this way , test for possible glitches in the calibrated timing hardware and potential nonlinearities caused as a result of microchannel plate saturation , thus assessing any unusual temporal activity in the pulsar time series . regarding the problem of @xmath38 saturation it was determined that , within the bounds of variable sky transparancy , no non - poisson behaviour at the 99% level was noted , for a range of datasets over the @xmath0 wavebands ( golden 1999 ) .
this was repeated throughout the light curve , indicating that the emission process is consistent with dominant incoherent synchrotron processes .
giant radio pulses , which are of order 10@xmath39 times more intense than the average radio pulse , occur in approximately 1 % of radio pulse periods ( lundgren et al .
there were no deviations in the photon statistics accumulated in these observations consistent with such giant pulse activity . regarding the temporal stability of the timing system , we previously determined the timing accuracy to within @xmath7 1 @xmath1s of utc , and for long ( 4 - 5 days ) timeseries datasets , it is possible to test the timing against the crab pulsar itself via epoch folding or fourier methods . for the former case , this will be as accurate as the timing ephemeris provided - during our observing period the jodrell ephemeris was accurate to 50 @xmath1s . in the latter ,
we are restricted to the accuracy with which one can determine to have detected a harmonic at high @xmath40 . via epoch folding ,
the best reduced @xmath41 was determined for the given rotational parameters from jodrell within 100 @xmath1s of the given reference time stamp given by the gps system . using an origin2000 supercomputer , detection of the high harmonics ( typically @xmath7 50 ) for several daily @xmath18 datasets
yielded frequency estimates that were accurate to the corrected ephemeris frequency to within 1 in 10@xmath42 .
consequently , we were satisfied with the temporal integrity of these mama datasets taken in russia in january 1996 . in this vein , the unusual low frequency signal of @xmath7 0.016 hz reported by cadez & galicic ( 1996 ) was put to the test .
fft analysis on each night s dataset binned at 1 , 0.1 & 0.2 hz were tested , but no significant signal was apparent within the 0.01 - 1 hz window - or at greater frequencies .
consequently we must remain somewhat guarded as regards the conclusions of cadez & galicic ( 1996 ) .
the definitive discovery of a 60 second precessional signal from an isolated rotating neutron star would have profound implications for fundamental physics and thus demands conclusive and undeniable evidence for its justification .
we must state that we see no such evidence to date .
by the use of developed epoch folding algorithms , it was possible to take the deep summed time series for all three wavebands and re - construct a phase - resolved image for a specific phase region .
figure [ mos_crab ] shows such an _ ad hoc _ mosaic of phase - resolved images , with the phase resolution here 0.05 of phase .
it is apparent that there appears to be residual emission from the crab pulsar during the traditional ` off ' phase ( essentially row 4 in figure 2 ) , as was originally noted by peterson et al .
the fact that we can obtain deep phase - resolved images which have been corrected for telescope drift and wobble with a sufficient field of view enables us to rigorously characterise the time - resolved nature of the pulsar s emission via standard image processing techniques .
there are a number of ways to proceed with this analysis .
principally , we wish to determine 1 ) what the total background is in a given aperture per colour band and 2 ) what are the respective fluxes for phase - resolved components of emission associated with the crab pulsar .
the first step is perhaps the most important as it may then be applied to the standard derived light curves so as to correct for the background per bin and thus allow one to characterise the so - called ` steady ' or ` off ' phase of emission . in golden et al .
( 2000 ) , the image processing techniques were outlined in some considerable detail for this analysis .
it is sufficient to state that the iraf daophot package was used to remove the full crab image ( and thus obtain the background per aperture per colour band in addition to the full crab flux ) , and similarly for the various phase - resolved light curve components , which are localised in terms of phase in table [ eik_comps ] .
this process involved the empirical fitting of a psf to the crab and making the subsequent extraction .
cl phase & region + peak 1 & -0.12 - 0.12 + peak 1 ( leading edge ) & -0.12 - 0.0 + peak 1 ( trailing edge ) & 0.01 - 0.12 + bridge & 0.121 - 0.27 + peak 2 & 0.271 - 0.658 + peak2 ( leading edge ) & 0.271 - 0.409 + peak2 ( trailing edge ) &
0.410 - 0.658 + off phase & 0.75 - 0.825 + the ` off ' phase here was localised in the following manner . correcting for the estimated background above , the ` off ' regions for the various coloured light curves at 11 @xmath1s resolution ( 3000 bins )
were tested via a @xmath41 algorithm . here
, the bin region was iteratively reduced , and tested to see if the observed flux was consistent with that expected from a non - varying source over that phase region , at the 95 % level .
as regards any nebular contribution to this photometric analysis , no significant change was noted between the estimated flux of the ` off phase ' made via extraction using the full cycle psf , and the flux determined by empirically fitting a psf to the ` off phase ' itself .
this indicates that the local nebulosity does not significantly contribute to the determined fluxes in this work .
actual phase - resolved photometry was done by normalizing the total crab flux per colour band to a known integrated photometric reference , and then determining the contribution per phase region based on the fractional fluxes estimated for that particular phase - resolved component .
the extinction corrected @xmath0 ground - based photometry of percival et al .
( 1993 ) was used as the reference data for this purpose table [ eik_anal3 ] displays the tabulated fractional fluxes estimated following the photometric analysis detailed previously within each colour band - the fluxes were derived via renormalizing the fractional flux with respect to the percival et al.s integrated data .
cccc parameters & & waveband & + & u ( mjy ) & b ( mjy ) & v ( mjy ) + peak 1 & 2.0 @xmath3 0.1 & 1.9 @xmath3 0.1 & 1.9 @xmath3 0.1 + peak 2 & 1.1 @xmath3 0.1 & 1.1 @xmath3 0.1 & 1.1 @xmath3 0.1 + bridge & 0.09 @xmath3 0.01 & 0.10 @xmath3 0.01 & 0.11 @xmath3 0.01 + off phase & 0.014 @xmath3 0.002 & 0.017 @xmath3 0.002 & 0.019 @xmath3 0.002 + p1 ( hwhm ) lead & 1.1 @xmath3 0.1 & 1.2 @xmath3 0.1 & 1.1 @xmath3 0.1 + p1 ( hwhm ) trail & 0.8 @xmath3 0.1 & 0.8 @xmath3 0.1 & 0.8 @xmath3 0.1 + p2 ( hwhm ) lead & 0.46 @xmath3 0.03 & 0.49 @xmath3 0.03 & 0.46 @xmath3 0.03 + p2 ( hwhm ) trail & 0.59 @xmath3 0.04 & 0.65 @xmath3 0.04 & 0.65 @xmath3 0.04 + cc dataset & power - law + & @xmath43 @xmath44 + integrated uv / u / b / v / r & 0.11 @xmath3 0.09 + integrated u / b / v & -0.07 @xmath3 0.19 + peak 1 & 0.07 @xmath3 0.19 + peak 2 & -0.06 @xmath3 0.19 + bridge & -0.44 @xmath3 0.19 + off & -0.60 @xmath3 0.37 + peak 1 hwhm ( leading ) & -0.15 @xmath3 0.19 + peak 1 hwhm ( trailing ) & -0.11 @xmath3 0.19 + peak 2 hwhm ( leading ) & -0.04 @xmath3 0.19 + peak 2 hwhm ( trailing ) & -0.18 @xmath3 0.19 + in table [ slopes ] the estimated power - law parameter @xmath2 has been determined via a weighted linear least - squares fit for each individual spectral dataset , with associated errors .
we have re - calculated @xmath2 for the both the full range ( @xmath45 ) & @xmath0 ground - based percival et al .
( 1993 ) datasets to compare with the other power - law fits . included in this table
are power - law fits to the fwhm and hwhm components .
as has been outlined previously , there has not been to date a consistent set of observational datasets that are temporally accurate ( to 1 @xmath1s of utc ) , and this has to some extent compromised previous efforts .
the data taken in january 1996 provides the first such thorough sample , and we consequently first concentrate on its temporal analysis . with an effective estimate of the background obtained for all three colour bands as previously outlined
, one can proceed with a rigourous morphological analysis of the light curves themselves .
this is essentially pursued by use of both least - squares fits to the poisson ` corrupted ' light curve profile , and the use of monte carlo techniques to ascertain the errors on any ideal topological fit to the data . in this way , we can hope to obtain information on a given peak s precise maximum position in terms of phase , the extent of a plateau within the peak s maximum region , and the peak s fwhm and hwhm characteristics .
the basic algorithm fits third order polynomials to the leading and trailing edges of the given peak ( from 30% to 90% maximum intensity ) , and then a sixth order polynomial across the peak cusp ( from 90% to 90% ) .
this latter fit is then used as a template in the following manner .
a monte carlo routine uses the template as a basis from which to ` add ' poisson noise dependent on the template counts per specific bin . having done this , a subsequent sixth order polynomial is fit via least - squares . from this , the phase bin with the maximum intensity is determined , and ` noted ' .
this maximum phase is then used in the calculation of the fwhm and hwhm parameters , using the previously determined fits to the leading and trailing edges of a given peak .
an iterative routine then starts with the two adjacent bins to the nominated maximum bin , computes the local average per bin , and computes the @xmath41 to test for a deviation from this local average i.e. the end of a ` plateau ' region .
the range is extended by a bin on each side , and the process is continued until significant deviations are noted at the 95% and 99% confidence intervals .
finally , within the defined bin regions defined by these confidence ranges , a final routine starts with 2 bins , and sweeps through the ranges , noting the @xmath41 statistic , and incrementing the bin number .
the @xmath49 number of bin sizes that deviate from the required confidence levels are noted .
the entire process is repeated from the initially derived template @xmath50 times . for each chosen parameter ,
the average is calculated , and the standard error on the @xmath50 samples is used to represent 1@xmath46 errors on that mean .
the above outlined approach has its origin in the analytical techniques of ransom , eikenberry and others ( ransom et al . , ( 1994 ) , eikenberry et al .
( 1996 ) ) with the monte carlo derived estimate of a plateau region incorporated .
the completed temporal analysis for the light curves in @xmath0 is tabulated in table [ eik_anal1 ] .
[ obs ] cccc parameter & & waveband & + & u & b & v + & & & + & phase & phase & phase + peak 1 position & 0.002 @xmath3 0.003 & 0.003 @xmath3 0.001 & 0.0024 @xmath3 0.0003 + peak - to - peak separation & 0.407 @xmath3 0.003 & 0.405 @xmath3
0.002 & 0.405 @xmath3 0.002 + peak 1 fwhm & 0.043 @xmath3 0.003 & 0.043 @xmath3 0.001 & 0.045 @xmath3 0.000 + peak 2 fwhm & 0.078 @xmath3 0.003 & 0.086 @xmath3 0.001 & 0.089 @xmath3 0.002 + peak 1 hwhm ( lead ) & 0.028 @xmath3 0.004 & 0.030 @xmath3 0.001 & 0.030 @xmath3 0.001 + peak 1 hwhm ( trail ) & 0.015 @xmath3 0.003 & 0.014 @xmath3 0.001 & 0.015 @xmath3 0.001 + peak 2 hwhm ( lead ) & 0.037 @xmath3 0.002 & 0.038 @xmath3 0.001 & 0.039 @xmath3 0.001 + peak 2 hwhm ( trail ) & 0.041 @xmath3 0.002 & 0.047 @xmath3 0.002 & 0.050 @xmath3 0.002 + for each of the two peaks per colourband , as stated above , the iterative monte carlo routine determines the positions of the ` best fitted ' peak maxima , and collates the difference between the two in phase , as well as recording the position in phase of the first peak .
figure [ p2p_d ] displays the peak - to - peak difference as a function of frequency .
it would seem that there is no clear trend for the peaks phase difference to deviate within this restricted photon energy regime .
based on the errors associated with our data , we do not see any departure from this trend reported elsewhere in the literature that the peak - to - peak phase difference is seen to contract with increasing energy ( e.g. eikenberry et al .
1997 ; ramanamurthy 1994 ) .
similar error concerns restrict any real attempt to definitively state that the @xmath0 main peaks arrive at differing phase intervals . the trend , if any , is that the radio - optical main peak difference in arrival phase increases with photon energy - broadly consistent with percival et al .
( 1993 ) and other authors .
the fwhm of a pulsar s pulsed components as a function of energy have been used for many years as a way of understanding the emission characteristics of these objects , being undoubtedly related to the geometry of the actual emission mechanism .
the expected fwhm behaviour with photon energy is thus a strong function of the model under scrutiny - for example , the standard outer gap formalism predicts a growing fwhm with decreasing photon energy ( eikenberry et al . 1997 ) , and early rather monolithic models envisaging emission occuring from an open cone interpret the morphologies as the collected open ` cones ' of synchtrotron emitting population of electrons . in this latter case
, a similar effect is expected .
our monte carlo results , as tabulated and presented in figure [ fwhm_ph ] for both peaks substantiate this trend - although there seems to be a statistically significant greater gradiant associated with peak 2 compared with peak 1 . from the previous work of eikenberry et al .
( 1997 ) using the hsp dataset , it is impossible to make such a differentiation . in their analysis ,
eikenberry et al . (
1997 ) found statistically significant differences between the leading and trailing half - widths for both peaks , these deviations clearly evident at the lowest photon frequencies ( @xmath48 ) , with the trend at our optical wavebands being an increase in measured phase extent and decreasing photon energy .
figures [ hwhm1_ph ] and [ hwhm2_ph ] display the results following the monte carlo analysis ( tabulated in table [ eik_anal1 ] ) , and there is agreement with these earlier conclusions within the spread of errors .
the leading edges for both peaks display similar functional forms .
resolution of the main peak has been discussed many times in the literature as a diagnostic as regards the basic emission mechanism for principally the crab pulsar , as data with the required s / n is unlikely ( at this stage ) to come from any other object ; but as outlined earlier , crude temporal resolution and not sufficiently rigorous analytical techniques have always dogged this . with the background - subtracted datasets , the same monte carlo code was used to ascertain the extent of non - deviating ` best - fit ' 6th order polynomial across the cusp of the main peak with respect to the local noise ( at the 95% and 99% confidence levels ) .
the results of this approach are tabulated in table [ eik_anal2 ] , and in figure [ plat_times ] we show the estimated plateau duration as a function of photon energy .
[ obs ] cccc plateau parameters & & waveband & + & u & b & v + & & & + & phase & phase & phase + local average , 99@xmath51 & 0.0014 @xmath3 0.0004 & 0.0017 @xmath3 0.0001 & 0.0017 @xmath3 0.0001 + peak intensity ,
99@xmath51 & 0.0013 @xmath3 0.0004 & 0.0016 @xmath3 0.0001 & 0.0017 @xmath3 0.0001 + & @xmath1s & @xmath1s & @xmath1s + local average , 99@xmath51 & 47.5 @xmath3 12.7 & 55.1 @xmath3 4.0 & 55.5 @xmath3 3.0 + peak intensity , 99@xmath51 & 45.0 @xmath3 13.7 & 55.1 @xmath3 0.3 & 55.5 @xmath3 3.0 + within the errors , we can say to first order that the cusp of the main peak is consistent with the presence of a ` plateau ' of emission lasting some @xmath7 50 @xmath1s in exent , which is in agreement with the previous conclusions of komarova et al .
tantalisingly , a possible trend of decreasing plateau exent with increasing photon energy will have to await further datasets ; there do appear to be indications that this may be the case from our data .
figure [ peaks_bridge_fluxes ] shows as a function of photon energy the spectral forms based on the fluxes of each peak , and to the same scale and the emission from the bridge component .
it seems clear that certainly as regards the bulk flux from both peaks , that the functional forms , with @xmath2 @xmath7 0.07 @xmath3 0.19 for peak 1 and @xmath2 @xmath7 -0.06 @xmath3 0.19 for peak 2 , are to first order identical with that reported for the integrated emission , namely a flat power law ( @xmath2 @xmath7 -0.07 for the ground - based @xmath0 , @xmath2 @xmath7 0.11 for the full ground - based @xmath45 percival et al . (
1993 ) dataset ) .
this is not entirely surprising , as the dominant emission from the pulsar is localised in the peaks , and the fact that both are identical in form suggests a common population of electrons / magnetic field environment / lorentz factor .
figure [ peak1_peak2_flxrat ] substantiates this common source , there being no significant difference between the colour bands evident in these datasets .
there appears to be , however , evidence for a somewhat steeper power - law associated with the bridge component of emission ( @xmath2 @xmath7 -0.44 @xmath3 0.19 ) . in figure [ bridge_peaks_flux ] the fractional bridge / peak fluxes are shown as a function of photon energy , both suggesting similar functional forms , consolidating the common proportionality between each flux component .
this suggests to first order a different population of emitting electrons / magnetic field environment / lorentz factor in terms of the bridge component than that for the peaks .
however it should be stressed that any definitive differing power - law forms are only significant at the 2@xmath46 level - in this we are at the mercy of the reference data of percival et al .
( 1993 ) , where the substantial error component originates .
using other reference data of a higher photometric quality and further observations at other frequency ranges ( such as in the @xmath47 ) would constrain these fits significantly . in figures
[ peak1_hwhm_flux ] & [ peak2_hwhm_flux ] the hwhm flux estimates for both peaks 1 and 2 are shown - here the leading and trailing hwhm fluxes are derived for each peak .
it seems clear that the leading & trailing edges for both peaks present spectral forms that are in agreement with each other ( see table [ slopes ] ) , although we are somewhat restricted in our analysis by error considerations . for both cases ,
we see suggestions for power - laws with decreasing exponents as a function of frequency , although both are consistent with an essentially flat functional form .
these results are in agreement with estimates of eikenberry & fazio ( 1997 ) , whose analysis yielded results which incline towards a decrease in flux with increasing photon energy .
it is at @xmath48 wavebands that eikenberry et al .
( 1996 ) observed quite profound changes in the overall spectral forms for the leading & trailing edges of both peaks - certainly in the @xmath0 regime , we appear to be sampling similar synchrotron emitting populations .
putting it succinctly , despite the excellent temporal resolution of our data , our analysis and subsequent conclusions are limited by both the low throughput of the imaging system and the large errors associated with the reference integrated photometry of the crab pulsar .
nevertheless , we have resolved to an unprecendented photometric accuracy the flux fractions associated with the principal regions of the light curve , most especially that of the unpulsed component . within errors ,
the datasets confirm existing trends seen in previous analyses ; the intra - peak separation , peak fwhm / hwhm , and bridge flux as a function of photon energy . in this , we see nothing unusual to most model predictions .
however the following results provoke further scrutiny .
we have outlined in some considerable detail elsewhere the implications of the observed spectral form for the resolved unpulsed flux component ( golden et al .
clearly it is nonthermal , and spectrally similar to that of the bridge of emission . in our previous communication
we argued that the observed unpulsed component of emission has its source in a similar electron population / magnetic field / lorentz factor environment , but stated that an unequivocal link was not evident , and that viewing geometry was likely to play a substantial role in what we ultimately observe .
resolution of the main peak s cusp temporally has long been regarded as a possibly powerful diagnostic as regards determining the local physical conditions where the emission is occuring , which in the optical regime is believed to be as a result of incoherent synchrotron processes . in this analysis
we have obtained rigorous estimates of the extent of this plateau , namely @xmath5 55 @xmath1s with some indication of an increase with wavelength .
it now remains to place this result in some sort of theoretical context .
the models imply emission either occuring from a spatially extended region ( the outer gap of romani & yadigaroglu ( 1995 ) or from a localised cone/`trumpet ' like zone traced by the opening angle above a polar cap , and subsequently modified by the toroidal dipolar field .
thus for the former , the peaks are a consequence of many photons arriving from disparate locations but in phase due to relativistic effects , and in the latter we have a conceptually more appealing core / cone of emission . in both cases ,
the synchrotron emitting electron s are expected to have a common lorentz / local magnetic field , in order to satisfy the observed luminosity , although the intuitive concept of a ` beam ' is problematic for the former @xmath52 . for a temporal extent of 55 @xmath1s at the main peak
, one could argue that this implies that the dominant beaming from the population of synchrotron emitting electrons from some localised area in the magnetosphere has a total angular scale of @xmath5 10@xmath53 radians .
it follows ( sturrock et al .
1975 ) then that if one envisages that the total optical emission comes from these electrons with a lorentz factor @xmath6 and pitch angle @xmath54 , then @xmath55 suggesting a @xmath56 lorentz factor emission regime ( @xmath7 @xmath57 ) .
if the radiation is via normal ( large - angle ) synchrotron emission , then the emitted spectrum peaks at the frequency @xmath58 where @xmath59 is the gyrofrequency given by @xmath60 in the small angle approximation , the spectrum peaks at @xmath61 it is easily shown that for either case , @xmath62 from eikenberry et al . ( 1997 ) , the optical spectrum appears to show a maximum in the vicinity of the h band ( @xmath7 1.8 @xmath63 hz ) or @xmath64 .
applying this with the above suggests that the local magnetic field must be @xmath65 assuming a dipolar @xmath43 @xmath66 law , and the canonical 10 km neutron star possessing a derived surface magnetic field of 3.7 @xmath28 @xmath67 gauss , this corresponds to a radial distance of some 2256 km . in comparsion , the @xmath68
is estimated to be @xmath69 1523 km - thus emission must occur at some 1.5@xmath68 .
this emission is not possible for the ideally orthogonally aligned rotator of smith ( 1978 ) and others , but feasible for those pulsars whose angle @xmath2 between dipole & rotational axes is @xmath33 90@xmath70 .
it is clear that the assumption of restricted radial zones of emission would favour the more localised models over the spatially extended ( where zonally varying local @xmath6 & @xmath18 would contribute to the overall pulse profile ) and that this radial estimate has many intuitive links with the emission model of gil et al .
( 2000 ) . as a final comment
, we note some evidence for the plateau s temporal extent to grow with wavelength , errors notwithstanding .
this trend , if real , could be possibly explained most easily with reference to synchrotron self - absorption .
previous discussions on this topic assumed that the observed roll - over was due to this phenomenon , yet time - resolved observations of the @xmath23 wavebands by several authors ( middleditch et al .
1983 ; eikenberry et al .
1996 , 1997 ) do not neccesarily show evidence for a flattening of the main peak ( however , pennypacker 1981 ; penny 1982 ) .
this effect should be manifested in the @xmath23 wavebands , and it is somewhat difficult to justify at our low frequencies for a tightly localised synchrotron source - but it is possible in a radially or latitudinally extended model , where differing regions of the magnetosphere may experience the effect .
alternatively , we can define the cusp as a simple function of the local synchrotron emission processes . here , the lower energy electrons have a wider opening angle , and thus we observe a wider plateau .
accurate numerical modelling would test this latter hypothesis . from the tabulated data ( table 4 ) , we can say for the higher s / n data of @xmath18 & @xmath20 that the main peak pulses arrive to within 10 @xmath1s of each other , and that , when one considers the peak - to - peak separation , there does not seem to be any clear evidence for variation as a function of photon energy . in this
we are somewhat restricted by the overall poor bin s / n for these datasets .
deeper observations would provide temporally more accurate data , but to first order , it does seem that in these wavebands , we are seeing an essentially constant general behaviour , as regards pulse arrivals . we can not absolutely differentiate the radio / pulse lag time , due to the error on the former ( some 50 @xmath1s ) , but we can place a relative lag of order @xmath7 -60 @xmath1s - that is to say , the @xmath0 emission trails the arrival of the radio main pulse .
this might not be so difficult to understand if the radio emission had its source at some distance further in towards the neutron star surface with respect to the higher energy sources , located in the toroidally warped dipolar field , where both relativistic beaming and the local magnetic field vector might conspire to produce such a delayed arrival time ( assuming localised emission ) .
but this is not the case for the crab , whose radio emission is believed to occur in proximity to the optical sources close to the light cylinder .
observations with the figaro ii @xmath6-ray instrument ( masnou et al .
1994 ) have suggested that the ( 0.15 - 4.0 mev ) @xmath6-ray main peak @xmath71 that of the radio pulse by some 375 @xmath3 148 @xmath1s , which according to the authors , points to spatially different regions of emission between the radio & @xmath6-ray source regions of some 100 km in the magnetosphere .
recent rxte observations ( rots et al .
1998 ) at a superior temporal resolution with respect to utc suggest that the ( 5 - 200 kev ) x - ray main peak leads the radio pulse by 264 @xmath3 330 @xmath1s . for our ` best ' position , the @xmath20 band ,
the main peak is localised at + 79.2 @xmath3 10 @xmath1s with respect to the radio pulse - incorporating the quoted radio error enhances this to @xmath3 60 @xmath1s .
thus to first order , the optical trails the x - ray main peak , and the x - ray trails the @xmath6-ray main peaks , but their absolute location with respect to the radio is not clear , being dependent on the local epoch timing solution .
the discrepancies are undoubtedly as a result of the differing quality of the different datasets ( in terms of s / n ) , the accuracy of the radio ephemeris solution and the manner in which the main peak s centroid is defined . from our folded light curves
, it appears that the main peak in the @xmath0 is located @xmath72 the radio main pulse , but it is important to appreciate that there is a substantial error on the radio timing solution .
what is required are simultaneous deep observations in both the optical & radio so as to guarantee as accurate a timing solution as possible .
such observations have recently been made in collaboration with colleagues from the westerbork radio telescope observatory using the triffid instrument at la palma ; the data is currently under analysis .
realistic numerical modelling is possibly the only full consistent manner in which to reproduce such frequency dependent lag effects and so constrain the position(s ) of emission for a standard dipolar structure under a range of viewing geometries , for both spatially extended & localised model frameworks .
such modelling studies are underway at nui , galway and elsewhere .
definitive optical observations of the crab and other pulsars have to date been restricted - either by poor temporal resolution both in terms of resolution and with respect to absolute utc , and also by the limited spatial information which is crucial in distinguishing the pulsar s emission from the background . using the triffid high speed photometer ,
we have obtained datasets uncompromised by these previous limitations , and in addition to confirming many of the previous conclusions of eikenberry et al .
( 1997 ) , we have determined several new results uniquely due to the technology we have implemented , including the fluxing of the unpulsed component of emission , the absolute arrival times of the three light curves with respect to one another , and the spatial extent of the plateau on the main peak .
our photometric analysis was however limited by the reference data we used to renormalize the flux ratios thus determined .
it is clear that further analysis using more photometrically accurate reference data combined with further observations at extended wavebands , such as the @xmath21 or @xmath73 would yield fluxes that would provide excellent leverage to the least - squares fits and thus produce more significant power - law exponents . in terms of a more rigorous temporal analysis of the light curves , deeper observations in @xmath0
would place limits on the extent of the main peak s plateau , and possibly indicate if it is energy dependent .
furthermore , a greater sample of photons per pulse period may allow us to probe the possible correlation between the radio giant pulses and high energy emission mechanisms .
we note that the incorporation of polarimetric technology into this nascent field of 2-d high speed photometry would yield critical data which would undoubtedly reveal more clues to puzzle over , and thus contribute towards the eventual solution to this three decade old mystery of how exactly pulsars function .
_ eso is thanked for the provision of their mama detector .
we are grateful to the 6-m telescope program committee of the ras for observing time allocation .
we thank the engineers of sao ras , a. maksimov for help in equipment preparation for the observations and the director of sao ras yu .
balega for arranging the observations .
dr ray butler of nui , galway is thanked for his assistance during the course of the photometric analysis .
this work was supported by the russian foundation of fundamental research ( 98 - 02 - 17570 ) , state programme `` astronomy '' , russian ministry of science and technical politics , and the science - educational centre `` cosmion '' , and funded under the intas programme .
the support of enterpise ireland , the irish research and development agency , is very gratefully acknowledged .
finally , we would like to thank the anonymous referee s considerable effort and wise counsel in contributing towards the final version . _ | we report a phase - resolved photometric and morphological analysis of @xmath0 data of the crab pulsar obtained with the 2-d triffid high speed optical photometer mounted on the russian 6 m telescope . by being able to accurately isolate the pulsar from the nebular background at an unprecedented temporal resolution ( 1 @xmath1s )
, the various light curve components were accurately fluxed via phase - resolved photometry . within the @xmath0 range , our datasets are consistent with the existing trends reported elsewhere in the literature .
in terms of flux and phase duration , both the peak full width half maxima and half width half maxima decrease as a function of photon energy . this is similarly the case for the flux associated with the bridge of emission .
power - law fits to the various light curve components are as follows ; @xmath2 = 0.07 @xmath3 0.19 ( peak 1 ) , @xmath2 = -0.06 @xmath3 0.19 ( peak 2 ) and @xmath2 = -0.44 @xmath3 0.19 ( bridge ) - the uncertainty here being dominated by the integrated ccd photometry used to independently reference the triffid data .
temporally , the main peaks are coincident to @xmath4 10 @xmath1s although an accurate phase lag with respect to the radio main peak is compromised by radio timing uncertainties .
the plateau on the crab s main peak was definitively determined to be @xmath5 55 @xmath1s in extent and may decrease as a function of photon energy .
there is no evidence for non - stochastic activity over the light curves or within various phase regions , nor is there evidence of anything akin to the giant pulses noted in the radio .
finally , there is no evidence to support the existence of a reported 60 second modulation suggested to be as a consequence of free precession . |
compared to the case of pions the electromagnetic productions of kaons are less understood .
this is due to the fact that the strangeness quantum number is explicitly present in the final state of the process .
although the reactions involving kaons are slightly more complicated , the additional degrees of freedom created by the strangeness can give information not available from nucleons and pions .
for instance , due to the conservation of the strangeness , the production of kaons is always accompanied by the creation of hyperons , which can be used to deeply explore the structure of the nucleus since they are not blocked by the pauli principle . on the other hand
, this elementary process provides an important input for the calculation of hypernuleus photo- or electroproduction . along with the fact that some @xmath0 resonances that were predicted in quark models have only noticeable branching ratios into the @xmath1 channel ,
kaon photo- and electroproduction clearly have drawn many attentions for more than five decades .
the earliest attempt to theoretically explain kaon photo- and electroproduction was proposed by kawaguchi and moravcsik more than 50 years ago@xcite .
interestingly , all of the six possible isospin channels were already considered by utilizing only three feynman diagrams of the born terms , though hitherto no experimental data were available .
the results were obviously very modest and , in fact , the cross sections for the @xmath2 and @xmath3 channels were predicted to be zero since neither nucleon anomalous magnetic moments nor resonance contributions were taken into account . since that time considerable efforts were devoted to explain the appearing experimental data , although only thom who tried to seriously fit the data to an isobar model .
after that , there were a number of works using isobar model@xcite , dispersion relation@xcite , multipoles analysis@xcite , and regge approach@xcite .
after the mid - seventies , the interest in this field was temporarily dormant , mainly due to the lack of experimental facilities .
the interest in kaon photo- and electroproduction was revived by the constructions of modern accelerators such as those in jlab , mami , elsa , and others laboratories .
the work of adelseck , bennhold , and wright started the new era of phenomenological models in @xmath4 photoproduction@xcite .
this work was refined with the inclusion of more data and electroproduction process@xcite and extended to all isospin channels@xcite .
a chiral quark model with less parameters has been also put forward@xcite in an attempt to recover the low energy theorem .
the results of this model are modest and , in fact , the @xmath5 cross section is predicted to be larger than that for @xmath6 by a factor of approximately two , since contributions from the seagull and @xmath7-channel resonance have opposite signs .
modern calculations of kaon photoproduction exploit chiral perturbation theory@xcite and coupled channels analysis@xcite . in the higher energy regions ,
the regge@xcite or hybrid models@xcite ( a combination of the feynman diagrammatic technique and regge formalism ) turn out to be more appropriate .
nevertheless , for practical use such as for nuclear applications and other phenomenological studies , the single channel isobaric analysis is still proven to be powerful@xcite .
historically , experimental data of kaon photo- and electroproduction can be divided into two categories , old data ( published before 1980 ) and new data ( published after 1990 ) .
a list of references for old data is given , e.g. , in refs.@xcite , whereas the corresponding kinematical coverage for the photoproduction is shown in the left panel of fig .
[ fig : data ] .
surprisingly , the old data cover also forward angle regions , the case which is very difficult to achieve with the presently available technologies , as obviously shown in the right panel of fig .
[ fig : data ] . as discussed in the following section ,
this region is in fact very decisive for constraining the models and for nuclear applications .
nevertheless , the data quality is in general poor and , as a consequence , the cross sections data do not show resonance structures .
there were few data available for the recoiled @xmath8 polarization and two data points on the target asymmetry .
kaon photoproduction started to become more interesting after the publication of saphir data@xcite on the @xmath4 and @xmath6 channels in 1998 , since the corresponding error bars allow for an identification of resonance structures , especially in the @xmath4 channels .
further analysis of saphir data@xcite as well as the new clas data@xcite confirmed these structures .
note that in both saphir and clas data sets , there are plenty of data on the recoil polarization . together with the new data from the leps@xcite and graal@xcite collaborations , these data provide a strong constraint on the proliferation of phenomenological models . in the @xmath9 channels ,
new data have also appeared from saphir collaboration@xcite ( @xmath5 ) , and from the leps collaboration@xcite ( @xmath10 ) . the ability to reverse
the direction of photon helicity in the spring8 has led to the measurements of the photon asymmetry @xmath11 in the @xmath4 , @xmath6 , as well as @xmath10 channels@xcite .
since the polarization of the recoiled @xmath8 can be obtained without any additional apparatus , measurements of the double polarization observables @xmath12 and @xmath13 ( or @xmath14 and @xmath15 ) are possible , provided that the polarized photon beams are available .
the first measurement of these observables was performed by the clas collaboration@xcite and because the available beams at jlab are circularly polarized , the corresponding observables are @xmath12 and @xmath13 .
very recently , by utilizing the linearly polarized photon the graal collaboration has been able to measure the @xmath14 and @xmath15 observables@xcite from threshold up to @xmath16 gev .
the electroproduction experiments with high statistics have been also performed at jlab@xcite .
it is reported that the longitudinal and transverse components of the cross section can be nicely separated@xcite , whereas the transferred polarization in the @xmath17 process has been measured@xcite .
in spite of their unprecedented high statistics , the new clas@xcite and saphir@xcite data reveal a lack of mutual consistency in the forward and backward regions .
this problem can be summarized by the total cross section shown in fig .
[ fig : tot ] , where it is obvious that fitting a phenomenological model to both data sets simultaneously results in a model which is inconsistent to all data sets . as shown by ref.@xcite
this problem also hinders a reliable extraction of the resonance parameters , since the use of saphir and clas data , individually or simultaneously , leads to quite different resonance parameters .
therefore , the presently available data do not allow for a firm conclusion on the extracted `` missing resonances '' . by using four different isobar models ref.@xcite studied the statistical properties of both clas and saphir data .
it is found that the saphir data are coherently shifted down with respect to the clas and leps@xcite data , especially at forward kaon angles . a global scaling factor of 15%
is required to remove this discrepancy .
the phenomenon also implies that the leps data are more consistent with the clas data than the saphir ones .
interestingly , at forward angles both clas and leps data sets can be described reasonably well within the isobaric model without hadronic form factors .
this can be understood from the energy distribution of the cross sections at this kinematics , as will be discussed in the following subsections .
the statistical study recommend that a combination of clas , leps , and old for the definition of old data ] data sets in the fitting data base is the more preferred choice , while a combination of clas and saphir data should not be assumed in fixing parameters of models , especially at forward angles . despite considerable efforts to alleviate this problem ,
it is still difficult to determine which data set should be used to obtain a reliable phenomenological model .
the reason is that in all analyses the experimental data are fitted by adjusting a set of free parameters , while the precise values of these parameters are not well known .
furthermore , the extracted parameters are not unique and also sensitive to the number of resonances used in a model . in view of this , ref.@xcite proposed to use other quantities which can be predicted by the models and can be directly compared with the results from other measurements or model predictions .
one of the possible quantities is the contribution of the @xmath18 channel to the gerasimov - drell - hearn ( gdh ) sum rule @xcite , which relates the anomalous magnetic moment of the proton @xmath19 to the total photoabsorption cross sections @xmath20 ,
i.e. @xmath21 . by looking at the energy distributions of the @xmath20 obtained from models that fit to clas and saphir data shown in the left panel of fig .
[ fig : sigtt ] , it is obvious that the two data sets lead to different contribution to @xmath19 .
it is then concluded that contribution from the model that fits to the saphir data sets is more preferred@xcite .
nevertheless , experimental measurement of the @xmath20 is clearly required to confirm this claim .
since the saphir detector has been dismantled , measurements of the @xmath20 by the clas or mami collaboration seem to be the only choice .
the beam - recoil polarization observables data , @xmath12 and @xmath13 , published by the clas collaboration@xcite recently , indicate that the @xmath8 polarization is predominantly in the direction of the spin of the incoming photon .
it is interesting that recent analyses found that these data seems to be difficult to explain . in ref .
@xcite it is reported that the inclusion of these data reduces the complicated structure of the total cross sections @xmath20 as shown in the right panel of fig .
[ fig : sigtt ] , which indicates that the @xmath12 and @xmath13 data select only certain important resonances , i.e. the @xmath22 , @xmath23 , @xmath24 , and @xmath25 resonances .
this finding corroborates the result of a recent coupled channels analysis .
@xcite the first version of saphir data on @xmath26 released in 1994 did not give any hint about resonance structures in the cross section , since the statistics was very limited@xcite .
however , by collecting more statistics the second version of saphir data@xcite released in 1998 display two distinct peaks in the total cross section at @xmath27 mev and 1900 mev .
more accurate data published by the saphir@xcite and clas@xcite collaborations 5 and 8 years later , respectively , confirm the existence of the two peaks ( see fig.[fig : tot ] ) . by comparing the extracted resonance widths and branching fractions with the predictions of a constituent quark model@xcite , ref.@xcite concluded that the second peak corresponds to the @xmath28 nucleon resonance that is missing from the pdg list@xcite .
although many subsequent analyses used this spin 3/2 nucleon resonance in their models , there was a suggestion@xcite that this peak could come from other resonances such as @xmath29 , @xmath30 or @xmath31 .
however , the possibility of using these resonances , instead of the @xmath28 , has been also discussed in ref.@xcite , but it was ruled out after comparing the extracted widths and branching fractions with the predictions of quark model .
further study using a multipoles approach@xcite found that , in spite of their differences , both saphir@xcite and clas@xcite data sets indicate that the peak originates from the @xmath32 resonance with a mass between @xmath33 mev .
the necessity of the @xmath32 in the analyses of @xmath4 photoproduction with a mass of around 1900 mev has been also corroborated by recent coupled channels analyses@xcite .
note that in these analyses the recoil polarization data have been also taken into account .
it is quite interesting , however , that a recent effective field theory analysis suggests that the peak should correspond to the @xmath30 state which is a bound state of @xmath34 with a mixture of @xmath35 and @xmath36 components@xcite .
to confirm this idea , two experiments are proposed , i.e. first to exclude contributions of the spin 3/2 states and second to find an indication of the @xmath34 bound state . on the other hand
, a preliminary study@xcite performed by constraining the free parameters in a multipole analysis@xcite to the pdg values indicates that the peak originates from the contributions of both @xmath29 and @xmath30 states with masses around 1920 mev . in this study
it is found that the contribution of the @xmath29 state is stronger than that of the @xmath30 and the result does not depend on which data set is being used in the fitting database . since the dominant contributions to the total cross section shown in fig .
[ fig : tot ] come from the forward angles data , the largest discrepancies between the clas and saphir data obviously show up at the forward angles .
unfortunately , the available phenomenological models vary wildly at these kinematical regions .
this situation is clearly shown in fig .
[ fig : forward ] , in which we can see that for @xmath37 our best knowledge on kaon photoproduction can not tell us about the actual values of differential cross sections .
meanwhile , theoretical predictions of the hypernuclear photo- or electroproduction rely heavily on the elementary operator extracted from kaon photo- or electroproduction data at very forward angles . as an example , in the electromagnetic production of the hypertriton , the magnitude of differential cross section is only realistic for experimental measurement at @xmath38 , whereas at @xmath39 the cross section is practically zero@xcite . in view of this ,
accurate and reliable experimental data as well as theoretical formulation of kaon photoproduction in the forward region are the first condition toward a reliable prediction of the hypernuclear photoproduction cross section .
it has been well known that the background terms become divergent at high energies .
therefore , the use of hadronic form factors to reduce the background contribution is desired .
it is also known that the use of this form factor leads to the violation of gauge invariance of the amplitude . as a result
, several methods have been put forward in the last decades to overcome this problem .
two of the well known prescriptions are due to ohta @xcite and haberzettl @xcite . as can be seen from the left panel of fig .
[ fig : had_ff ] , the latter is superior since it can provide a reasonable description of experimental data , i.e. producing relatively much smaller @xmath40 , using values for the leading couplings constants close to the su(3 ) values . in ohta s method
this is not possible due to the absence of a hadronic form factor in the electric current contribution .
since then , most of the phenomenological models include hadronic form factors in their background terms . on the other hand ,
ref.@xcite proposed to use certain hyperon resonances , the @xmath41 and @xmath42 , for counterbalancing the strength of the born terms .
this is achieved through a destructive interference between these @xmath43-channel resonances and the born terms , thereby reducing the strength to a realistic level .
however , the method is not so convincing for the huge number of presently available experimental data and , on the other hand , the use of hadronic form factor is more flexible and simultaneously takes into account the fact that nucleons are composite objects , and not point - like .
recently , ref.@xcite has demonstrated that the behavior of forward - angle cross sections as a function of the energy for models that include the hadronic form factors is far from realistic .
this finding is depicted in the right panel of fig .
[ fig : had_ff ] , where the hadronic form factors drastically suppress the cross sections at @xmath44 gev .
although no data are available for comparison at this kinematics , such a strong damping has not been observed in the experimental data near the forward angles . as can be seen in figs . 4 and 6 of ref.@xcite differential cross sections for
@xmath45 tend to be flat up to @xmath46 gev .
therefore , the concept of hadronic form factors in meson photoproduction needs to be revisited in the future . by considering the conservation of quantum numbers ,
there are six possible reaction channels for kaon photoproduction on the nucleon , three on the proton and three on the neutron . except for the @xmath18 channel , these possible reactions
are shown in fig .
[ fig : iso ] , where the calculations obtained from kaon - maid are compared with the available data .
note that there are few data points available just recently for the @xmath47 differential cross section released by the leps collaboration@xcite .
thus , there are no data for the neutral kaon photoproduction on the neutron , @xmath48 and @xmath49 , although from fig .
[ fig : iso ] it is clear that the data can severely constrain the model , especially in the @xmath48 channel .
furthermore , the investigation of the @xmath50 potentials through the @xmath51 reaction requires the information from these two channels@xcite .
we note that important progress in this direction has been indicated by the newly published @xmath52 data by the tohoku s group@xcite .
it is also important to note that the @xmath53 photoproduction excludes the @xmath54-channel in the born terms .
thus , by comparing photoproductions of @xmath55 and @xmath53 , one can study the influence of the kaon propagator in kaon photoproduction .
it is widely known that the electroproduction process can give more information not available from the photoproduction one .
one of the important information is the electromagnetic structure ( form factor ) of the kaon , which is hiding in the longitudinal cross section ( @xmath56 ) .
in contrast to the pion case , where the mass of pion is much smaller and therefore the @xmath54-channel can dominate the process , so that an independent extraction is possible@xcite , the extraction of kaon electromagnetic form factor requires a reliable phenomenological model .
consequently , accurate measurements of @xmath56 are required for this purpose . however , as seen from the lower panels of fig .
[ fig : electro ] , the presently available data still do not allow for the extraction of this form factor . furthermore , to suppress contaminations from nucleon resonances , measurements with @xmath57 gev are recommended .
the recent progress and some important issues in photo- and electroproduction of kaon on the nucleon have been briefly presented .
it is quite apparent that more experimental and theoretical works are needed to settle the present problems in this field .
nevertheless , with the modern concept of accelerator and spectrometer technologies available at jlab , mami , elsa , spring8 , and other laboratories , we are quite optimistic that these problems can be solved in the near future .
the author acknowledges the support from the universitas indonesia .
9 m. kawaguchi and m. j. moravcsik , phys . rev . * 107 * , 563 ( 1957 ) .
b. b. deo and a. k. bisoi , _ phys .
d _ * 9 * , 288 ( 1974 ) .
n. f. nelipa , nucl . phys . * 82 * , 680 ( 1966 ) .
a. r. pickering , _ nucl .
phys . _ * b66 * , 493 ( 1973 ) . w. schorsch , j. tietge , and w. weilnbck , _ nucl . phys . _ * b25 * , 179 ( 1970 )
. n. levy , w. majerotto , and b. j. read , _ nucl .
phys . _ * b55 * , 493 ( 1973);*b55 * , 513 ( 1973 ) .
r. a. adelseck , c. bennhold , and l. e. wright , _ phys .
c _ * 32 * , 1681 ( 1985 ) .
r. a. adelseck and b. saghai , _ phys .
c _ * 42 * , 108 ( 1990 ) ; j .- c .
david , c. fayard , g. h. lamot , and b. saghai , _ phys .
c _ * 53 * , 2613 ( 1996 ) .
t. mart , c. bennhold , and c. e. hyde - wright , _ phys .
c _ * 51 * , r1074 ( 1995 ) ; t. mart and c. bennhold , _ nucl .
phys . a _ * 585 * , 369c ( 1995 ) .
zhenping li , phys .
c * 52 * , 1648 ( 1995 ) .
s. steininger and u. g. meissner , _ phys .
b _ * 391 * , 446 ( 1997 ) ; b. borasoy , p. c. bruns , u. g. meissner and r. nissler , _ eur .
phys . j. a _ * 34 * , 161 ( 2007 ) .
n. kaiser , t. waas and w. weise , _ nucl .
a _ * 612 * , 297 ( 1997 ) ; t. feuster and u. mosel , _ phys .
c _ * 59 * , 460 ( 1999 ) .
a. v. anisovich _ et al .
_ , _ eur .
j. a _ * 25 * , 427 ( 2005 ) ; b. julia - diaz , b. saghai , t. s. lee and f. tabakin , _ phys .
c _ * 73 * , 055204 ( 2006 ) .
m. guidal , j. m. laget and m. vanderhaeghen , _ nucl .
phys . a _ * 627 * , 645 ( 1997 ) .
t. corthals , j. ryckebusch and t. van cauteren , _ phys .
c _ * 73 * , 045207 ( 2006 ) , t. mart and t. wijaya , _ acta phys .
b _ * 34 * , 2651 ( 2003 ) .
t. mart and c. bennhold , _ phys .
c _ * 61 * , 012201 ( 1999 ) .
f. x. lee , t. mart , c. bennhold and l. e. wright , _ nucl .
phys . a _ * 695 * , 237 ( 2001 ) ; b. s. han , m. k. cheoun , k. s. kim and i. t. cheon , _ nucl .
a _ * 691 * , 713 ( 2001 ) ; a. de la puente , o. v. maxwell and b. a. raue , arxiv:0809.3805 [ nucl - th ] .
adelseck and b. saghai , _ phys .
c _ * 42 * , 108 ( 1990 ) .
m. q. tran _
et al_. , _ phys .
b _ * 445 * , 20 ( 1998 ) .
k. h. glander _ et al .
_ , _ eur .
j. a _ * 19 * , 251 ( 2004 ) r. bradford _
_ , _ phys .
c _ * 73 * , 035202 ( 2006 ) m. sumihama _ et al .
_ , _ phys .
c _ * 73 * , 035214 ( 2006 ) k. hicks _ et al .
_ , _ phys .
c _ * 76 * , 042201 ( 2007 ) .
a. lleres _ et al .
_ , _ eur .
j. a _ * 31 * , 79 ( 2007 ) r. lawall _ et al . _ , _ eur .
j. a _ * 24 * , 275 ( 2005 ) h. kohri _ et al .
_ , _ phys . rev .
lett . _ * 97 * , 082003 ( 2006 ) r. bradford _
_ , _ phys .
c _ * 75 * , 035205 ( 2007 ) .
a. lleres _ et al .
_ , eur .
j. a * 39 * , 149 ( 2009 ) r. m. mohring _ et al . _ , _ phys .
c _ * 67 * , 055205 ( 2003 ) d. s. carman _ et al .
_ , _ phys . rev
. lett . _ * 90 * , 131804 ( 2003 ) p. ambrozewicz _
_ , _ phys .
c _ * 75 * , 045203 ( 2007 ) t. mart and a. sulaksono , _ phys .
c _ * 74 * , 055203 ( 2006 ) p. bydovsk and t. mart , _ phys .
c _ * 76 * , 065202 ( 2007 ) .
t. mart , _ few body syst .
_ * 42 * , 125 ( 2008 ) . s.b .
gerasimov , _ sov .
_ * 2 * , 430 ( 1966 ) ; s.d .
drell and a.c .
hearn , _ phys .
* 16 * , 908 ( 1966 ) .
a. v. anisovich _
_ , _ eur .
j. a _ * 34 * , 243 ( 2007 ) m. bockhorst et al . , _
c _ * 63 * , 37 ( 1994 ) .
s. capstick and w. roberts , _ phys . rev .
d _ * 49 * , 4570 ( 1994 ) ; * 58 * , 074011 ( 1998 ) . c. caso _
et al_. , _ eur .
j. c _ * 3 * , 1 ( 1998 ) .
s. janssen _ et al .
_ , _ eur .
j. a _ * 11 * , 105 ( 2001 ) a. martinez torres , k. p. khemchandani , u. g. meissner and e. oset , arxiv:0902.3633 [ nucl - th ] ( 2009 ) .
r. nelson and t. mart , arxiv:0904.3598 [ nucl - th ] , to be published .
t. mart and b. i. s. van der ventel , _ phys .
c _ * 78 * , 014004 ( 2008 ) ; t. mart , _ nucl .
a _ * 815 * , 18 ( 2009 ) . k. ohta , _ phys .
c _ * 40 * , 1335 ( 1989 ) .
h. haberzettl , _ phys .
c _ * 56 * , 2041 ( 1997 ) .
h. haberzettl , c. bennhold , t. mart , and t. feuster , _ phys . rev .
c _ * 58 * , r40 ( 1998 ) .
t. mart , c. bennhold , h. haberzettl , and l. tiator , kaon - maid , available at http://www.kph.uni-mainz.de/maid/kaon/kaonmaid.html .
_ , _ phys .
c _ * 74 * , 044004 ( 2006 ) .
k. tsukada _
_ , _ phys .
c _ * 78 * , 014001 ( 2008 ) .
t. mart , _ mod .
lett . a _ * 23 * , 3317 ( 2008 ) t. mart and a. sulaksono , arxiv : nucl - th/0701007 ( 2007 ) . | the present status of kaon photo- and electroproduction on the nucleon is briefly reviewed .
some current important issues in this field are discussed . |
the @xmath0 topology on the skorokhod space @xmath2)$ ] of cdlg functions has emerged as a result of chain of observations made in eighties and nineties of the twentieth century . in 1984
meyer and zheng @xcite considered certain conditions on truncated variations of stochastic processes and proved that these conditions give uniform tightness of the processes in some topology ( nowadays called the meyer - zheng topology ) on @xmath3 .
a year later stricker @xcite proved that it is possible to relax meyer - zheng conditions to uniform tightness of random variables @xmath4 and @xmath5 , for each pair of levels @xmath6 ( where @xmath7 is the number of up - crossings of levels @xmath8 ) .
stricker still operated with the meyer - zheng topology .
it was clear for kurtz @xcite that such conditions give much more .
but an _ ad hoc _ device constructed by kurtz _ did not have a topological character_. the @xmath0 topology was constructed by the author in @xcite .
this step was final in the sense that stricker s conditions are _ equivalent _ to the uniform tightness in the @xmath0 topology
. it should be emphasized , that the @xmath0 topology , as considered in @xcite , is sequential and it is still not known if it is completely regular .
so at the moment of its creation there was no formalism to deal with it within the probability theory .
such formalism has been provided in @xcite . and an efficient tool - the almost sure skorokhod s representation in non - metric spaces , given in @xcite - made the @xmath0 topology an operational and useful device in many problems .
the very first application of the @xmath0 topology was given by the author in @xcite , where convergence of stochastic integrals was considered .
later the @xmath0 topology was used in problems related to homogenization of stochastic differential equations ( e.g. @xcite , @xcite , @xcite , @xcite ) , @xcite , @xcite ) , diffusion approximation of solutions to the poisson equation ( @xcite ) , skorokhod problem ( @xcite , @xcite , @xcite ) , econometrics ( @xcite ) , control theory ( @xcite , @xcite ) , linear models with heavy - tails ( @xcite ) , continuity of semilinear neumann - dirichlet problems ( @xcite ) , generalized doob - meyer decomposition ( @xcite ) , modeling stochastic reaction networks ( @xcite ) and even in some considerations of more general character ( @xcite , @xcite ) . in the present paper
we provide some complementary information related to the @xmath0 topology .
section 2 restates the definition and basic properties of the @xmath0 topology .
the @xmath0 topology was defined in @xcite by means of so - called @xmath9-convergence @xmath10 , which leads to the topological convergence @xmath11 via the kantorovich - vulih - pinsker - kisynski recipe ( we refer to section [ appendix ] for a primer on sequential spaces ) . in section [ seccompact ]
we provide a compact formula for the @xmath11 convergence .
in section [ sechier ] we find a position for @xmath0 in the hierarchy of topologies on @xmath3 , by showing that it is _ essentially _ finer than any linear topology which is coarser than the skorokhod s @xmath1 topology .
all results in this section are taken from @xcite .
we need some notation first .
1 . @xmath2)$ ] denotes the skorokhod space , i.e. a family of functions @xmath12
\to { { \mathbb{r}}}^1 $ ] , which are right - continuous at every @xmath13 and admit left - limits at every @xmath14 $ ] .
@xmath15 is the _ sup - norm _ on @xmath3 .
3 . for @xmath8
, @xmath7 is _ the number of up - crossings of levels @xmath16 and @xmath17_. 4 . for @xmath18
, @xmath19 is _ the number of @xmath20-oscillations _ on @xmath21 $ ] .
@xmath22 is _ the total variation of @xmath23 on @xmath21 $ ] _ :
@xmath24 where the supremum is taken over all @xmath25
@xmath26 .
[ l : criteria ] let @xmath27 .
assume that @xmath28 then the following conditions are equivalent : @xmath29 @xmath30 moreover , any of the pairs of conditions ( [ 2e1])+([2e2 ] ) and ( [ 2e1])+([2e3 ] ) is equivalent to the following statement : for every @xmath31 and every @xmath32 there exists @xmath33 such that @xmath34 and @xmath35 we shall write @xmath36 if for every @xmath31 one can find elements @xmath37 , @xmath38 which are @xmath39-uniformly close to @xmath40 s and weakly-@xmath41 convergent : @xmath42 recall , that @xmath43 means that @xmath44 } f(t ) \,dv_n \to \int_{[0,t ] } f(t)\,dv_0(t),\ ] ] for each continuous function @xmath45 \to { { \mathbb{r}}}^1 $ ] .
( criterion of relative @xmath0-compactness ) [ t : criteria ] let @xmath46 .
we can find in every sequence @xmath47 of elements of @xmath48 a subsequence @xmath49 such that @xmath50 , as @xmath51 , if , and only if , one of the equivalent sets of conditions mentioned in the previous lemma is satisfied .
the @xmath0 topology is the sequential topology on @xmath3 generated by the @xmath9-convergence @xmath10 .
`` @xmath10 '' is not an @xmath52-convergence . in the next section
we give a compact form of @xmath11 " , obtained from @xmath10 by the kvpk recipe ( theorem [ kvpk_recipe ] ) .
let @xmath53 be a family of _ continuous _ functions of _ finite variation _ ( @xmath54 ) \cap { \mathds{v}}$ ] ) , satisfying @xmath55 .
let @xmath56 .
we will say that @xmath57 , if @xmath58 } |a_n(t ) - a_0(t)| \to 0,\ ] ] and @xmath59 this is a `` mixed topology '' on @xmath60 ) \cap { \mathds{v}}$ ] .
@xmath61 if , and only if , @xmath62 and @xmath63 for each sequence @xmath64 .
suppose that @xmath61 .
then @xmath62 is a direct consequence of the definition of @xmath10 .
now let @xmath64 .
then ( [ eq : int ] ) is a very particular ( deterministic ) case of theorem 1 ( or theorem 5 ) in @xcite .
so let us assume that @xmath62 and ( [ eq : int ] ) holds for every sequence @xmath64 .
we claim that it is enough to establish relative @xmath0-compactness of @xmath47 . indeed , then in every subsequence @xmath65
we can find a further subsequence @xmath66 such that @xmath67 , for some @xmath68 .
take a function @xmath69)$ ] and define @xmath70
. then @xmath71 and by the first part of the proof , @xmath72 hence for each integrable @xmath73 we have @xmath74 and , consequently , @xmath75 almost surely . since they are cdlg functions , @xmath75 on @xmath76 . and @xmath77 by assumption .
we have proved that in every subsequence @xmath65 we can find a further subsequence @xmath66 along which @xmath78 .
hence , by the kvpk recipe , @xmath61 . in order to prove relative @xmath0-compactness of @xmath47
it is necessary to adjust integrands @xmath79 in a way suitable for the particular functional determining the relative @xmath0-compactness via lemma [ l : criteria ] .
for example , consider condition ( [ 2e1 ] ) .
suppose that @xmath80 then there exists a subsequence @xmath81 and numbers @xmath82 such that @xmath83 . without loss of generality
we may assume that @xmath84 . by the right - continuity of @xmath85 we can find numbers @xmath86 such that @xmath87 and @xmath88.\ ] ] let @xmath89 .
consider function @xmath90}$ ] and the corresponding function @xmath91 .
we have @xmath92 while @xmath93
it follows that ( [ eq : int ] ) can not be satisfied . in a similar way we deal with condition ( [ 2e3 ] ) .
we leave this simple , but arduous , reasoning to the reader .
@xmath61 if , and only if , @xmath62 and for each relatively @xmath94-compact set @xmath95 @xmath96 clearly , ( [ eq : unif ] ) implies ( [ eq : int ] ) . to prove the converse , assume ( [ eq : int ] ) and suppose
that ( [ eq : unif ] ) does not hold for some relatively @xmath94-compact set @xmath95 .
this means that for some @xmath18 there exists a subsequence @xmath65 and elements @xmath97 of @xmath98 such that for all @xmath99 @xmath100 passing to a @xmath94-convergent subsequence @xmath101 we obtain a contradiction with ( [ eq : int ] ) .
denote by @xmath102 the locally convex topology on @xmath3 given by the seminorm @xmath103 and the seminorms @xmath104 where @xmath98 runs over relatively @xmath94-compact subsets of @xmath53 .
then @xmath61 if , and only if , @xmath105 and so @xmath106 , for @xmath0 is sequential .
is it true that @xmath107 ?
positive answer would allow to state that @xmath108 is a locally convex linear topological space .
let us begin with listing some facts on topologies on @xmath3 . 1 .
@xmath3 with norm @xmath109 is a banach space , but non - separable . 2 .
the skorokhod s @xmath1 topology is metric separable and @xmath110 is topologically complete . for definition and properties of @xmath1 we refer to billingsley s classic book @xcite rather than to its second edition .
it is easy to show that @xmath111 implies @xmath112 , hence the @xmath0 topology is coarser than @xmath1 .
it was shown in @xcite that the @xmath0 topology is coarser than skorokhod s @xmath113 topology ( see @xcite for definitions of four skorokhod s topologies ) .
@xmath0 is incomparable with skorokhod s @xmath114 topology !
@xmath110 _ is not a linear topological space _
, for addition is not sequentially @xmath1-continuous , as figure 1 shows .
7 . on the contrary ,
the sequence @xmath115 defined in figure 1 is @xmath0-convergent to @xmath116 and exhibits a typical for @xmath0 phenomenon of self - cancelling oscillations .
addition is _ sequentially continuous _ in @xmath0 !
we do not know , whether addition is _ continuos _ , as a function on the product @xmath117 with _ product topology _ @xmath118 ( in general sequential continuity does not imply continuity ) .
therefore we do not know , whether @xmath108 is a linear topological space .
[ rys1 ] is not linear , width=377 ] [ finest ] suppose @xmath119 is a topology on @xmath2)$ ] which satisfies the following assumptions .
@xmath120}\,;\ , |a| \leq a , u \in [ 0,t]\}$ is relatively $ \sigma$-compact . } \label{th4:c}\end{aligned}\ ] ] then @xmath119 is coarser than the @xmath0 topology .
we claim that it is enough to prove that @xmath112 implies @xmath121 along some subsequence @xmath65 . indeed
, this implies that any @xmath122-closed set is also @xmath0-closed ( for @xmath0 is sequential ) , and so @xmath122 is coarser than @xmath0 .
since we always have @xmath123 , our claim follows .
so let us assume that @xmath112 . for @xmath124 and @xmath125 let us define @xmath126 ( where by convention @xmath127 ) and let @xmath128 , k=0,1,2,\ldots .\ ] ] then by the very definition @xmath129 similarly , if we set @xmath130 then @xmath131 and by theorem [ t : criteria ] and lemma [ l : criteria ] we have @xmath132 since @xmath133 varies only through @xmath134 jumps , it can be represented as a sum of @xmath135 terms : @xmath136},\ ] ] where @xmath137 and @xmath138 by ( [ eq : aa ] ) and ( [ eq : em ] ) we obtain @xmath139 it follows that the sequence @xmath140 lives in the algebraic sum @xmath141 where @xmath142}\,;\ , |a| \leq \big(m_{\varepsilon } + 1\big ) a_{\varepsilon } , u \in [ 0,t]\}$ ] is relatively @xmath119-compact . since @xmath143 is linear @xmath144 is relatively @xmath119-compact as well .
this means that in every subsequence @xmath65 one can find a further subsequence @xmath66 such that @xmath145 , for some @xmath146 .
but we can say more : by the special form ( [ form ] ) of elements @xmath147 ( bounded number of jumps with bounded amplitudes ) we may extract a further subsequence @xmath148 such that @xmath149 .
now choose @xmath150 and apply the diagonal procedure to extract a subsequence @xmath99 such that for each @xmath151 we have along @xmath65 @xmath152 for some @xmath153 .
notice that @xmath154 implies @xmath155 . by corollary 2.10 in @xcite @xmath156 for later purposes we may write @xmath157 , where @xmath158 for @xmath159 .
our final task consists in proving @xmath121 .
let @xmath160 be a @xmath119-open neighborhood of @xmath161 . by the linearity there exists a @xmath119-open neighborhood @xmath162 of @xmath116 such that @xmath163 .
since @xmath119 is coarser than the uniform topology , there exists @xmath164 such that @xmath165 .
let @xmath166 be such that @xmath167 . then for @xmath99
large enough we have @xmath168 let us consider the space @xmath169 ) = l^p\big([0,t ] , { { \mathcal l}}|_{[0,t ] } , \ell|_{[0,t]}\big)$ ] , @xmath170 $ ] of lebesgue - measurable functions on @xmath21 $ ] ( here @xmath171 stands for the lebesgue measure ) .
of course , for each @xmath172 we have @xmath173 ) \subset l^p([0,t])$ ] . moreover , the induced metric converts @xmath173)$ ] into a normed space ( if @xmath174 ) or a metric linear space ( if @xmath175 $ ] ) . clearly , assumptions ( [ th4:a ] ) - ( [ th4:c ] ) are satisfied and by our theorem [ finest ] all mentioned metric topologies are coarser than @xmath0 .
if we replace @xmath171 with another _ atomless _ finite measure @xmath176 on @xmath21 $ ] , then again the metric topologies induced by spaces @xmath169 , \mu)$ ] , @xmath177 , are coarser than the @xmath0 topology .
this is not so , if we admit atoms for @xmath176 , for ( [ th4:c ] ) is then violated .
in introduction we suggested that the @xmath0 topology is _ almost _ finer than any linear topology which is coarser than skorokhod s @xmath1 topology .
the delicate point is that condition ( [ th4:c ] ) does not hold for @xmath1 .
the corresponding typical example is given in figure 2 , with parameters @xmath178 .
to overcome this difficulty we shall introduce a variant of the @xmath1 topology , called @xmath179 ( @xmath166 - for modified ) , which slightly weakens the original topology and for which condition ( [ th4:c ] ) is satisfied .
[ rys2 ] , but converges in @xmath179,width=377 ] fix @xmath31 and consider a one - to - one embedding @xmath180 ) \ni x \mapsto \widetilde{x } \in { \mathds{d}}([-\varepsilon , t + \varepsilon])\ ] ] given by the formula @xmath181 take the complete metric @xmath182 on @xmath183)$ ] ( see @xcite ) ) and define @xmath184 then @xmath185 becomes a metric space and the corresponding topology will be called @xmath179 . *
( i ) * @xmath185 is a polish ( i.e. complete and separable ) metric space . *
( ii ) * a subset @xmath46 is relatively @xmath179-compact iff it is uniformly bounded : @xmath186 } | x(t ) |
< + \infty,\ ] ] and @xmath187 where we use the conventions @xmath188 and @xmath189 . * ( iii ) * for each @xmath190 the set @xmath191}\,;\ , |a| \leq a , u \in [ 0,t]\}$ ] is relatively @xmath179-compact .
[ emjeyone ] it is easy to see that if @xmath192 converges in @xmath193 ) , d_{sk}\big)$ ] to some @xmath194 , then the limit is of the form @xmath195 , for some @xmath196 .
hence @xmath185 is homeomorphic to the closed subset of the polish space @xmath193 ) , d_{sk}\big)$ ] , and so it is polish itself . part ( ii ) is a specification of theorem 14.4 in @xcite . and part ( iii ) is a direct consequence of our definition ( see figure 2 ) .
taking into account theorems [ finest ] and [ emjeyone ] we obtain the following interesting result , positioning the @xmath0 topology in the hierarchy of topologies on @xmath3 .
every linear topology on @xmath3 , which is coarser than @xmath197 , is coarser than the @xmath0 topology as well .
were @xmath198 a linear topological space , @xmath0 would be the finest linear topology on @xmath3 `` below '' @xmath179 .
following frchet , we say that @xmath199 is a space of type @xmath9 , if among all sequences of elements of @xmath199 a class @xmath200 of convergent " sequences is distinguished in such a way that : say that @xmath211 is @xmath212-closed if limits of @xmath210-convergent sequences of elements of @xmath213 remain in @xmath213 , i.e. if @xmath214 and @xmath202 , then @xmath215 .
the topology given by @xmath212-closed sets is called the sequential topology generated by the @xmath9-convergence @xmath210 and will be denoted by @xmath212 .
it must be stressed that for a sequential topology to be defined only the extremely simple properties ( ii ) and ( iii ) of convergence @xmath210 are required . on the other hand , obtained this way topology has in general extremely poor separation properties .
it is only t1 space due to the fact that in view of ( ii ) above each one - point set @xmath216 is @xmath212-closed .
but it is enough for the topology @xmath212 to define a new ( in general ) convergence , `` @xmath217 '' say , which , after urysohn , is called the convergence a posteriori " , in order to distinguish from the original convergence (= convergence a priori " , i.e. `` @xmath210 '' ) .
so @xmath201 converges _ a posteriori _ to @xmath218 if for every @xmath212-open set @xmath219 containing @xmath161 eventually all elements of the sequence @xmath201 belong to @xmath219 .
then the convergence @xmath210 is called an @xmath52-convergence .
it is an immediate consequence of theorem [ kvpk_recipe ] that if we start with an @xmath52-convergence then the convergences _ a posteriori _ and _ a priori _ coincide .
it follows that given an @xmath9-convergence @xmath210 " we can _ weaken _ it to an @xmath52-convergence @xmath222 " which is already the usual convergence of sequences in the topological space @xmath223 .
at least two examples of such a procedure are commonly known .
let @xmath225 and take a sequence @xmath226 say that @xmath227 if for each @xmath228 i.e. @xmath40 converges to @xmath161 at the given rate @xmath229 then @xmath222 " means the usual convergence of real numbers .
it is worth to realize that a set @xmath230 is relatively @xmath210-compact ( i.e. in each sequence @xmath231 one can find a subsequence @xmath65 such that @xmath232 , for some @xmath233 ) iff it is relatively @xmath222-compact .
let us notice that if @xmath234 is a hausdorff topological space , then @xmath235 and in general this inclusion may be strict ( like in the case of weak topology on an infinite dimensional hilbert space ) .
jakubowski , a. : from convergence of functions to convergence of stochastic processes . on skorokhod s sequential approach to convergence in distribution . in : korolyuk , v. , portenko , n. and syta , h. , eds .
, skorokhod s ideas in probability theory , _ insitute of mathematics , national academy of sciences of ukraine _ , kyiv 2000 , 179 - 194 .
ouknine , y. , pardoux , e. : homogenization of pdes with nonlinear boundary condition . in : dalang .
c. et al . : seminar on stochastic analysis , random fields and applications iii , _ birkhuser _ , basel 2002 , 229242 . | the @xmath0 topology on the skorokhod space was introduced by the author in 1997 and since then it proved to be a useful tool in several areas of the theory of stochastic processes .
the paper brings complementary information on the @xmath0 topology .
it is shown that the convergence of sequences in the @xmath0 topology admits a compact description , exhibiting the locally convex character of the @xmath0 topology .
it is also shown that @xmath0 is , up to some technicalities , finer than any linear topology which is coarser than skorokhod s @xmath1 topology . + * keywords : * functional convergence of stochastic processes ; @xmath0 topology ; @xmath1 topology ; skorokhod space ; sequential spaces + * ams msc 2000 : * 60f17 ; 60b11 ; 54d55 ; 54a10 |
in the total photoabsorption cross section on the proton above 400 mev the two pion production becomes possible , as it is clearly shown by several data sets from bubble chamber real photon experiments @xcite , as well as from exclusive electron scattering experiments @xcite and recent measurements with real photon beams with the daphne @xcite and saphir@xcite detectors . from the data
one can see that the cross section for this process shows a steep rise at photon energies above threshold and then exhibits a smooth decrease , although still representing a large fraction of the total hadron photoproduction at @xmath2 gev .
it is well established @xcite that a large contribution to this process is given by the intermediate formation of @xmath3(1236 ) , @xmath4(770 ) and their subsequent decay in the cascade processes @xmath5\ ] ] @xmath6 where indices @xmath7 stand for real and virtual photons . an important role in quasi two - body reactions ( 1),(2 )
is played by nucleon resonances , whose excitation is clearly seen in the total photohadronic cross section @xcite as well as in exclusive channels like single pion photoproduction @xcite .
many of such nucleon resonances have an appreciable branching ratio in the @xmath8 and @xmath9 channels ( see table i ) and for @xmath1 heavier than 1.7 gev multipion decay channels become dominant . therefore reactions ( 1),(2 ) offer a promising opportunity to study the structure of high lying resonances composed by light @xmath10 and @xmath11 quarks .
moreover , quark model calculations based on the three constituent quark picture @xcite predict more states than those found in experimental searches@xcite .
some parameters for such `` missing '' states are shown in table ii : many of them are expected to have a strong branching ratio in the two - pion final state and weak or absent coupling to the single pion : this could be a reason for them having escaped detection in experiments with pion beams and in single pion photoproduction .
an additional prediction for these unobserved states is that some of them could have sizeable electromagnetic couplings , similar to other states observed in electromagnetic production .
therefore , sizeable electromagnetic couplings and strong decays through @xmath8 or @xmath9 channels could make poorly known states more visible as well as make some of the `` missing '' states appear in measurements of photo- or electroproduction of the two - pion final states .
it is the aim of a wide experimental program at new facilities like tjnaf @xcite to investigate @xmath1 structure and to perform a search for missing states by measuring the multipion production exclusive channels produced via electromagnetic interaction . in this paper
we will focus on the particular two - pion production channel that proceeds through the @xmath12 intermediate state .
an important feature in the study of resonances in reactions ( 1),(2 ) is the strong contribution from non - resonant processes @xcite , creating a `` continuum '' in which resonant states are embedded .
for this reason , common methods for resonance investigation such as multipole amplitude analysis @xcite , which essentially use resonance dominance , may be not very effective for reactions ( 1),(2 ) .
different approaches to describe double pion photoproduction have been presented in a few papers@xcite , where models based on a variety of tree - level diagrams and a few baryon resonances have been used to calculate the two pion production .
the limited number of resonances included however makes them applicable only for w lower than about 1.6 gev ; moreover non - resonant terms are not always corrected for unitarity absorption effects : actually , a very important problem in the investigation of @xmath1 in exclusive meson production by photons is the description of non - resonant processes at @xmath13 gev , where many competitive channels open up and coupled channel calculations@xcite appear to be very difficult due to restricted knowledge of hadron couplings as well as to computational problems . a phenomenological description of the interaction with open inelastic channels in the initial and final states ( isi and fsi ) has been proposed first in @xcite . in this approach particle absorption in the initial and final states accounts for the complexity of interactions with inelastic channels ;
the absorptive coefficients have been usually obtained@xcite assuming a diffractive character of the interaction of ingoing and outgoing particles ; this assumption could be justified at w values above @xmath1 excitation region , while at @xmath14 gev other mechanisms like especially s - channel processes should contribute in isi and fsi significantly . moreover , the high value of the effective coupling coefficients , close to the physical limit , adopted in @xcite , at relatively small @xmath15 gev looks somewhat artificial and not physically justified .
the need to develop a method for @xmath1 electromagnetic form factor investigation from the experimental two - pion photo- and electroproduction data led us to the development of a phenomenological approach based on minimal model ingredients .
this was accomplished parametrizing , under simple assumptions , the main two - pion photo- and electroproduction mechanisms and using experimental data to determine the corresponding parameters .
this paper is our first step in establishing such phenomenological approach : we focused our attention on the particular quasi - two - body channel ( 1 ) and set up a simple model including all known resonant states contributing to this reaction and non - resonant processes starting from a minimal set of mechanisms proposed in @xcite .
we used existing data for the pion electromagnetic form factor to provide a description of non - resonant terms for @xmath16 , as well as data about the strong form factor in @xmath17 vertex to take into account the particle size in hadronic vertices .
unitarity effects manifesting in the competition of many hadronic channels in non - resonant processes were taken into account effectively , implementing initial and final state interactions ( isi and fsi ) , in form of absorptive corrections ; a new feature of our approach is that we developed a specific parametrisation of isi and fsi mechanisms responsible for channel coupling at w @xmath18 gev , based on experimental information about hadronic scattering amplitudes in this energy region . in this regard , we examined some aspects of the @xmath1 electromagnetic vertex dressing as related to our implementation of higher order corrections in form of isi and fsi .
we studied also the transition to the higher energy region , w @xmath19 2 gev , where the pure tree - level scheme corrected for finite size ( strong form factor ) and unitarity effects ( isi - fsi ) fails in reproducing the experimental cross section : we found that a regge trajectory exchange , in the particular form recently proposed@xcite , can be a valid continuation into the higher energy region .
our model relates @xmath1 electromagnetic helicity couplings @xmath20 , @xmath21 , @xmath22 and measured cross sections for reaction ( 1 ) induced both by real and virtual photons , therefore offering a way to attempt the measurement of @xmath1 contributions from a comparison to experimental data or a fit .
the helicity representation has been chosen to describe the amplitudes for reaction ( 1 ) .
the differential cross section for reaction ( 1 ) induced by real or virtual photons in the one - photon exchange approximation can be expressed as @xcite : @xmath23 \frac{1}{(2\pi)^{2}}\frac{p_{\pi}^{*}}{4w}\ ] ] @xmath24 where @xmath25 is cm total energy , @xmath26 is the pion three momentum modulus in the cm frame , @xmath27 is the `` equivalent '' photon energy , @xmath28 is the nucleon mass , @xmath29 = - @xmath30 where @xmath31 is the real - virtual photon four - momentum , @xmath32 is the photon polarization parameter , @xmath33 is the leptonic tensor well - known from qed@xcite . the information about hadronic production is contained in the hadronic tensor @xmath34 which is a bilinear combination of hadronic currents : @xmath35 related to the helicity amplitudes @xmath36 according to : @xmath37 where @xmath38 , @xmath39 , @xmath40 are helicities of @xmath3 , proton and photon , respectively , @xmath41 is the four vector associated to the photon polarization state @xmath40 and all variables are evaluated in the reaction cm frame . being the @xmath3 an unstable particle , it is necessary to fold ( 3 ) on its mass distribution as follows@xcite : @xmath42 where @xmath43 is the squared running invariant mass of @xmath3 , while @xmath44 , @xmath45 are @xmath3 mass and width , respectively .
the mechanisms of reaction ( 1 ) were described by a minimal set of feynman tree - level diagrams presented in fig .
1 . these diagrams can be subdivided into @xmath1 contributions ( fig.1a ) and a group of non - resonant processes , or born terms ( fig.1b - e ) .
the resonant part ( fig.1a ) involves all relevant @xmath1 and @xmath46 excitations in s - channel .
the non - resonant amplitudes correspond to the same set of mechanisms considered in @xcite .
new features of our approach in this respect are : 1 . implementation of electromagnetic vertex functions to describe the behaviour of non - resonant amplitudes at @xmath16 and 2 .
implementation of a strong @xmath47 form factor , relativistically invariant and based on the @xmath48 scattering analysis , to take into account the finite size of hadrons involved in non - resonant mechanisms .
the non - resonant processes are composed by the `` seagull '' or `` contact '' term ( fig .
1b ) ( which also naturally arises when considering the @xmath17 vertex as described by an effective meson - baryon lagrangian and then introducing the interaction with the electromagnetic field by the minimal coupling prescription ) , the t - channel pion - in - flight diagram ( fig.1c ) , the u - channel delta - in - flight diagram ( fig.1e ) and by the s - pole nucleon term ( fig.1d)@xcite . to treat photon , pion and delta off - shell we introduced form factors in the corresponding vertices .
therefore born terms ( fig.1b - e ) are functions of @xmath29 and mandelstam t variables . for the vertex function evaluation we used a compilation of experimental data about electromagnetic and strong form factors @xcite
; these data provide a reliable vertex function evaluation in the relevant region of @xmath29 and t ( @xmath49 @xmath50 , @xmath51 @xmath50 ) . in momentum space ,
helicity amplitudes corresponding to the contact term are given by @xmath52\ ] ] where w is the usual invariant cm energy , @xmath53 is the pion production angle in the hadronic cms , @xmath54 and @xmath55 are the target nucleon and @xmath3-particle four momenta , @xmath56 is the photon four momentum and @xmath57 , @xmath58 and @xmath10 are rarita - schwinger spinor - tensor for @xmath3-particle with @xmath38 helicity , polarisation vector for photon with @xmath40 helicity and spinor for target nucleon with @xmath39 helicity , respectively ; @xmath59 is an effective `` seagull '' term vertex function .
the pion - in - flight contribution in momentum representation reads : @xmath60 where @xmath61 is the pion momentum , @xmath62 is the pion mass , @xmath63 is the product of strong and electromagnetic vertex functions : @xmath64 the electromagnetic vertex function in this case was described by the well - known pole fit of pion form factor @xcite @xmath65 where @xmath66 corresponds to pion production in hadronic cms at zero degree angle ; the factor @xmath67 reflects the way of @xmath68 form factor extraction from single pion electroproduction data presented in @xcite .
analysis in @xcite yielded @xmath69 @xmath50 and we used this value in calculations . however considering uncertainties in @xcite as well as other data about pion electromagnetic form factor @xcite
, we could allow for a variation of @xmath70 cut - off parameter within @xmath71 gev@xmath72 . concerning the @xmath47 t - dependence , we introduce as vertex function a strong form factor successfully applied in @xmath73 relativistic transition potentials@xcite @xmath74 the interaction constant @xmath75 and cut - off parameter @xmath76 are : @xmath77 and @xmath78 gev@xcite .
we found that best data reproduction corresponds to @xmath79 gev .
gauge invariance implies equality of the `` seagull '' term coupling @xmath59 and the pion - in - flight term coupling @xmath80 ; we stress that , instead of applying specific electromagnetic and strong vertex functions to the contact term ( which would imply to introduce some free parameters to be fitted ) , we rather fixed the @xmath81 of them @xmath59 requiring gauge invariance . the s - channel nucleon contribution ( diagram of fig.1d )
is given by @xmath82 where s is mandelstam invariant @xmath83 , @xmath84 is the nucleon mass , @xmath85 is the electromagnetic vertex function ( we put the strong vertex function for s - channel equal to unity ) , described by the well - known dipole fit@xcite : @xmath86 the last contribution to the born terms that we considered was the @xmath3-in - flight ( diagram of fig.1e ) . the expression for such process is : @xmath87 where u is mandelstam variable corresponding to the crossed invariant momentum transfer @xmath88 ; factor 2 accounts for the double electric charge of the @xmath89 particle , while again @xmath90 is the product of electromagnetic and strong vertex functions . due to lack of information about @xmath3 s elastic electromagnetic form factors we followed two ways to evaluate @xmath90 vertex function : first
, it was derived from gauge invariance of the total born amplitude : @xmath91 second , we used function ( 11 ) for @xmath47 vertex , substituting mass @xmath62 by @xmath92 . in the latter choice we found the @xmath3-in - flight contribution to be negligible , because according to @xcite @xmath76 cut - off parameter does not exceed 1.3 gev .
of course in the second evaluation of @xmath3-in - flight term gauge invariance of the total born amplitude is lost but , considering that our description of non - resonant terms aims to be phenomenological , we can assume that gauge invariance could be restored by other mechanisms not contributing significantly to the cross section .
moreover , a negligible @xmath3-in - flight contribution seems to be supported by the comparison with the experimental data .
it is well known @xcite that the tree - level calculation is not able to reproduce the data extracted from experimental analysis of @xmath0 production .
one reason is that the lowest order mechanisms for reaction ( 1 ) do not incorporate unitarity and hence interaction with other open channels in initial and final states must be taken into account .
direct description of the unitarity effects in the initial and final states would require a simultaneous treatment of all relevant hadronic channels in a coupled channel calculation using hadronic elastic and inelastic scattering amplitudes data .
the present status of the hadronic amplitudes knowledge as well as computational capabilities restrict coupled channel calculations to w lower than 1.6 gev@xcite , while investigation of the cross section for process ( 1 ) above 1.6 gev is very important for @xmath1 physics .
the description of open channels competition in non - resonance processes remains a challenging open problem of @xmath1 structure investigation in exclusive meson photo- and electroproduction .
the modern status of strong interaction theory does not allow to evaluate hadronic amplitudes and couplings from fundamental principles ( i.e. qcd lagrangian ) .
therefore we chose to adopt an effective treatment of unitarity effects in initial and final state ; the initial and final state interactions responsible for the modification of the amplitude for reaction ( 1 ) are schematically depicted in fig .
2 . in the spirit of vector meson dominance ( vdm)@xcite ,
the isi was assumed to proceed through a transition between photon and vector meson @xmath93 .
the black blobs describe all transition processes , between @xmath94 initial and @xmath95 intermediate states as well as between @xmath96 intermediate and @xmath97 final states .
we evaluated isi and fsi effects in @xmath98 reaction in the frame of a simple phenomenological recipe described in @xcite , where ingoing and outgoing particle interaction is described by penetration factors in the initial and final states @xmath99 , @xmath100 , which depend on the reaction total angular momentum @xmath101 ; they determine the amplitude for @xmath94 initial state to be transformed into intermediate @xmath95 state and the amplitude for @xmath102 intermediate state to be transformed into final @xmath97 state . according to @xcite ,
the total born amplitude @xmath103 was numerically decomposed in partial waves with total angular momentum @xmath101 : @xmath104 the isi & fsi modify the @xmath105 decomposition coefficients as : @xmath106 where @xmath107 represent decomposition coefficients for the born term amplitude modified by isi & fsi . @xmath108 and @xmath109 factors , under the assumptions of ref .
@xcite , can be related with @xmath110-matrix elements of @xmath0 and @xmath94 elastic scattering amplitudes at total angular momentum @xmath101 as : @xmath111 where @xmath112 , @xmath39 , @xmath38 are helicities for @xmath4-meson ( equal to photon helicity in the vdm picture ) , for proton and @xmath3-isobar , respectively ; s and t matrix elements are related as : @xmath113 the penetration coefficients @xmath108 and @xmath109 are unambiguously determined by @xmath97 and @xmath94 elastic scattering amplitudes , therefore these amplitudes effectively describe all mechanisms ( resonant and non - resonant ) responsible for transitions between the initial @xmath94 and intermediate @xmath95 states as well as between the intermediate @xmath102 and final @xmath97 states ( blobs in fig .
as mentioned in the introduction , a new feature of our approach is that , instead of using a diffractive ansatz like in the previous literature@xcite , better suited for higher energies , t - matrix elements for @xmath0 and @xmath94 elastic scattering were more appropriately evaluated using data about hadronic scattering in the resonance region , as briefly described in the following paragraphs .
t - matrix elements for @xmath0 and @xmath94 elastic scattering were evaluated in an isobar model @xcite assuming them to be a superposition of relevant @xmath1 contributions and a smooth background as depicted in fig .
the resonant part of the corresponding amplitudes was evaluated in breit
wigner ansatz described in details in sect .
3.3 . to obtain @xmath114 amplitude normalization
, we considered a schematical situation , when complexity of all transitions @xmath115 and @xmath116 processes is limited to a single @xmath1 excitation with elastic scattering as the only open channel ; for this case , ingoing and outgoing particle absorption should be absent at resonant point ( @xmath117 ) and the moduli of @xmath108 and @xmath109 coefficients should be equal to unity ; from unitarity conditions and from the standard resonance phase behavior , we determined @xmath118 amplitude normalization as @xmath119 at @xmath120 , therefore @xmath121 this way we obtained the following relations between elastic @xmath122 and @xmath94 scattering resonant amplitudes @xmath123 , @xmath124 and @xmath1 decay helicity amplitudes @xmath125 , @xmath126 : @xmath127 \left [ \frac{p^{c}_{\pi(\rho)}}{8\pi ( 2j + 1 ) w } \right]\end{aligned}\ ] ] where @xmath128 , @xmath129 are masses and w dependent @xmath1 decay widths , @xmath130 , @xmath131 are three momenta moduli of pion and @xmath4 .
summation in ( 21 ) is performed over all @xmath1 contributing to the partial wave of total angular momentum @xmath101 and presented in table iii ; @xmath132 , @xmath133 are decomposition coefficients of @xmath125 and @xmath126 @xmath1 decay amplitudes through the states of total angular momentum j and related with @xmath1 partial decay widths in helicity representation @xmath134 and @xmath135 according to ( 31 ) .
partial decay widths were taken from analysis @xcite and transformed from ls to helicity representation .
our approach should be somewhat modified if dressed @xmath136 electromagnetic vertices are used for the resonant part of reaction ( 1 ) amplitude ( sect 3.3 ) . indeed , in the complexity of transition mechanisms in isi & fsi shown in fig . 2 by black blobs , also the sequence of processes reported in fig . 4
is effectively taken into account ; that clearly represents a dressing mechanism of the @xmath136 electromagnetic vertex .
but for instance , @xmath1 electromagnetic helicity amplitudes @xmath20 , @xmath21 extracted from pion photoproduction data analysis@xcite are usually interpreted as corresponding to dressed @xmath136 verticies .
therefore , a double counting problem can arise in this case , as dressing effects of the electromagnetic @xmath1 vertex can be sizeable , as discussed in @xcite . to avoid double counting ,
one possibility is to exclude @xmath137 and @xmath138 mechanisms from isi & fsi . to work out a phenomenological prescription for such exclusion we considered the above mentioned schematical situation , where the complexity of @xmath115 and @xmath139 transition mechanisms is restricted to single @xmath1 excitation with only elastic scattering as open channel .
unitarity for breit wigner formula can be expressed as @xmath140 where @xmath141 , @xmath142 are total and partial @xmath1 decay widths .
for single @xmath1 with only elastic open channel @xmath143 where @xmath141 , @xmath144 are total and elastic @xmath1 decay width ; our assumption insures that all transition mechanisms in @xmath115 , @xmath139 are represented by @xmath137 and @xmath138 processes ; therefore exclusion of @xmath1 excitation from isi & fsi should lead to complete absorption of ingoing and outgoing particle or to zero values for @xmath108 , @xmath109 penetration factors .
as it follows from ( 19 ) the substitution @xmath145 would correspond to vanishing @xmath108 , @xmath109 coefficients and , therefore , could be considered as an empirical prescription for excluding @xmath137 and @xmath138 mechanisms from isi & fsi treatment .
the complete absorption of born term amplitudes takes place only for such schematical situation ( single @xmath1 excitation with only elastic open channel ) on which we based our exclusion prescription .
in actual situation prescription ( 24 ) gives only partial absorption of ingoing and outgoing particles , since @xmath1 contributing to @xmath97 and @xmath94 elastic scattering have many open decay channels , while relations ( 21 )
( 24 ) give zero @xmath108 , @xmath109 coefficients only in our schematic assumption .
inelasticities ( @xmath146 ) provide instead non - zero @xmath108 , @xmath109 coefficients even at resonant point ; @xmath1 off - shell excitations as well as non - resonant processes also give non - zero values of @xmath108 and @xmath109 coefficients .
we assumed that this partial absorption of born terms can represent isi & fsi corrections , being the contribution from @xmath137 and @xmath138 mechanisms excluded .
of course our exclusion procedure ( substitution ( 24 ) ) should be applied only if for the resonant part of reaction ( 1 ) dressed electromagnetic @xmath136 vertices are used .
actually , @xmath136 vertices calculated in quark models are generally assumed to be `` bare '' , i.e. free from higher order corrections ( see for instance @xcite ) : in this case our calculation should not be affected by any double counting and therefore no exclusion procedure is necessary .
the information on @xmath1 electromagnetic form factors from the measured cross section could be extracted both with and without application of exclusion procedure : in the first case we would obtain information about dressed @xmath136 vertices ; in the second about @xmath136 vertices without contribution from higher order corrections depicted in fig .
therefore our approach provides the flexibility of choosing one procedure or the other .
evaluation of non - resonant part of @xmath97 and @xmath94 scattering amplitude is described in more detail in @xcite , but we want to report here the main features of our evaluation .
we used data@xcite on partial - wave @xmath147 elastic cross - sections with definite total angular momentum and isospin ; the amplitudes for these partial waves were described by a superposition of all relevant @xmath1 ( table 3 ) and a particular background for each orbital momentum l and total spin s. the resonant part of @xmath147 elastic scattering amplitudes was treated in the same manner as @xmath148 and @xmath94 elastic scattering processes ; data@xcite were used for @xmath149 couplings .
background for each ls wave was parametrized as a function of w by a simple linear dependence @xmath150 parameters @xmath151 and @xmath152 were determined from our fit of @xmath147 elastic scattering data@xcite . to calculate background amplitudes for @xmath97 and @xmath94 elastic scattering starting from @xmath147 elastic scattering background we used su(3 ) flavour symmetry relations@xcite .
the fit results for @xmath147 partial elastic scattering cross section are shown in fig .
dotted curves represent resonant contributions , dashed lines are background contributions , while solid lines correspond to complete amplitudes . as follows from fig . 5 , in @xmath1 excitation region nucleon resonances provide the main contribution in @xmath147 elastic amplitudes , therefore also in penetration coefficients ( 17 ) .
this contribution was completely neglected in previous evaluations @xcite .
hence , implementation of s channel @xmath1 excitation amplitude performed in our approach is particularly important for isi & fsi treatment in @xmath1 excitation region .
for w above 2.0 gev the contribution of resonant part in @xmath147 elastic scattering partial waves ( fig .
5 ) falls down drastically and non resonant processes provide the main contribution for most partial waves .
therefore a diffractive approximation for isi and fsi absorptive corrections @xcite could be justified above @xmath1 excitation region . to check the reliability of our approach to evaluate isi and fsi absorptive factors , we also performed calculations of such effects implementing @xmath0 and @xmath94 elastic scattering amplitudes directly provided by the authors of analysis @xcite , based on a global unitary fit of @xmath147 scattering t - matrix .
the gross features of isi and fsi absorptive factors as obtained from these two procedures are in reasonable coincidence @xcite .
it is also worth to note that using the above mentioned description of born terms , together with our isi - fsi description , non - resonant processes in our approach do not have any free parameters to be determined from reaction ( 1 ) .
actually , our approach for isi & fsi treatment is mostly phenomenological and our assumptions can be mainly justified by comparison with experimental data ( we will discuss it in more detail in the next sections ) : pure born terms give rise to a cross section that does not reproduce real photon data at high w by a large factor , while absorption - corrected calculations are able to give a good account of data up to w around 2 gev .
implementation of pion regge trajectory exchange according to prescription @xcite in born terms allows to obtain a satisfactory data description up to w=3 gev as discussed in section 4 . assuming vector dominance as main mechanism responsible for @xmath29 evolution of coupling with other hadronic channels in the initial state we obtained the following expression for the @xmath99 absorptive factor @xmath29 dependence : @xmath153 where @xmath154 gev@xmath72@xcite .
a simple breit - wigner ansatz was chosen to describe the coherent superposition of all relevant @xmath1 , @xmath46 resonant amplitudes@xcite ( see table i , fig .
@xmath155 where @xmath156 , @xmath157 are resonance mass and energy - dependent total width and @xmath158 , @xmath159 are electromagnetic production and strong decay amplitudes of @xmath1 with helicity @xmath160 , respectively .
the @xmath1 off - shell effects were taken into account by the breit - wigner propagator in ( 23 ) as well as by the w - dependence of resonance width and strong decay amplitudes . following a phenomenological approach we related @xmath161 and @xmath162 amplitudes with observables extracted from experimental data analysis .
electromagnetic @xmath1 amplitudes @xmath163 were expressed in terms of commonly used helicity couplings @xmath20 , @xmath21 and @xmath22@xcite . comparison of the cross section for only one isolated @xmath1 state calculated according to ( 3)-(6 ) with breit - wigner formula gives : @xmath164 where @xmath165 and @xmath166 are photon three - momentum moduli in the cm frame at resonance point ( @xmath167 ) and at running w , respectively .
couplings @xmath168 , @xmath169 , @xmath170 completely describe @xmath1 electromagnetic excitation and their values , calculated in a model or taken from some experimental analysis , can be used to calculate the resonant part of cross section for reaction ( 1 ) . on the other hand , considering these couplings as parameters our approach could be used to attempt their extraction from a fit of the measured cross section .
the @xmath1 strong decay amplitudes @xmath159 were parametrised through the projection on the set of states with definite total angular momentum j : @xmath171 where @xmath172 is the decomposition coefficient , that does not depend on the resonance helicity state due to rotational invariance , @xmath173 is the cm pion emission angle , @xmath174 and @xmath26 are cm pion three - momentum moduli at resonance and running w , respectively .
we then related the @xmath172 parameters with the partial decay widths @xmath175 of @xmath176 in ls- representation extracted in analysis @xcite .
this way we obtained the decay amplitudes at the resonance point .
general requirements for the amplitude threshold behaviour@xcite as well as data analysis of single - pion production by photon and pion beams@xcite suggest the introduction of a w dependence of @xmath1 decay amplitudes in ( 27 ) .
assuming as usually that barrier penetration effects provide a good description of @xmath25- evolution in @xmath177 decay amplitudes and using the parametrization from @xcite we arrived to : @xmath178^{1/2}\end{aligned}\ ] ] where @xmath179 , @xmath180 are bessel s and neumann s functions and the factor in square brackets in ( 30 ) appears from barrier penetration ratio@xcite for a particle emitted with relative orbital momentum @xmath181 , while r is an interaction radius whose value was set to 1 fm . the decay amplitude ( 30 )
was then transformed into the helicity representation @xmath182 and using the relationship between two - body decay amplitude and width @xmath134 we obtained : @xmath183 where the integration over running @xmath3 mass squared @xmath43 takes into account the unstable character of @xmath3 particle and the quantity @xmath184 contains the above mentioned w evolution according to ( 30 ) .
the total @xmath1 decay width ( @xmath185 in ( 27 ) ) was assumed to be a sum over all partial widths presented in @xcite .
the w evolution of each partial width @xmath186 was evaluated again based on a barrier penetration ansatz : @xmath187 where @xmath26 and @xmath174 are three - momenta moduli of meson at running w and at resonance point respectively .
the total amplitude of @xmath188 process was then evaluated as a coherent superposition of resonant and total born amplitude @xmath189 , corrected for ingoing and outgoing channel absorption : @xmath190
using the above described approach we performed a cross section calculation for reaction ( 1 ) .
the resonances included in the evaluation are listed in table i. the relative contribution of a particular @xmath1 in the cross section can be described by the factor of merit @xmath191@xmath192 also reported in table i. all three- and four - star resonances@xcite with @xmath193 were included .
as follows from table i , @xmath194 , @xmath195 and @xmath196 resonances give maximum contribution and reaction ( 1 ) presents a promising opportunity for investigation of their structure . as mentioned in sect .
3.1 , we adopted two ways for @xmath3-in - flight term evaluation : a ) from gauge invariance requirements ; b ) neglecting this term due to proximity of @xmath3 mass and @xmath76 cut - off parameter .
neglecting @xmath3-in - flight term we obtained a better data description at high w value and pion emission angles .
the results are presented in fig . 6 by solid lines .
the data are reasonably reproduced in the overall w region .
a remarkable point is that our calculations predict cross sections at @xmath197 and @xmath2 gev lower than 1.2 @xmath198b / sr ; such strong cross section suppression namely results from absorption in the initial and final states due to interactions with open hadronic channels . for w = 1.62 gev
the calculated cross sections are a bit below the data points ; this discrepancy could be ascribed to simultaneous uncertainties in the cut - off parameter appearing in the strong vertex function used for the born terms , in the @xmath199 , @xmath200 elastic amplitudes appearing in the absorption parameterisation , as well as to uncertainties in some s - channel @xmath1 electromagnetic and strong parameters . the born term contribution in angular distributions for reaction ( 1 )
is shown by dashed lines in fig .
6 , therefore the maximum @xmath1 contribution takes place for cm pion emission angle above @xmath201 .
the evaluation of angular distributions performed by varying @xmath1 strong decay couplings inside uncertainties of analysis @xcite demonstrated that the cross section variation does not exceed a few percent , being negligible compared with data uncertainties : therefore , our approach for the non - resonant background appears to be quite stable with respect to @xmath1 strong decay parameter variation ; this is an important feature for the extraction of @xmath1 electromagnetic form factors from measured cross sections .
the comparison between the calculated total cross section for reaction ( 1 ) with recent saphir data@xcite as well as with old abbhhm collaboration data@xcite is presented in fig . 7 . decomposition of the total cross - section for reaction ( 1 ) in resonant , non - resonant processes and interference terms is also shown in fig .
the maximum contribution of @xmath1 ( at level @xmath202 ) is found at @xmath203 gev and decreases as w increases .
the region @xmath203 gev also corresponds to maximum contribution of interference effects in coincidence with results presented in @xcite . to demonstrate the effects of @xmath47 form factor implementation as well as isi and fsi absorptive corrections , we also reported in fig . 8 calculation results assuming : a ) @xmath47 form factor equal to unity and absence of absorptive corrections ( dashed line in fig . 8) ; b ) @xmath47 form factor from @xcite and absence of absorptive corrections ( dotted line in fig .
8) ; c ) complete model ( solid line ) .
calculation a ) is not able to reproduce data at all : at @xmath13 gev the difference between calculated and measured cross - section marks the complete failure of the pure born terms calculation .
implementation of @xmath47 strong form factor turns out to be very important ( dotted line of fig .
8) : we stress that such strong vertex function was taken from n
n scattering analysis without further tuning and assumed to be appropriate to describe the @xmath47 vertex in the t - channel pion exchange diagram , while for the contact term which is of similar magnitude as the pion exchange , we used gauge invariance as guidance for the vertex functions evaluation .
however without initial and final state absorptive corrections it is not possible to reproduce data at w above 1.7 gev . on the other hand ,
even neglecting isi and fsi effects our approach is able to describe reasonably the data at @xmath203 gev .
implementation of isi and fsi absorptive correction factors provides a reasonable agreement between measured and calculated cross sections for w below 1.8 gev , while for w above 1.9 gev calculated cross sections are systematically higher than data .
the reason for such deviation could be connected to the description of @xmath204 , @xmath94 elastic scattering amplitudes in isi - fsi : this aspect could be improved by adding contributions from isobar states with higher mass and spin ; another explanation could be that for w above 1.9 gev we may start to observe a transition from the corrected born terms picture to other processes , as those discussed in @xcite ; in this case , the framework for non - resonant reaction mechanisms in @xmath98 reaction should be modified at w above 1.9 gev .
we discuss these aspects in the next paragraph , to examine the limits of applicability of our meson - baryon approach and study how to continue the description of reaction ( 1 ) in the higher energy region .
following the recipe of @xcite , we replaced the pion exchange amplitude in born term ( fig .
1c ) by a pion regge trajectory exchange .
the results are presented in fig . 9 ( dotted line ) and compared with the non - reggeized pion exchange ( solid line ) .
regge trajectory exchange implementation provides a better data description at high w values .
however , even in this case the calculated cross section appears to be systematically higher than the data .
the possible reason could be the need to modify also the contact term in the way proposed in @xcite , thereby restoring gauge invariance .
we found that such contact term modification led to a sizeable cross section reduction at w below 1.6 gev , where descriptions in the frame of pion and regge trajectory exchange have to coincide . to provide this coincidence we had to require @xmath205
( pion regge trajectory ) effective coupling to be a factor 1.2 higher than @xmath47 coupling .
the cross section calculated under this assumption is shown in fig .
9 with a dashed line and reproduces reasonably the data up to 2.5 gev .
as mentioned above , non resonant part for @xmath97 and @xmath94 elastic scattering amplitudes in isi - fsi was estimated from @xmath147 scattering partial wave data fit@xcite , containing data up to w @xmath206 2.1 gev .
this is the main reason to restrict our calculation to w @xmath206 2.5 gev ( 3 gev photon energy in lab .
frame ) . implementation of any data on @xmath97 and @xmath94 elastic scattering amplitudes above 2 gev would allow a model extension toward higher w. to investigate the influence of @xmath136 vertex dressing on our cross section , we reported in fig .
10 the calculations with @xmath1 electromagnetic helicity amplitudes taken from pdg@xcite and from quark models@xcite . in the pdg case , we assumed the effects of @xmath136 vertex dressing at level of meson baryon degrees of freedom to be included in the numbers quoted , while in the quark models case we assumed such effects not to be accounted for ( `` bare '' @xmath136 vertices ) . therefore , using the pdg amplitudes we applied our prescription ( 24 ) for removal of @xmath1 dressing in isi - fsi , while in cross section calculations with @xmath136 vertices from quark models we kept @xmath1 excitation terms in isi & fsi mechanisms with no modification .
in principle , one would expect these two results approximately to coincide , assuming that the same dressing effects are introduced one way or another .
our results show that this is not the case , being the cross section calculated using the `` dressed '' pdg values systematically higher than the quark model results ; moreover , if we removed completely the @xmath1 from the isi - fsi description , the discrepancy would be even bigger ; we also found that the difference between the two quark model cross sections is significantly smaller than the deviation between quark model and pdg results .
actually , the difference between the cross section evaluated with `` bare '' and `` dressed '' vertices could be due to specific approximations of quark model approaches@xcite ; or it could be an indication that additional dressing effects are present in the experimentally extracted photocouplings from @xcite .
this difference does not exceed 30 % , with a maximum at w between 1.5 1.7 gev .
this w range actually corresponds to maximum @xmath1 contribution .
the discrepancy appears to be negligible at w above 2.0 gev and below 1.4 gev , where @xmath1 contributions are less pronunced . to see the influence of particular dressing effects
@xmath1 excitation in @xmath97 and @xmath94 elastic scattering processes shown in fig .
we performed calculations using `` dressed '' @xmath1 electromagnetic vertices@xcite , but keeping the @xmath1 excitation in isi & fsi mechanisms for born terms .
the results are shown in fig .
10 by dotted line ; of course in this case some double counting of the @xmath1 electromagnetic vertex dressing takes place , but surprisingly the cross section seems to be in better agreement with the `` bare '' quark model results . in any case , the difference between solid and dotted curves could be considered as estimation of this particular @xmath136 vertex dressing contribution in the cross section .
this contribution is lower than 20 % and vanishes at w above 2.0 gev , where @xmath1 excitation is negligible . to extract @xmath1 helicity amplitudes for @xmath16
it is important to have a good description of the @xmath29 dependence for non - resonant processes .
we checked this point by comparing the total cross section for reaction ( 1 ) calculated in our approach with data reported in @xcite .
for the resonant part we used results from a single quark transition model ( sqtm ) fit@xcite about the @xmath29 evolution of @xmath1 photocouplings . the comparison between data@xcite and our calculations is presented in fig .
11 where dashed lines correspond to the contribution of non resonant processes only , while the complete evaluation with @xmath1 included is shown by solid lines . considering this comparison , two important points must be stressed : first , bins for experimental data@xcite , both in w and @xmath29 were very wide , implying a strong cross section averaging over the measured kinematical range ; second , experimental data about high - lying @xmath1 photocouplings are rather scarce and do not allow to check the validity of sqtm predictions .
however , our approach reproduces the measured cross - section for w in the 1.3 - 1.5 gev and in the 1.5 - 1.7 gev interval ; for w in the 1.7 - 2.0 gev bin our calculated cross - section is higher than the measurements ; the reasons for such deviation could be namely the poor knowledge of the @xmath1 contributions or the kinematical average effects , as well as the above mentioned limitations in the treatment of born terms absorption in the higher w region .
new high precision data upcoming from new facilities like tjnaf are therefore definitely needed for a better understanding of the @xmath29 behaviour of @xmath1 electromagnetic form factors .
an important prediction of our evaluations is a strong @xmath1 contribution in the w = 1.5 1.7 gev interval , representing over 60 % of cross section at @xmath207 gev@xmath72 .
therefore the two pion exclusive channel seems to present a promising opportunity for @xmath1 structure investigation by photons with high virtuality . to investigate @xmath1 electromagnetic vertex dressing effects on the cross section at @xmath16
, we performed the calculation with @xmath1 electromagnetic form factors taken from different approaches@xcite .
form factors in approach@xcite were determined from exclusive single pion production data analysis , imposing symmetry relations between form factors of @xmath1 s belonging to a particular su(6 ) multiplet : therefore they can be assumed to represent `` dressed '' vertices . instead
, quark models @xcite should not contain higher order corrections and we considered their results as representing `` bare '' vertices .
again we expected to have approximate coincidence of results obtained using these different ingredients .
actually in this case , according to fig .
12 , the difference between cross sections calculated with the two `` bare '' vertices @xcite and @xcite is higher than the difference between results obtained using form factors from @xcite and quark models results .
also in this case we calculated the cross section using `` dressed '' vertices from @xcite , with and without prescription ( 24 ) for @xmath1 removal in isi - fsi ( solid and dotted curve in fig .
12 , respectively ) ; in the latter case the result is supposed to be affected by double counting , but the difference can give an indication of the extent of dressing introduced by our effective unitarity implementation ; as shown in fig .
12 , such effect appears to be negligible , thereby indicating that , for the @xmath16 evolution in our framework , quark model theoretical uncertainties could be more important than dressing effects .
therefore , comparison of cross sections calculated in our approach with forthcoming precise experimental data at @xmath16 can provide a promising opportunity to select between model approaches for @xmath1 structure description in non perturbative qcd region . as mentioned above
, we also compared the cross section calculated using `` dressed '' @xmath1 form factors@xcite with / without implementation of @xmath1 excitation in isi & fsi mechanisms ( dotted and solid lines in fig .
the contribution of this particular dressing mechanism is negligible for w@xmath208 1.4 and 1.85 gev and vanishes at @xmath209 gev@xmath72 for w @xmath210 gev .
there are actually reasons for such behaviour : a ) the relative contribution of non resonant processes drastically falls down as @xmath29 increases ( from more than 80 % at the photon point to lower than 50 % at @xmath29 above 1 gev@xmath72 ; b ) the isi effects also go down as @xmath29 increases due to suppression of transitions between photon and vector meson .
we have extensively studied a phenomenological model for the two pion photo- and electroproduction through the intermediate @xmath211 channel , as part of a broader effort in establishing a basis for analysis and interpretation of the upcoming data from tjnaf@xcite , where this as well as other exclusive electromagnetic production channels will be studied with unprecedented accuracy .
our calculation included a minimal set of born amplitudes , with appropriate absorptive corrections to effectively take into account interaction with open channels in the initial and final states , plus a large number of nucleon resonances that , according to the existing data , are thought to give a sizeable contribution to the cross section . a strong form factor for @xmath17 vertex was introduced according to nn scattering experiments .
data about pion and proton electromagnetic form factors were also used to evaluate the pion - in - flight term and the s - channel nucleon pole term , respectively , for @xmath212 , while the @xmath29 behaviour of the contact term was obtained imposing gauge invariance ; the delta - in - flight term was calculated both from gauge invariance and using a @xmath17 vertex determined from nn scattering analysis .
strong and electromagnetic @xmath1 couplings were related with experimental observables : electromagnetic helicity couplings @xmath20 , @xmath21 , @xmath22 and partial hadronic decay widths @xmath213 .
we found that strong absorptive corrections were essential to reproduce experimental data for cm pion emission angle above 30@xmath214 at w above 1.6 gev .
we developed a specific approach for isi and fsi absorptive corrections , relating absorptive factors in ingoing and outgoing channels with @xmath0 and @xmath94 elastic scattering amplitudes .
these elastic strong amplitudes were evaluated in a simple isobaric model , so that our approach did not have any free parameters to be determined from reaction ( 1 ) . any other approach for @xmath0 and @xmath94 elastic amplitude evaluation can be easily implemented in our calculation .
reasonable reproduction of pion angular distributions in the photon point as well as good agreement between measured and calculated total cross sections for @xmath16 demonstrate the ability of our approach to reproduce the main features of @xmath0 production by real and virtual photons in the resonance region .
moreover the proposed method for isi and fsi description can be considered as a reasonable phenomenological way to describe non - resonant processes in electromagnetic meson production for @xmath13 gev , where the competition of many open hadronic channels makes a rigorous background evaluation very difficult .
our unitarity corrections effectively introduce a dressing of the resonance amplitudes ; therefore we elaborated an empirical prescription to remove the dressing from isi - fsi when using experimental resonance photocouplings , assumed to be already `` dressed '' ; we also performed calculations without this prescription when using resonance amplitudes from quark models , asssumed to be `` bare '' . in the real photon case
, we found that the calculations performed with two different quark models show a good agreement , while the results obtained using the experimental pdg amplitudes show a systematic disagreement with the previous ones ; the reason is unclear and could be connected to other dressing effects than @xmath1 excitation in @xmath115 @xmath139 transitions as well as to quark model uncertainties and also to model - dependence in the experimental resonance amplitude extraction .
we performed the same analysis for the virtual photon cross section , but in this case we found a disagreement between the two quark models adopted , while the calculation based on resonance amplitudes extracted from an analysis of experimental data showed less sensitivity to our `` dressing '' effects : the quark models discrepancies appeared to be more important , leading us to believe that in the virtual photon case the sensitivity to different quark model ingredients could be more pronunced , although the reason for the different behaviour of real and virtual photon is not evident .
the calculation presented could serve as a first basis for interpreting the data coming from new facilities like jefferson lab and moreover , allowing the resonance parameters to vary , it could be the starting point of a fitting procedure with the goal of investigating electromagnetic form factors for high mass @xmath1 s ( @xmath215 gev ) , as well as for attempting to discover new states .
the model presented is currently being used as a foundation in our development of a full three - body final state description for the two - pion production off the nucleon by real and virtual photons@xcite . our special thanks to dr .
v. burkert from jefferson laboratory , for continuos interest , useful discussions and support .
particular thanks to prof .
s. dytman from the university of pittsburgh , for kindly providing the elastic hadronic amplitudes from his global fit and to prof .
m. giannini from the university of genova , italy , for kindly providing the @xmath1 electromagnetic transition amplitudes from his quark model .
we also want to thank prof .
mukhopadhyay from rensselaer polytechnic institute , for useful discussions and suggestions
. 99 cambridge bubble chamber group , phys .
155 , 1477(1967 ) ; abbhhm collaboration , phys .
175 , 1669 ( 1968 ) .
d. luke and p. soding , springer tracts in mod .
59 ( 1971 ) .
i. damman et al .
b54 , 355(1973 ) .
k. wacker et al . , nucl .
b144 , 269(1978 ) .
a. braghieri et al . , phys.lett .
b363 , 46 - 50(1995 ) .
b. krusche , in proceedings of the `` joint ect*/jlab workshop on n * physics and nonperturbative qcd '' , trento , italy , may 18 - 29 , 1998 , ed . by s. simula , b. saghai , n.c .
mukhopadhyay , v.d .
burkert , few body systems , suppl . 11 , 1999 .
klein , bonn university thesis bonn - ir-96 - 08 .
t.a armstrong et al . , phys .
d5 , 1640 ( 1972 ) ; t.a .
armstrong et al . , nucl .
b41 , 445 ( 1972 ) .
r. walker , phys .
182 , 1729(1969 ) ; r.g .
moorhouse et al . , phys .
d9 , 1(1974 ) .
r. koniuk , n. isgur , phys .
44 , 845(1980 ) ; r. koniuk , n. isgur , phys . rev .
d21 , 1868(1980 ) ; r. koniuk , nucl .
b195 , 452(1982 ) .
s. capstick , w. roberts , phys .
d49 , 4570(1994 ) .
r.bijker , f. iachello , a. leviatan , phys .
c54 , 1935(1996 ) .
particle data group , phys .
d54 , 1(1996 ) .
burkert , in proceedings of the `` joint ect*/jlab workshop on n * physics and nonperturbative qcd '' , trento , italy , may 18 - 29 , 1998 , ed . by s. simula , b. saghai , n.c .
mukhopadhyay , v.d .
burkert , few body systems , suppl . 11 , 1999 .
m. ripani , in proceedings of the `` joint ect*/jlab workshop on n * physics and nonperturbative qcd '' , trento , italy , may 18 - 29 , 1998 , ed .
by s. simula , b. saghai , n.c .
mukhopadhyay , v.d .
burkert , few body systems , suppl . 11 , 1999 .
berens , a. donnachie , nucl .
b84 , 342(1975 ) .
arndt , i.i .
strakovsky , r.l .
workman , phys .
c53 , 430(1996 ) .
a. bartl , w. majerotto , d. schildknecht , nuovo cimento 12a , 703(1972 ) .
murphy , j.m .
laget , dapnia - sphn-96 - 10 , mar .
gomez - tejedor and e. oset , nucl.phys .
a571 , 667(1994 ) ; j.a .
gomez - tejedor and e. oset , nucl.phys
. a600(1996 ) .
k. ochi , m. hirata , t. takaki , phys .
rev . , c56 , 1472(1997 ) . s. nozawa , b. blankleider , t .- s.h lee , nucl .
a513 , 459(1990 ) .
surya , f. gross , preprint cebaf - th-95 - 04&wm-95 - 101 . k.gottfried and j.d.jackson , nuovo cimento .
34 , 736(1964 ) .
m. guidal , j .-
laget , m. vanderhaeghen , phys .
b400 , 6 ( 1997 ) .
e. amaldi , s. fubini , g. furlan , pion electroproduction .
springer tracts in modern physics , 83(1989 ) .
m. pilkuhn , relativistic particle physics .
springer verlag ( 1979 ) .
r. machleidt in advances in nuclear physics v.19 ( 1979 ) .
bebek et al . ,
d17 1693(1978 ) .
l. y. glozman , et al .
, pin newslett .
14 , 99 ( 1998 ) p. brauel , t.canzler , e. a. , phys .
lett . , 69b , 253 ( 1977 ) .
r. machleidt , k holinde , ch .
elster , phys .
149 , 1(1987 ) .
m.jacob and g.c wick , ann . of phys
. 7 , 404(1959 ) .
m. battaglieri et al . , bull . of moscow state university
( accepted for publication ) m. anghinolfi et al . , physics of atomic nuclei 63(2000 )
d.m.manley , e.m.salesky , phys .
d45 , 4002(1992 ) .
koch , e. moniz , ann .
154 , 99(1984 ) t. s. h. lee in proceedings of the `` joint ect*/jlab workshop on n * physics and nonperturbative qcd '' , trento , italy , may 18 - 29 , 1998 , ed . by s. simula , b. saghai , n.c .
mukhopadhyay , v.d .
burkert , few body systems , suppl . 11 , 1999 , p157 .
m. ferraris et al .
b364 , 231 ( 1995 ) ; m. aiello et al . , j. phys .
24 , 753 ( 1998 ) ; m. m. giannini , e. santopinto in proceedings of the `` joint ect*/jlab workshop on n * physics and nonperturbative qcd '' , trento , italy , may 18 - 29 , 1998 , ed . by s. simula , b. saghai , n.c .
mukhopadhyay , v.d .
burkert , few body systems , suppl .
37 , 1999 .
s. a. dytman , t. p. vrana , t. s. h. lee , pin newslett , 14 , 17(1998 ) .
t. p. vrana , s. a. dytman and t. s. h. lee , submitted to phys . rev . c. m. giannini , rep .
54 , 453(1990 ) .
newton , scattering theory of waves and particles .
mcgraw - hill ( 1969 ) r.r .
longacre , j. dolbeau , nucl .
phys . , b122 , 493(1977 ) .
blatt , v.f .
weisskopf , theoretical nuclear physics , new york - london ( 1952 ) . v.
d. burkert , czech .
journ . of phys . ,
vol.46 , 627(1996 ) v. mokeev et al . , in proceedings of the `` joint ect*/jlab workshop on n
* physics and nonperturbative qcd '' , trento , italy , may 18 - 29 , 1998 , ed .
by s. simula , b. saghai , n.c .
mukhopadhyay , v.d .
burkert , few body systems , suppl . 11 , 1999 . fig 5a .
@xmath147 partial wave cross section from @xcite decomposed in resonant ( dotted lines ) background ( dashed lines ) parts according to procedure described in sect . 3.5 .
the fit results are shown by solid lines .
fig 6 . the calculated and measured@xcite angular distributions for @xmath188 reaction assuming negligible contribution for @xmath3-in- flight term ( see text sect .
dashed lines correspond to born terms alone , while solid lines represent the complete calculation with @xmath1 contribution included .
decomposition of @xmath188 reaction cross section in born terms ( dashed line ) , @xmath1 terms ( dotted line ) and interference term ( dash - dotted line ) contributions .
data are from @xcite ( squares ) and from @xcite ( circles and triangles ) .
total cross section for @xmath188 reaction at the photon point .
dashed line represents calculation results with no @xmath47 form factor and no isi and fsi absorptive corrections .
dotted line corresponds to @xmath47 strong form factor taken from @xcite and no isi and fsi absorptive corrections .
solid line represents cross section evaluation in the complete model .
data as in fig . 8 .
the calculation with pion regge trajectory exchange@xcite .
solid line represents calculations performed with born terms evaluated in single pion exchange picture .
dotted line represents the results obtained after substitution of pion exchange by pion regge trajectory exchange , dashed line corresponds to an additional modification of contact term according to prescription of @xcite .
data as in fig . 8 .
the influence of @xmath136 vertex dressing effects at the photon point .
our calculations with pdg@xcite @xmath136 vertices ( `` dressed '' ) and @xmath1 excitations excluded ( solid line ) and included ( dotted line ) in isi & fsi treatment for born terms are presented .
the calculation results with @xmath136 vertices calculated in quark models ( `` bare '' vertices ) and with @xmath1 excitation included in isi & fsi are shown by : dashed line for model @xcite , dot dashed line for model @xcite .
data as in fig . 8 .
@xmath29 dependence of @xmath136 vertex dressing effects .
our calculations with @xmath136 vertices from @xcite ( `` dressed '' ) and @xmath1 excitations excluded ( solid line ) and included ( dotted line ) in isi & fsi treatment are presented .
the calculation results with @xmath136 vertices from quark models ( `` bare '' vertices ) and with @xmath1 excitation included in isi & fsi treatment are shown by : dashed line for model @xcite , dot dashed line for model @xcite .
data from @xcite . | the @xmath0 production on the nucleon by real and virtual photons is discussed as initial step in a simple model approach for the two pion photo- and electroproduction on the nucleon , with emphasis on nucleon resonance excitation which is of interest for new facilities like tjnaf . a calculation for @xmath0 channel in resonance excitation region
is presented and compared to existing experimental data along with a discussion of physical effects that we find to be of relevance .
the calculation is proposed as a starting basis for the investigation of @xmath1 electromagnetic form factors using experimental data about two pion production by real and virtual photons . |
the large magellanic cloud ( lmc ) contains an extensive population of massive star clusters , with ages ranging from the very newly formed ( e.g. , r136 with age @xmath11 myr ) to those with ages comparable to the galactic globular clusters ( e.g. , hodge 11 ) .
these objects are extremely useful not only for studies of cluster development , but also for tracing the formation history of the lmc and its related kinematic , structural , and chemical evolution . of particular interest in this respect
is the distribution of star cluster ages .
a number of authors have explicitly demonstrated , using high quality colour - magnitude diagrams , that the oldest globular clusters in the lmc are coeval with the oldest galactic globular clusters , to within @xmath12 gyr ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
with the recent addition of ngc 1928 and 1939 , the census of objects constituting this group of ancient clusters numbers @xmath13 @xcite .
there also exists a very substantial population of young and intermediate age clusters ; however , examination of the age distribution reveals an almost complete dearth of clusters older than @xmath14 gyr but younger than the ancient ensemble ( e.g. , * ? ? ?
this curious distribution constitutes the well known lmc cluster age - gap .
recent research aimed at explaining the existence of the age - gap has focused on the impact of the periodic strong tidal interactions between the lmc and its companion the small magellanic cloud ( smc ) @xcite .
only one cluster is suspected to lie within the age gap .
this is the remote northern lmc member eso 121-sc03 , which lies at a projected angular separation of @xmath15 from the lmc centre .
eso 121-sc03 was first studied in detail by @xcite , who used @xmath16 photometry to estimate an age in the range @xmath17 gyr .
subsequent work has not altered this conclusion ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
given its apparently unique age , and interesting spatial location , eso 121-sc03 clearly constitutes an object worthy of detailed study .
we were therefore surprised to discover that it apparently has not been the subject of deep high - resolution imaging : we could only locate two independent published cmds in the literature the @xmath16 photometry of @xcite , and the washington photometry of @xcite .
neither of these studies reaches more than @xmath12 mag below the main - sequence turn - off ( msto ) , and indeed in both , the turn - off is only measured at low signal - to - noise . in this paper
we take advantage of the existence of observations of eso 121-sc03 taken with the advanced camera for surveys ( acs ) on _ hst _ as part of our recent snapshot survey of lmc and smc star clusters ( program 9891 ) , to make deep , high quality photometric measurements of this cluster .
our resulting @xmath18 cmd reaches more than @xmath0 magnitudes below the main - sequence turn - off , representing by far the deepest observations of eso 121-sc03 to date .
we use these data to conduct a detailed study of the age of eso 121-sc03 relative to two comparison clusters the young accreted sagittarius dsph globular cluster palomar 12 ( also imaged with acs in program 9891 ) , and the old galactic globular cluster 47 tuc .
we find that eso 121-sc03 is significantly younger than 47 tuc , but very similar in age to pal .
12 , confirming the earlier conclusion of @xcite and @xcite that eso 121-sc03 does indeed lie squarely within the lmc cluster age - gap .
the observations were taken during _ hst _
cycle 12 using the acs wide field channel ( wfc ) . as snapshot targets ,
the clusters were observed for only one orbit each .
this allowed two exposures to be taken per cluster
one through the f555w filter and one through the f814w filter .
details of the individual exposures are listed in table [ t : observations ] .
the acs wfc consists of a mosaic of two @xmath19 site ccds with a scale of @xmath20 arcsec per pixel , and separated by a gap of @xmath21 pixels .
each image therefore covers a field of view ( fov ) of approximately @xmath22 arcseconds .
the clusters were centred at the reference point wfc1 , located on chip 1 at position @xmath23 .
this allowed any given cluster to be observed up to a radius @xmath24 from its centre , while also ensuring the cluster core did not fall near the inter - chip gap .
our observations were made with the acs / wfc gain parameter set to @xmath25 .
this allows the full well depth to be sampled ( as opposed to only @xmath26 per cent of the full well depth for gain @xmath27 ) with only a modest increase in read noise ( @xmath28 e@xmath29 extra rms ) , thus increasing the dynamic range of the observations by greater than @xmath30 mag over that obtained with the default gain .
in addition , we offset the second image of each cluster by @xmath25 pixels in both the @xmath31 and @xmath32 directions , to help facilitate the removal of hot pixels and cosmic rays . with only two images per cluster , through different filters , it is not possible to completely eliminate the inter - chip gap using such an offset . [
cols="<,^,^,^,^,^,^,^,^ " , ] next , we applied our theoretical age calibration in order to obtain relative age estimates from the horizontal method .
the results may be seen in table [ t : ageshoriz ] .
as previously , we used the @xmath1fe@xmath2h@xmath3 = -0.83 $ ] and @xmath33fe@xmath3 = 0.0 $ ] polynomial fit for pal .
12 ; the @xmath1fe@xmath2h@xmath3 = -0.83 $ ] and @xmath33fe@xmath3 = + 0.3 $ ] polynomial fit for 47 tuc ; and the polynomials with @xmath1fe@xmath2h@xmath3 = -1.01 $ ] and @xmath33fe@xmath3 = 0.0 $ ] or @xmath34 for eso 121-sc03 . from table
[ t : ageshoriz ] , it can be seen that the horizontal indicator is apparently more sensitive to @xmath6-element abundance than is the vertical indicator , although we note that the supposed variation with @xmath6-abundance is of the same order as the quoted errors .
overall , using the horizontal indicator we found eso 121-sc03 to have an age @xmath35 per cent that of 47 tuc , in good agreement with our result from the vertical indicator .
similarly , eso 121-sc03 has an age @xmath36 per cent that of pal .
12 . for comparison , we find pal .
12 to be @xmath37 per cent of the age of 47 tuc , which is again in agreement with the relative age we derived using the vertical indicator ( @xmath38 per cent ) . based on the absolute ages for 47 tuc discussed earlier
, we find that from our horizontal method calculations , the age of eso 121-sc03 lies in the range @xmath39 gyr .
in this paper we have presented photometric measurements from acs / wfc imaging of eso 121-sc03 in the f555w and f814w passbands .
the resulting cmd represents the deepest and most precise published photometry for this unique lmc globular cluster .
we have also presented new acs / wfc photometry for the accreted sagittarius dsph globular cluster palomar 12 . using these data ,
we have conducted a detailed study of the age of eso 121-sc03 .
specifically , we have derived its age relative to the comparison clusters pal .
12 and 47 tuc using vertical and horizontal dating methods calibrated against the stellar evolution models of @xcite .
these are well sampled in metallicity and allow us to account for the uncertain run of @xmath6-elements in eso 121-sc03 , as well as the fact that pal .
12 is deficient in @xmath6-elements compared to many galactic globulars , including 47 tuc .
our main result is that eso 121-sc03 is significantly younger than 47 tuc , but rather similar in age to pal .
this conclusion is independent of assumed @xmath6-element abundance and age - dating method . taking a straight error - weighted mean of our age measurements
yields eso 121-sc03 to be @xmath7 per cent the age of 47 tuc , and @xmath8 per cent the age of palomar 12 .
palomar 12 is in turn @xmath9 per cent as old as 47 tuc , in good agreement with previous measurements . as discussed in section [ ss : vertical ] , recent results suggest the oldest galactic globular clusters have ages of @xmath40 gyr
. however , there is some disagreement in the literature as to whether 47 tuc is coeval with the oldest galactic globulars , or @xmath41 per cent younger .
this uncertainty translates to an absolute age range for eso 121-sc03 , spanning @xmath42 gyr .
similarly , the absolute age of pal .
12 lies in the range @xmath43 gyr .
typical errors on all these ages are @xmath44 gyr . our new , deep photometry and precise age estimate fully confirms eso 121-sc03 as the only known lmc cluster with an age placing it squarely within the age - gap . in this respect , it is clearly a unique object which may well be able to tell us important information about the formation and evolution of the lmc .
does this cluster represent the tail - end of ancient globular cluster formation in the lmc , just as pal .
12 and terzan 7 do for the sagittarius dsph galaxy ?
if so , then how can we reconcile this picture and the observed metallicity of eso 121-sc03 with its remote location in the very outer lmc halo ?
other ancient globular clusters at similar radii ( such as ngc 1841 and reticulum ) are considerably more metal poor . prompted by the observation that eso 121-sc03 is surrounded by a field population which shares the same properties as the cluster , @xcite speculated that eso 121-sc03 may represent part of a recently accreted lmc building block ( or dwarf galaxy ) .
we note however , that @xcite found the age and metallicity of this cluster to be consistent with the age - metallicity relation they derived from lmc field star observations ; hence they concluded the accretion of eso 121-sc03 not a necessity .
even so , it would clearly be of interest to examine in more detail the possibility that eso 121-sc03 is a captured object . by analogue with known accreted clusters in the galactic halo ( such as pal .
12 ) , we may expect three characteristics of eso 121-sc03 if it is indeed a captured cluster .
first it may exhibit a peculiar velocity , at odds with that expected from the observed dynamics of the lmc .
second , we may hope to observe the presence of some stream or anomalous stellar population surrounding the cluster .
finally , it may be possible that eso 121-sc03 possesses elemental abundance patterns different to those observed in other lmc globular clusters just as pal .
12 has been demonstrated to possess abundance ratios unlike galactic globular clusters or field stars , but in agreement with those observed for the sagittarius dsph @xcite .
regarding the first characteristic , we note that @xcite measured a radial velocity of @xmath45 km@xmath46s@xmath47 for eso 121-sc03 . at a position angle ( measured north through east ) of @xmath48 @xcite and angular separation of @xmath49 from the lmc centre
, this velocity fits very well with the observed carbon star kinematics and circularly - rotating disk model summarized by @xcite in their figure 5 .
the observed velocity of eso 121-sc03 also fits well with the rotational velocities exhibited by other old outer lmc globular clusters ( see * ? ? ?
* figure 6d ) .
these results suggest that eso 121-sc03 may not be an accreted cluster , unless its parent building - block shared the kinematics characteristic of the outer regions of the lmc .
even so , regarding the second and third points above , it would clearly be of significant interest to target the area surrounding eso 121-sc03 with deep wide - field imaging , in order to better survey the field star populations in this region
. in would also be of great value to obtain high resolution spectral measurements with the aim of determining abundance ratios in this cluster .
together such observations may offer important clues to the origin of eso 121-sc03 with strong implications for the formation and evolution of the lmc .
this paper is based on observations made with the nasa / esa hubble space telescope , obtained at the space telescope science institute , which is operated by the association of universities for research in astronomy , inc .
, under nasa contract nas 5 - 26555 .
these observations are associated with program # 9891 .
adm is grateful for financial support from a pparc postdoctoral fellowship .
many thanks also to francesca de angeli for useful discussions about blue stragglers , and the anonymous referee whose comments helped improve the paper . 99 armandroff t. e. , da costa g. s. , 1991 , aj , 101 , 1329 bekki k. , couch w. j. , beasley m. a. , forbes d. a. , chiba m. , da costa g. s. , 2004 , apj , 610 , l93 bica e. , clari j. j. , dottori h. , santos jr .
j. f. c. , piatti a. e. , 1996 , apjs , 102 , 57 bica e. , geisler d. , dottori h. , clari j. j. , piatti a. e. , santos jr . j. f. c. , 1998 , aj , 116 , 723 binney j. , tremaine s. , 1987 , galactic dynamics .
princeton univ . press ,
princeton , nj brown j. a. , wallerstein g. , zucker d. , 1997 , aj , 114 , 180 burstein d. , heiles c. , 1982 , aj , 87 , 1165 carney b. w. , 1996 , pasp , 108 , 900 chaboyer b. , 1999 , in heck a. , caputo f. , eds .
, post-_hipparcos _ cosmic candles .
kluwer , dordrecht , p. 111
cohen j. g. , 2004 , aj , 127 , 1545 da costa g. s. , armandroff t. e. , 1990 , aj , 100 , 162 da costa g. s. , armandroff t. e. , 1995 , aj , 109 , 2533 de angeli f. , piotto g. , cassisi s. , busso g. , recio - blanco a. , salaris m. , aparicio a. , rosenberg a. , 2005 , aj , 130 , 116 dirsch b. , richtler t. , gieren w. p. , hilker m. , 2000 , a&a , 360 , 133 djorgovski s. g. , 1993 , in djorgovski s. g. , meylan g. , eds .
, asp conf .
50 , structure and dynamics of globular clusters .
, san francisco , p. 373 dolphin a. e. , 2000 , pasp , 112 , 1383 geisler d. , bica e. , dottori h. , clari j. j. , piatti a. e. , santos jr
. j. f. c. , 1997 , aj , 114 , 1920 gratton r. g. , ortolani s. , 1988 , a&as , 73 , 137 gratton r. g. , bragaglia a. , carretta e. , clementini g. , desidera s. , grundahl f. , lucatello s. , 2003 , a&a , 408 , 529 harris w. e. , 1996 , aj , 112 , 1487 hill v. , franois p. , spite m. , primas f. , spite f. , 2000 , a&a , 364 , l19 johnson j. a. , bolte m. , hesser j. e. , ivans i. i. , stetson p. b. , 2004 , in mcwilliam a. , rauch m. , eds . , origin and evolution of the elements , from the carnegie observatories centennial symposia .
carnegie observatories , pasadena , p. 29
johnson j. a. , bolte m. , stetson p. b. , hesser j. e. , somerville r. s. , 1999 , apj , 527 , 199 mackey a. d. , gilmore g. f. , 2003a , mnras , 338 , 85 mackey a. d. , gilmore g. f. , 2003b , mnras , 340 , 175 mackey a. d. , gilmore g. f. , 2004 , mnras , 352 , 153 mateo m. , hodge p. , schommer r. a. , 1986 , apj , 311 , 113 olsen k. a. g. , hodge p. w. , mateo m. , olszewski e. w. , schommer r. a. , suntzeff n. b. , walker a. r. , 1998 , mnras , 300 , 665 olszewski e. w. , schommer r. a. , suntzeff n. b. , harris h. c. , 1991 , aj , 101 , 515 piotto g. , de angeli f. , king i. r. , djorgovski s. g. , bono g. , cassisi s. , meylan g. , recio - blanco a. , rich r. m. , davies m. b. , 2004 , apj , 604 , l109 puzia t. h. , perrett k. m. , bridges t. j. , 2005 , a&a , 434 , 909 recio - blanco a. , piotto g. , de angeli f. , cassisi s. , riello m. , salaris m. , pietrinferni a. , zoccali m. , aparicio a. , 2005 , a&a , 432 , 851 rich r. m. , shara m. m. , zurek d. , 2001 , aj , 122 , 842 riess a. , mack j. , 2004 , acs instrument science report 2004 - 006 rosenberg a. , saviane i. , piotto g. , held e. v. , 1998 , a&a , 339 , 61 rosenberg a. , saviane i. , piotto g. , aparicio a. , 1999 , aj , 118 , 2306 salaris m. , weiss a. , 2002 , a&a , 388 , 492 sarajedini a. , 1994 , aj , 107 , 618 sarajedini a. , lee y .- w . , lee d .- h . , 1995 , apj , 450 , 712 schlegel d. j. , finkbeiner d. p. , davis m. , 1998 , apj , 500 , 525 schommer r. a. , suntzeff n. b. , olszewski e. w. , harris h. c. , 1992 , aj , 103 , 447 sirianni m. , et al . , 2005 , pasp , 117 , 1049 stetson p. b. , hesser j. e. , smith g. h. , vandenberg d. a. , bolte m. , 1989 , aj , 97 , 1360 vandenberg d. a. , bergbusch p. a. , dowler p. d. , 2006 , apjs , 162 , 375 van der marel r. p. , alves d. r. , hardy e. , suntzeff n. b. , 2002 , aj , 124 , 2639 | we present the results of photometric measurements from images of the lmc cluster eso 121-sc03 taken with the advanced camera for surveys on the _ hubble space telescope_. our resulting colour - magnitude diagram ( cmd ) reaches @xmath0 magnitudes below the main - sequence turn - off , and represents by far the deepest observation of this cluster to date .
we also present similar photometry from acs imaging of the accreted sagittarius dsph cluster palomar 12 , used in this work as a comparison cluster . from analysis of its cmd
, we obtain estimates for the metallicity and reddening of eso 121-sc03 : @xmath1fe@xmath2h@xmath3 = -0.97 \pm 0.10 $ ] and @xmath4 , in excellent agreement with previous studies .
the observed horizontal branch level in eso 121-sc03 suggests this cluster may lie @xmath5 per cent closer to us than does the centre of the lmc .
eso 121-sc03 also possesses a significant population of blue stragglers , which we briefly discuss .
our new photometry allows us to undertake a detailed study of the age of eso 121-sc03 relative to pal .
12 and the galactic globular cluster 47 tuc .
we employ both vertical and horizontal differential indicators on the cmd , calibrated against isochrones from the victoria - regina stellar models .
these models allow us to account for the different @xmath6-element abundances in pal .
12 and 47 tuc , as well as the unknown run of @xmath6-elements in eso 121-sc03 . taking a straight error - weighted mean of our set of age measurements
yields eso 121-sc03 to be @xmath7 per cent the age of 47 tuc , and @xmath8 per cent the age of pal .
12 .
pal .
12 is @xmath9 per cent as old as 47 tuc , consistent with previous work .
our result corresponds to an absolute age for eso 121-sc03 in the range @xmath10 gyr , depending on the age assumed for 47 tuc , therefore confirming eso 121-sc03 as the only known cluster to lie squarely within the lmc age gap .
we briefly discuss a suggestion from earlier work that eso 121-sc03 may have been accreted into the lmc system .
[ firstpage ] galaxies : star clusters
globular clusters : individual : eso 121-sc03 , palomar 12 magellanic clouds |
Subsets and Splits